Browse the corpus

tative muscarinic cholinergic receptor. If so, the high doses of atropine and glycopyrrolate which were given for the prolonged period to the patient in the present report may have been able to block the muscarinic receptor in Aspergillus, thus eliminating the acetylcholine-induced protection against biofilm formation. Pathogenic fungi have the capacity to form biofilm structures that are notoriously unresponsive to anti-fungal therapies. Fungal biofilms are located all over the human host, including the upper and lower airways as well as the gastrointestinal and genitourinary tracts [13]. Thus, it must have been detrimental to lose the capacity to inhibit biofilm formation, particularly in this patient of the present report who had a history of diabetes mellitus. Invasive primary colonic aspergillosis, which had been complicated by multiple colon perforations, has been reported in the immunocompetent host without classical risk factors except for diabetes mellitus [14]. Since neutrophil-mediated innate immunity is the first line of host defense against the invasive aspergillosis and neutrophil function is impaired in diabetic patients [15], diabetes mellitus has been considered as an important risk factor for invasive aspergillosis. In the present case, diabetes mellitus may have contributed to the development of enteroinvasive aspergillosis in part.

A 78-old-man with history of diabetes mellitus and cerebral infarction was transferred to intensive care unit of Korea University Anam Hospital from nursing hospital. He presented with acute respiratory failure requiring mechanical ventilation caused by pneumonia and empyema. We inserted chest tube for empyema on the right side. A few days later, pleural effusion occurred on the left side. Thus, pleural catheter was inserted into left seventh intercostal space at the mid axillary line after marking of site using ultrasound. Chest simple radiography showed that the catheter direction had been inserted too downward (Figure 1A). A subsequent computed tomography scan revealed that the catheter first entered into the pleural space, passed through diaphragm, and the tip was located in the abdominal cavity (Figure 1B). The catheter was removed immediately with a close monitoring. After catheter removal, the patient was still stable and showed no signs or symptoms of any complication. The rate of chest tube malposition is less than 3% and 0.6% especially for small drain [1,2]. Pleural catheter malposition was very rarely reported [3]. Pleural catheter into the abdominal cavity through diaphragm is an exceptional complication. Various complications from chest tube misplacement into the abdominal cavity were reported [4]. Catheter malposition into the abdominal cavity, although rare and less severe, also can lead to injury of diaphragm and any intra-abdominal organ such as stomach, liver, spleen, bowel. Clinical manifestation includes enteric content drainage, peritonitis, bleeding, hemodynamic instability, respiratory insufficiency. The position of the pleural catheter must be checked through chest radiograph after insertion. Also, proper training and supervision are needed as well as ultrasound guidance to reduce complications associated with pleural procedures.

drainage, peritonitis, bleeding, hemodynamic instability, respiratory insufficiency. The position of the pleural catheter must be checked through chest radiograph after insertion. Also, proper training and supervision are needed as well as ultrasound guidance to reduce complications associated with pleural procedures. No potential conflict of interest relevant to this article was reported. Figure 1. (A) Chest radiography showed the pleural catheter directed downward, marked with arrow. (B) Computed tomography scan showed that the catheter first entered into the pleural space, passed through diaphragm and the tip was located in the abdominal cavity, marked with arrows. G: stomach; P: pleural effusion; B: bowel; S: spleen.

Methidathion, S-(5-methoxy-2-oxo-2,3-dihydro-1,3,4-thiadiazol-3-yl) methyl O,O-dimethyl phosphorodithioate, is a phosphorus ester used as an insecticide and an acaricide. Methidathion in itself is a poor cholinesterase inhibitor and it requires oxidation of the P-S bond to become active [1]. Unlike other organophosphate insecticides, for the poisoning of which administration of parasympatholytic agents and mechanical ventilator support are the mainstay of management, sympathetic ganglion blockade by methidathion can be profound and long-lasting. This sympathetic ganglion blockade activates renin-angiotensin axis, which produces a tremendous change in splanchnic blood flow [2]. A decrease in splanchnic blood flow allows bacterial and endotoxin translocation via gastrointestinal mucosa [3]. The presence of splanchnic ischemia can increase possibility of developing an invasive biofilm-producing fungal infection such as aspergillosis. Catecholamine inotropes, which are used to overcome septic shock or hypotension, would stimulate the growth and biofilm formation of biofilm-forming bacteria and fungus [4-6]. In addition, the effects of excessively accumulated acetylcholine on non-neuronal nicotinic acetylcholine receptors, including alpha 7 nicotinic receptor on proximal immune cells [7], and the effects of high doses of parasympatholytic drugs, which are used to overcome severe bradycardia, on the non-neuronal muscarinic acetylcholine receptor in the biofilm-producing fungus [8] would contribute to the increase in the chance of biofilm-forming fungal infection. In diabetic patients with methidathion poisoning, the fatal outcome from secondary infection of biofilm-forming organism may be inevitable.

ardia, on the non-neuronal muscarinic acetylcholine receptor in the biofilm-producing fungus [8] would contribute to the increase in the chance of biofilm-forming fungal infection. In diabetic patients with methidathion poisoning, the fatal outcome from secondary infection of biofilm-forming organism may be inevitable. Case Report A 55-year-old man was brought to emergency room after accidental ingestion of insecticide, methidathion (also known as methion in Korea) while he was drunken. According to his family member, he ingested a paper cup full of 40% original solution. He had history of non-insulin dependent diabetes mellitus. His body weight was 75 kg and height was 173 cm.

emergency room after accidental ingestion of insecticide, methidathion (also known as methion in Korea) while he was drunken. According to his family member, he ingested a paper cup full of 40% original solution. He had history of non-insulin dependent diabetes mellitus. His body weight was 75 kg and height was 173 cm. In the emergency room, he was semi-comatose, but had intermittent myoclonus-like movements. Blood pressure was 67/50 mmHg and heart rate was 48 beats per minute. His trachea was intubated and he was placed on mechanical ventilation. Initial resuscitation efforts included intravenous infusion of atropine (50 mcg/minute) after a bolus of 1 mg, norepinephrine (0.25 mcg/Kg of body weight/minute), vasopressin (0.3 units/minute), dobut-amine (30 mcg/Kg/min), epinephrine (0.25 mcg/Kg/min), and PAM-A (Pralidoxime chloride, cholinesterase reactivator, 2.5 mg/min), in addition to intravascular volume expansion with crystalloids, in efforts to keep the mean arterial pressure higher than 65 mmHg. Also given were midazolam, remifentanyl, and vecuronium. Gastric lavage was done. The initial arterial blood gas analysis while he was being mechanically ventilated with 100% of oxygen showed that pH was 6.96, PCO2 98 mmHg, PO2 85 mmHg, HCO-3 19 mEq/L, and BE -14.8. His ethanol level in blood was 144 mg/dl. Albumin was 4.4 g/dl, glucose 359 mg/dl, Hemoglobin A1c 11.4%, and total CO2 content 16.9 mEq/L. The levels of total bilirubin, aminotransferases, sodium, chloride, potassium and troponin I were normal, so were prothrombin time, international normalized ratio, and activated thromboplastin time. Leukocyte count was 17,400 cells/mm3, hemoglobin 15.4 g, hematocrit 46% and platelets 266,000/mm3 . The repeated analysis of arterial blood gases showed that pH was 7.09, PCO2 29 mmHg, PO2 66 mmHg, HCO-3 20 mEq/L, BE –11.3 and lactate 7.8 mEq/L. Urine output had been in the range of 200 to 600 mL/hour. Serum cholinesterase level was 12 (normal, 620-1,370) units/L.

in 15.4 g, hematocrit 46% and platelets 266,000/mm3 . The repeated analysis of arterial blood gases showed that pH was 7.09, PCO2 29 mmHg, PO2 66 mmHg, HCO-3 20 mEq/L, BE –11.3 and lactate 7.8 mEq/L. Urine output had been in the range of 200 to 600 mL/hour. Serum cholinesterase level was 12 (normal, 620-1,370) units/L. The patient was transferred to intensive care unit. Tests for hepatitis A, B, and C were negative, so was human immunodeficiency virus test. Free T4 level was 3.84 mcg/L. Arterial blood gases showed that pH was 7.23, PCO2 59 mmHg, PO2 78 mmHg, HCO-3 24 mEq/L and lactate 1.7 mEq/L. Blood cultures, of which blood samples were taken from the central venous line, arterial line and percutaneous punctures, were all negative. Intravenous infusion of vecuronium was discontinued. On the 2nd hospital day, intravenous infusion of epinephrine was tapered off. The infusion of norepinephrine was reduced to 0.15 mcg/kg/min, vasopressin to 0.15 units/min and dobutamine to 20 mcg/kg/min. There was no longer metabolic acidosis. Urine output decreased to 300 ml/h. On the 4th hospital day, intravenous infusion of atropine was discontinued. Glycopyrrolate was given intermittently whenever heart rate became less than 60 beats per min. Urine output was running about 100 ml/h. Daily chest radiography had been unremarkable. Sputum culture revealed the growth of Klebsiella pneumoniae, which was sensitive to all antibiotics except ampicillin. Piperacillin/tazobactam was started. On the 5th hospital day, doses of vasopressors and inotropic were not able to be reduced. Interleukin-6 was 335 pg/ml and procalcitonin 92 ng/ml. Intravenous infusion of remifentanil and midazolam had been maintained for sedation. Despite the use of high doses of vasopressors and inotropic, the patient’s hands and feet were warm and pink, and there was no metabolic acidosis, thus indicating that there was no peripheral vasoconstriction or poor peripheral perfusion. On the 6th hospital day, lactate level was 1.3 mEq/L. Fibrinogen was 966 mg/dl and d-dimer 6.53 mg/L. The absence of deep vein thrombosis was confirmed with ultrasonography and transthoracic echocardiography ruled out an acute pulmonary embolism (no right ventricular dysfunction). Total parenteral nutrition was started. Enteral feeding was not considered appropriate at this time. On the 9th hospital day, intravenous midazolam infusion was discontinued. Over the following several days, vasopressors and inotropic were able to be reduced gradually.

onary embolism (no right ventricular dysfunction). Total parenteral nutrition was started. Enteral feeding was not considered appropriate at this time. On the 9th hospital day, intravenous midazolam infusion was discontinued. Over the following several days, vasopressors and inotropic were able to be reduced gradually. On the 11th day, all vasoactive medications were completely tapered off. The patient opened eyes on command. Intravenous infusion of ketamine was added. On the 14th hospital day, ulnar nerve stimulation test using train-of-four technique to assess neuromuscular junction activity revealed the presence of depolarizing block, indicating the persistent muscle paralysis by cholinergic crisis. Sputum culture showed the growth of methicillin-resistant Staphylococcus aureus. Vancomycin was started. On the 17th hospital day, pH of arterial blood decreased 7.29 from 7.41 and PCO2 increased to 56 mmHg from 42 mmHg in the afternoon, compared with those of morning values. Leukocyte count increased to 17,750 cells/mm3 from 12,000. Platelet count also increased to 765,000/mm3 from 576,000/mm3. Ultrasonography of abdomen showed a large amount of ascites. CT scan of abdomen revealed multiple perforations of small bowel. While the exploratory laparotomy was being arranged, pulmonary artery catheter was inserted via right subclavian vein under ultrasonographic guidance. Pulmonary artery pressure was 36/18 mmHg with pulmonary capillary wedge pressure of 12 mmHg. On laparotomy, almost the whole small bowel was found gangrenous and there were multiple perforations of small bowel. The portion of bowel from the distal duodenum to the ascending colon was resected. Since there was no gross evidence of peritonitis, the peritoneum was closed. Postoperatively, vasopressors and inotropic were required overnight, but quickly tapered off by the following morning. Urine output was running in the range of 100 to 300 ml/h. Analysis of arterial blood gases showed that pH was 7.44, PCO2 44 mmHg, PO2 83 mmHg and HCO-3 29.2 mEq/L, while the patient was on mechanical ventilator with 10 cmH2O and 0.6 of FiO2. Lactate was 1.8 mEq/L. Creatinine was 1.15 mg/dl. Leukocyte count was 17,690 cells/mm3 and platelet 352. Fibrinogen level was 305 mg/dl. On the 20th hospital day, glycopyrrolate was given for the last time. Cultures from the peritoneal drain and blood were all negative.

hanical ventilator with 10 cmH2O and 0.6 of FiO2. Lactate was 1.8 mEq/L. Creatinine was 1.15 mg/dl. Leukocyte count was 17,690 cells/mm3 and platelet 352. Fibrinogen level was 305 mg/dl. On the 20th hospital day, glycopyrrolate was given for the last time. Cultures from the peritoneal drain and blood were all negative. Procalcitonin was 5.03 ng/dl, interleukin-6 92.2 pg/dl, C-reactive protein 16.32, fibrinogen 432 mg/dl and platelet count 230,000/mm3. Repeated ulnar nerve stimulation test using train-of-four technique revealed the presence of depolarizing block, indicating the persistent muscle paralysis by cholinergic crisis. Pathologic report of surgical specimen revealed the aspergillosis infection causing gangrene and abscess formation with multiple bowel perforations. Serum Aaspergillus antigen (galactomannan) was negative. Voriconazole was started. On the 27th hospital day (10th postoperative day), culture of peritoneal drain fluid showed the growth of methicillin-resistant S. aureus. Blood culture was negative. On the 31th hospital day, the second laparotomy revealed a diffuse peritonitis, gangrenous colon and abscess formation with perforations. Total colectomy was done. Postoperatively, the patient developed a severe septic shock, requiring vasopressors and inotropic in escalating doses. Intravascular volume expansion was done with guidance of pulmonary artery catheter. Thermodilution cardiac output was running in the range of 10 to 12 L/min. The patient developed a severe metabolic acidosis with rising level of lactate. On the 38th hospital day, lactate level rose to 10.7 mEq/L. Repeated test for Aspergillus antigen (galactomannan) became positive (5.42 index). The patient developed a sudden cardiac arrest. A prolonged effort of cardiopulmonary resuscitation became futile.

oped a severe metabolic acidosis with rising level of lactate. On the 38th hospital day, lactate level rose to 10.7 mEq/L. Repeated test for Aspergillus antigen (galactomannan) became positive (5.42 index). The patient developed a sudden cardiac arrest. A prolonged effort of cardiopulmonary resuscitation became futile. Discussion Compared with other organophosphate insecticides, methidathion appears to be the most dangerous organophosphate substance in producing toxicity such as a prolonged muscle paralysis requiring mechanical ventilator support, profound circulatory collapse requiring high doses of vasopressors and inotropics, severe bradycardia requiring high doses of parasympatholytic agents, and secondary enteroinvasive fungal infection.

defense against the invasive aspergillosis and neutrophil function is impaired in diabetic patients [15], diabetes mellitus has been considered as an important risk factor for invasive aspergillosis. In the present case, diabetes mellitus may have contributed to the development of enteroinvasive aspergillosis in part. Platelet count was very high just before the first laparotomy. Platelets bind plasma-opsonized hyphae and degranulate. The interaction of platelets with hyphae results in reduced hyphal galactomannan release [16]. This antifungal activity of human platelets against aspergillus species may explain the negative result of galactomannan test, which was done right after the first surgery, despite extensive invasion of small intestine by aspergillus infection. Interestingly, platelet count was not elevated around the second surgery. According to a previous study [17], a galactomannan index cutoff value ≥ 2.0 is able to classify patients with a poor outcome in the invasive aspergillosis with a sensitivity of 100%. As such, the repeated galactomannan test right after second surgery, which was 5.42 index, predicted an inevitable poor outcome in the present case.

nophosphate substance in producing toxicity such as a prolonged muscle paralysis requiring mechanical ventilator support, profound circulatory collapse requiring high doses of vasopressors and inotropics, severe bradycardia requiring high doses of parasympatholytic agents, and secondary enteroinvasive fungal infection. Acetylcholine is regarded as a classical neurotransmitter (neuronal cholinergic). Nicotinic acetylcholine mediates chemical neurotransmission at neurons, ganglia, interneurons and the motor endplate. Nicotonic receptors are ligand-gated ion channels, when bound to acetylcholine, these receptors undergo a conformational change that allows the entry of sodium ions, resulting in the depolarization of the effector cell. In the presence of acetylcholinesterase inhibitor such as organophosphate, acetylcholine is accumulated at the nicotinic receptor, resulting in the continuous depolarization of effector cell, thus leading into dysfunction of the effector cell or organ. Dysfunction of this neuronal cholinergic system causes central nervous system disorders, such as coma and seizure, sympathetic ganglion blockade leading to circulatory collapse, and the motor endplate dysfunction resulting in muscle paralysis. Muscarinic acetylcholine mediates chemical neurotransmission at neurons and effector organs such as heart, smooth muscle fibers and glands. Muscarinic receptors are G-protein coupled receptors. Thus, the accumulated acetylcholine at muscarinic receptor, in the presence of acetylcholinesterase inhibitor, results in the augmented agonistic effect (parasympathetic) on the effector cell or organ.

effector organs such as heart, smooth muscle fibers and glands. Muscarinic receptors are G-protein coupled receptors. Thus, the accumulated acetylcholine at muscarinic receptor, in the presence of acetylcholinesterase inhibitor, results in the augmented agonistic effect (parasympathetic) on the effector cell or organ. The preganglionic neuron for sympathetic ganglion is cholinergic and the acetylcholine receptor on postganglionic neuron is nicotinic. In the presence of acetylcholinesterase inhibitor, the accumulated acetylcholine induces cholinergic crisis, resulting in sympathetic ganglion blockade, thus leading to circulatory collapse. Surprisingly, in this patient of the present case report, the effect of sympathetic ganglionic blockade was so profound that the patient required very high doses of vasopressors and inotropic agent for a prolonged period, to maintain mean arterial pressure at the level of 65 mmHg or higher, despite an aggressive intravascular volume expansion. Although there was no evidence of poor peripheral blood flow or metabolic acidosis while high doses of vasopressors and inotropic agent were running, the possibility of a decrease in splanchnic blood flow could not be excluded. Because of the pathophysiologic prioritization of systemic over local splanchnic hemodynamic needs, the response of the splanchnic vasculature to circulatory shock leads to splanchnic organ ischemic injury. The selectively decrease in splanchnic blood flow induced by circulatory shock is a direct consequence of the hypersensitivity of the splanchnic vasculature to the renin-angiotensin axis as shown in Fig. 1 [2]. As such, the absence of metabolic acidosis in systemic circulation may not exclude the possibility of suboptimal splanchnic perfusion which can be reflected by low gastrointestinal intramucosal pH [9]. A decrease in splanchnic blood flow allows bacterial and endotoxin translocation via gastrointestinal mucosa, leading to endotoxemia. This endotoxemia may explain increases in inflammatory markers such as interleukin 6, procalcitonin and C-reactive protein on the 5th hospital day. Thus, the patient in the present case report may have had splanchnic ischemia, which could have contributed to a rapid progression of enteroinvasive aspergillosis.

toxemia. This endotoxemia may explain increases in inflammatory markers such as interleukin 6, procalcitonin and C-reactive protein on the 5th hospital day. Thus, the patient in the present case report may have had splanchnic ischemia, which could have contributed to a rapid progression of enteroinvasive aspergillosis. The use of catecholamine may promote the growth of biofilm-forming bacteria and fungus. Catecholamine inotropes such as norepinephrine and epinephrine have been shown to be potent stimulators of growth and film formation of Pseudomonas aeruginosa [4] and Staphylococcus species [5], via a mechanism involving inotrope delivery of transferrin ion, internalization of inotrope, and upregulation of the key pseudomonal siderophore pyoverdine, which is a mechanism to increase uptake of iron from transferrin by bacteria for growth and biofilm formation. Aspergillus species use the similar mechanism including siderophore system to increase iron uptake for growth and biofilm formation [6]. As such, catecholamine inotrope should be able to stimulate the formation of biofilm by Aspergillus fungus. Thus, it is most likely that the prolonged use of high doses of catecholamine inotropes in effort to overcome the profound hypotension in the patient of the present report would have contributed to the development of enteroinvasive aspergillosis, independent of their effects on splanchnic perfusion. Interestingly, the culture of peritoneal fluid after first bowel perforation showed the growth of biofilm-forming S. aureus instead of Gram-negative bacteria, thus supporting the above-mentioned argument.

ntributed to the development of enteroinvasive aspergillosis, independent of their effects on splanchnic perfusion. Interestingly, the culture of peritoneal fluid after first bowel perforation showed the growth of biofilm-forming S. aureus instead of Gram-negative bacteria, thus supporting the above-mentioned argument. Cholinergic signaling in non-neuronal cells is comparable to cholinergic neurotransmission. In mammalian non-neuronal cells, including those of humans, all components of the cholinergic system such as choline acetyltransferase, acetylcholine, nicotinic and muscarinic acetylcholine receptors have been demonstrated. Blockade of nicotinic and muscarinic acetylcholine receptors on non-innervated cells causes cellular dysfuncftion and/or cell death. The cholinergic anti-inflammatory pathway regulates immune responses to pathogens and is mediated by acetylcholine, involving the alpha-7 nicotinic receptors on proximal immune cells [7], although investigations into the role of the cholinergic anti-inflammatory pathway in bacterial infections have shown contradictory findings. It is most likely that the blockade of non-neuronal nicotinic acetylcholine receptors by organophosphate like methidathion would allow cholinergic proinflammatory pathway to be activated. Furthermore, a previous study has demonstrated that the acetylcholine protects against Candida albicans infection by inhibiting biofilm formation and promoting hemocyte function in a Galleria mellonella infection model [10]. Likewise, since Aspergillus fumigatus is another biofilm-producing fungus, acetylcholine would protect against Aspergillus infection by inhibiting biofilm formation and yeast-to-hyphae transition. A previous study has suggested that Candida albicans possesses putative cholinergic receptor genes [11], and pharmacological evidence suggests that this organism may possess a receptor that is homologous to human muscarinic receptors [12]. The acetylcholinesterase enzymatic activity inhibited by the neostigmine and partly physostigmine has been found in extracts from mycelium of aspergillus fungus [8]. Therefore, it can be speculated that Aspergillus also may possess putative muscarinic cholinergic receptor. If so, the high doses of atropine and glycopyrrolate which were given for the prolonged period to the patient in the present report may have been able to block the muscarinic receptor in Aspergillus, thus eliminating the acetylcholine-induced protection against biofilm formation.

a previous study [17], a galactomannan index cutoff value ≥ 2.0 is able to classify patients with a poor outcome in the invasive aspergillosis with a sensitivity of 100%. As such, the repeated galactomannan test right after second surgery, which was 5.42 index, predicted an inevitable poor outcome in the present case. The mainstay of treatment for organophosphate poisoning includes administration of atropine and mechanical ventilatory support. Reactivation of inhibited acetylcholinesterase may improve clinical pictures. Oximes retard the acetylcholinesterase aging rate. However, acetylcholinesterase aging (loss of alkyl side chain that prevents reactivation by oximes [18]) is particularly rapid with dimethyl phosphoryl compounds such as methidathion and may thwart the effective reactivation by oximes, particularly in excessive dose. In methidathion poisoning, therefore, it seems that there is no point to keep maintaining oxime after the initial dose. Removal of organophosphate from body, such as gastric lavage and hemoperfusion, may be another way to improve clinical pictures. According to the previous report [19], hemoperfusion was ineffective and did not provide real clinical benefit to the patient. Because of relatively high fat solubility and hence a large apparent volume of distribution, hemoperfusion could remove only small portion of methidathion. Also, there was evidence that the redistribution of methidathion from fat to blood could take place when plasma level diminished [19]. In contrary, gastric lavage with bicarbonate may provide some benefits, because the toxicity of organophosphates is reduced in an alkaline medium. Methidathion is stable in a neutral or a weak acid medium, but hydrolyzed in an alkaline medium [20]. The patient had had a depolarizing block even more than 5 weeks after the ingestion of methidathion. Depolarizing block is an endplate-muscular block (Phase I), from which the block will gradually change into a desensitizing Phase II neuromuscular block, as the patient recovers.

t hydrolyzed in an alkaline medium [20]. The patient had had a depolarizing block even more than 5 weeks after the ingestion of methidathion. Depolarizing block is an endplate-muscular block (Phase I), from which the block will gradually change into a desensitizing Phase II neuromuscular block, as the patient recovers. It appears that the long-lasting profound hypotension, requiring high doses of vasoconstrictors and inotropes caused the fatal enteroinvasive Aspergillus fungal infection in the present case with methidathion poisoning. The profound hypotension lead to splanchnic ischemia, the use of high doses of inotropes stimulated biofilm formation by Aspergillus fungus, and parasympatholytic drugs eliminated the acetylcholine-induced protection against biofilm formation. With the history of diabetes mellitus in this patient, the fatal outcome from secondary infection of biofilm-forming organism may have been inevitable. No potential conflict of interest relevant to this article was reported. Figure 1. Responses to angiotensin II (AII) infusion. The splanchnic vascular bed responds selectively by disproportionate vasoconstriction to a central intravenous infusion of angiotensin II, mimicking its response to cardiogenic shock. TSR: total splanchnic vascular resistance; TPR: total peripheral vascular resistance; NSR: non-splanchnic vascular resistance. Adapted from Reilly and Bulkley. Crit Care Med 1993;21(2 Suppl):S55-68 with permission of Lippincott Williams & Wilkins [2].

Introduction Predicting the prognosis of patients in intensive care unit (ICU) is important for evaluating the quality of ICU, and making decisions regarding further management [1]. The accurate methods of predicting outcomes of patients have been needed, therefore many scoring systems for predicting prognosis including mortality of patients have been proposed. However, predicting the prognosis of critically ill patients whose condition fluctuates every other moments is challenging. Recently developed assessments based on physiologic variables have limitations due to requiring many variables which are not collected for all patients admitted to ICU [1]. Therefore, it is necessary to identify noninvasive, easy means to predict the prognosis of the patients in ICU, especially for pediatric patients. The critical condition of patients admitted to the ICU is vulnerable to oxidative stresses caused by reactive oxygen species which result in injuries to cells and tissues and activating extracellular antioxidant defense network consecutively [2,3]. There are many antioxidants in extracellular fluids including albumin, which is known as one of the most potent antioxidants [4,5].

lnerable to oxidative stresses caused by reactive oxygen species which result in injuries to cells and tissues and activating extracellular antioxidant defense network consecutively [2,3]. There are many antioxidants in extracellular fluids including albumin, which is known as one of the most potent antioxidants [4,5]. Therefore, serum albumin as an indicator of the disease severity and the mortality is suggested and studied in adult patients [6,7] and it has been found that the decrease in serum albumin concentration significantly raises the mortality [8]. Hypoalbuminemia (serum albumin <3.5 g/dl) is frequently observed in patients admitted to the pediatric intensive care unit (PICU) [2], but its role in pediatric patients has not been established. In this study, we evaluated the value of serum albumin level at PICU admission as a biomarker of poor prognosis including mortality in critically ill children regardless of underlying etiology and compared serum albumin level with other mortality predictive indices which are universally used in many PICU settings.

Therefore, serum albumin as an indicator of the disease severity and the mortality is suggested and studied in adult patients [6,7] and it has been found that the decrease in serum albumin concentration significantly raises the mortality [8]. Hypoalbuminemia (serum albumin <3.5 g/dl) is frequently observed in patients admitted to the pediatric intensive care unit (PICU) [2], but its role in pediatric patients has not been established. In this study, we evaluated the value of serum albumin level at PICU admission as a biomarker of poor prognosis including mortality in critically ill children regardless of underlying etiology and compared serum albumin level with other mortality predictive indices which are universally used in many PICU settings. Materials and Methods 1) Study population We enrolled 431 pediatric patients admitted to the ICU in Severance Hospital with the exception of cardiology and surgical unit where patients who need cardiologic surgeries admit separately between January 1, 2012 and December 31, 2015. Patients were over 1 month and under 18 years of age. Patients who expired within 24 hours after ICU admission, had hepatic or renal failure, or received albumin replacement before ICU admission were excluded. We defined infection by the presence of a pathogen-proven infection (positive culture, tissue stain, or polymerase chain reaction test) or a clinical syndrome associated with a high probability of infection, such as positive findings on clinical examination, imaging, or laboratory tests (e.g., white blood cells in a normally sterile body fluid, perforated viscus, chest radiograph consistent with pneumonia, petechial or purpuric rash, or purpura fulminans) [9]. This study was approved by the Institutional Review Board of Severance Hospital (protocol No. 4-2013-0207). The informed consent was waived.

ts (e.g., white blood cells in a normally sterile body fluid, perforated viscus, chest radiograph consistent with pneumonia, petechial or purpuric rash, or purpura fulminans) [9]. This study was approved by the Institutional Review Board of Severance Hospital (protocol No. 4-2013-0207). The informed consent was waived. 2) Data collection All data were collected and analyzed retrospectively. Initial blood sample results which obtained at the time of ICU admission were analyzed. Hypoalbuminemia was defined by serum albumin concentration <3.5 g/dl. Age, sex, nutritional status, mortality, length of ICU stay, reason for admission, sepsis, septic shock, requirement for respiration support in the first 1 hour of ICU, and requirement for mechanical ventilation within 24 hours of ICU admission were collected. 3) Statistical analysis Baseline characteristics of patients were compared using independent two sample t-test and Mann-Whitney U-test, as appropriate. Groups were compared by the chi-square test for categorical variables. Univariable and multivariable logistic regression model were used to identify independent predictors of mortality and to examine the relation between serum albumin and mortality. The correlation between serum albumin and other mortality predictive indices was assessed using Spearman’s method. Survival curves were determined using Kaplan-Meier method. Statistical analyses were performed with IBM SPSS version 23.0 (IBM Corp., Armonk, NY, USA).

ine the relation between serum albumin and mortality. The correlation between serum albumin and other mortality predictive indices was assessed using Spearman’s method. Survival curves were determined using Kaplan-Meier method. Statistical analyses were performed with IBM SPSS version 23.0 (IBM Corp., Armonk, NY, USA). The predictive values of serum albumin, Pediatric Index of Mortality (PIM) 3 score, and Pediatric Risk of Mortality (PRISM) III score were compared by calculating the area under the receiver operating characteristic curve (AUC). Receiver operating characteristic curve for discrimination and comparisons of AUCs between other mortality predictive indices and albumin incorporated models (PIM 3 with albumin, and PRISM III with albumin) were used by SAS version 9.3 (SAS Institute, Cary, NC, USA). Net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were performed to assess improvement in mortality prediction after incorporating albumin to PIM 3, PRISM III by using R software version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria). A P < 0.05 was considered statistically significant.

improvement (NRI) and integrated discrimination improvement (IDI) were performed to assess improvement in mortality prediction after incorporating albumin to PIM 3, PRISM III by using R software version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria). A P < 0.05 was considered statistically significant. Results 1) Patients’ characteristics A total of 431 pediatric patients had diverse underlying diseases and reasons for ICU admission (Table 1). The most common underlying disease was neurologic disease (248 patients, 58%), and others were hemato-oncologic disease (78 patients, 18%), pulmonary disease (23 patients, 5%), gastro-intestinal disease (15 patients, 3%), cardiologic disease (11 patients, 3%), endocrinologic disease (six patients, 1%), nephrologic disease (six patients, 1%), and so on. The most common reason for ICU admission is airway-lung problem (268 patients, 62%) and other reasons were neurologic problem (64 patients, 15%), postoperation or procedure care (37 patients, 9%), hemato-oncologic problem (15 patients, 3%), nephrologic problem (14 patients, 3%), gastrointestinal problem (nine patients, 2%), metabolic diseases (three patients, 1%) and others (21 patients, 5%) sequentially. The patients’ baseline characteristics are presented in Table 2. The median age of patients was 3.1 years, and the number of male patients was 261 (60.6%). The incidence of hypoalbuminemia was 55.0% (237/431).

roblem (nine patients, 2%), metabolic diseases (three patients, 1%) and others (21 patients, 5%) sequentially. The patients’ baseline characteristics are presented in Table 2. The median age of patients was 3.1 years, and the number of male patients was 261 (60.6%). The incidence of hypoalbuminemia was 55.0% (237/431). 2) Clinical characteristics by serum albumin level Patients were divided into the normoalbuminemia (n = 194) and the hypoalbuminemia (n = 237) group on the basis of serum albumin level (Table 3). No statistically significant difference was observed in the age, sex, the nutritional status, the length of ICU stay, and the main reason for ICU admission between two groups. The patients with hypoalbuminemia had higher delta neutrophil index (2.0% vs. 0.6%, P < 0.001), C-reactive protein (33.0 mg/L vs. 5.8 mg/L, P < 0.001), lactate level (1.6 mmol/L vs. 1.2 mmol/L, P < 0.001), PIM 3 (9.23 vs. 8.36, P < 0.001), PRISM III (7.0 vs. 5.0, P < 0.001), 28-day mortality rate (24.60% vs. 9.28%, P < 0.001), incidence of septic shock (11.8% vs. 2.6%, P < 0.001), and lower hemoglobin level (10.2 g/dl vs. 11.0 g/dl, P < 0.001) and platelet level (206,000/μl vs. 341,000/μl, P < 0.001) compared to the normoalbuminemia group (Table 3).

01), PRISM III (7.0 vs. 5.0, P < 0.001), 28-day mortality rate (24.60% vs. 9.28%, P < 0.001), incidence of septic shock (11.8% vs. 2.6%, P < 0.001), and lower hemoglobin level (10.2 g/dl vs. 11.0 g/dl, P < 0.001) and platelet level (206,000/μl vs. 341,000/μl, P < 0.001) compared to the normoalbuminemia group (Table 3). 3) Clinical characteristics in infection and non-infection group by serum albumin level Patients were divided into infection group (n = 299) and non-infection group (n = 132) as we defined in the Materials and Methods. No statistically significant difference was observed in the age, sex, the nutritional status, and the length of ICU stay in both infection and non-infection group (Supplementary Tables 1 and 2). The patients with hypoalbuminemia had higher delta neutrophil index, C-reactive protein, lactate level, PIM 3, PRISM III, 28-day mortality rate, incidence of septic shock, and lower hemoglobin level and platelet level compared to the normoalbuminemia patients in both infection and non-infection group (Supplementary Tables 1 and 2). This is consistent with the result of comparison of clinical characteristics in the total group by serum albumin level (Table 3).

, incidence of septic shock, and lower hemoglobin level and platelet level compared to the normoalbuminemia patients in both infection and non-infection group (Supplementary Tables 1 and 2). This is consistent with the result of comparison of clinical characteristics in the total group by serum albumin level (Table 3). 4) Clinical characteristics in infant and the other group by serum albumin level Patients were divided by the age, infant group (n = 96) and the other group (age, ≥1 year; n = 335). No statistically significant difference was observed in the age, sex, the nutritional status, and the length of ICU stay in both age group (Supplementary Tables 3 and 4). The patients with hypoalbuminemia had higher delta neutrophil index, C-reactive protein, lactate level, PIM 3, PRISM III, 28-day mortality rate, incidence of septic shock, and lower platelet level compared to the normoalbuminemia patients in both age groups as others (Supplementary Tables 3 and 4). Hemoglobin level was statistically significant in patients who were above 1 year old, but not for infant group. 5) Serial serum albumin levels for the first 4 days in ICU We compared serial serum albumin levels for the first 4 days in ICU between survivors and non-survivors in hypoalbuminemia and normoalbuminemia group separately (Figure 1). Serum albumin level in survivors remained higher than non-survivors in both groups. Serum albumin level of non-survivors was below 3.5 g/dl in the second, third, and fourth day in ICU in both groups.

t 4 days in ICU between survivors and non-survivors in hypoalbuminemia and normoalbuminemia group separately (Figure 1). Serum albumin level in survivors remained higher than non-survivors in both groups. Serum albumin level of non-survivors was below 3.5 g/dl in the second, third, and fourth day in ICU in both groups. 6) Albumin as a predictor of mortality in PICU patients Serum albumin level in the survival group is higher than that in the non-survival group (3.4 g/dl [interquartile range, 3 to 3.8 g/dl] vs. 2.9 g/dl [interquartile range, 2.3 to 3.4 g/dl], P < 0.001) (Figure 2). Figure 3 shows survival curves according to albumin level and it shows that the normoalbuminemia group has higher survival probability than the hypoalbuminemia group. In the univariable analyses, nine variables including albumin were correlated with the mortality (P < 0.001) (Table 4). We also considered the relationship between serum albumin at admission and pre-existing mortality prediction models used in ICU. Even though the correlation coefficients were low, each model was significantly correlated with serum albumin (Supplementary Table 5). The AUC was 0.702 for serum albumin, 0.802 for PIM 3, and 0.850 for PRISM III score (Figure 4, Supplementary Table 5).

dmission and pre-existing mortality prediction models used in ICU. Even though the correlation coefficients were low, each model was significantly correlated with serum albumin (Supplementary Table 5). The AUC was 0.702 for serum albumin, 0.802 for PIM 3, and 0.850 for PRISM III score (Figure 4, Supplementary Table 5). There was no improvement of predicting mortality in the corporation of serum albumin to pre-existing mortality prediction models; however, incorporating albumin to pre-existing mortality prediction models showed a tendency to be more predictive of mortality than PIM 3, PRISM III alone (Supplementary Table 6). The PIM 3 with serum albumin provided an NRI of 21.0% (P = 0.136) and IDI of 4.3% (P = 0.158), and the PRISM III with serum albumin yielded an NRI of 24.5% (P = 0.388) and IDI of 2.1% (P = 0.416) (Supplementary Table 7). Discussion We showed that serum albumin concentration at ICU admission was higher in survivors compared with nonsurvivors in children. Patients with hypoalbuminemia at ICU admission had higher mortality rate. In addition, serum albumin concentration was negatively correlated with pre-existing mortality prediction models including PIM 3 and PRISM III.

albumin concentration at ICU admission was higher in survivors compared with nonsurvivors in children. Patients with hypoalbuminemia at ICU admission had higher mortality rate. In addition, serum albumin concentration was negatively correlated with pre-existing mortality prediction models including PIM 3 and PRISM III. The prevalence of hypoalbuminemia in critically ill children is about 33%–57% from previous studies [10-12]. Hypoalbuminemia has long been a predictor of poor outcome, such as mortality, morbidity, and prolonged ICU and hospital stay [5]. There are many studies that low serum albumin concentration reflects disease severity and prognosis in critically ill adult patients [13-16]. Regarding pediatric patients, studies on this issue were limited to only some specific circumstances, such as cardiac surgery children, pediatric renal transplant recipients, showing that hypoalbuminemia is a risk factor for cardiac surgery and graft failure [17,18]. Serum albumin concentration as a predictor of clinical outcomes of critically ill children in general circumstances is so far inconclusive [10,11].

ances, such as cardiac surgery children, pediatric renal transplant recipients, showing that hypoalbuminemia is a risk factor for cardiac surgery and graft failure [17,18]. Serum albumin concentration as a predictor of clinical outcomes of critically ill children in general circumstances is so far inconclusive [10,11]. Albumin plays important role as an osmotic pressure determinant balancing between intravascular and interstitial space, a non-specific carrier protein, and an antioxidant by scavenging of oxygen free radicals to protect against oxidant injuries [19,20]. The important determinants of serum albumin concentration in critically ill patients are synthetic rate and leakage of albumin to interstitial space from intravascular space [19,21]. As the clinical course of critically ill patients progresses, the distribution of albumin from intravascular to interstitial spaces occurs resulting from increases of capillary leakage by the response of systemic inflammation [20]. Since the leakage of albumin into the extravascular spaces is related to the degree of systemic inflammation [22], the development of hypoalbuminemia is liable to be seen in patients with poor clinical status. Owing to potency of albumin as an antioxidant agent, in addition to this kinetic and dynamic alteration of albumin in critically ill patients, hypoalbuminemia is associated with poor clinical outcomes such as mortality, morbidity and prolonged hospital stay in critically ill patients [5,23].

with poor clinical status. Owing to potency of albumin as an antioxidant agent, in addition to this kinetic and dynamic alteration of albumin in critically ill patients, hypoalbuminemia is associated with poor clinical outcomes such as mortality, morbidity and prolonged hospital stay in critically ill patients [5,23]. Mortality is the worst outcome of clinical management in ICU setting. Early recognition of probability of mortality in critically ill patients can help to make prompt and precise decision to give the best treatment to patients by risk stratification. Therefore, predicting mortality risk is a major concern of ICU care. Several predictive indices have been suggested to predict the risk of death in critically ill children of ICU, but some predictive indices are too complicated demanding too much information to calculate the mortality risk. As a result, there has been an increasing need to establish a clinical value to use in children in ICU easily. Several biomarkers such as delta neutrophil index, C-reactive protein, and thrombocytopenia have been suggested as predictors of poor prognosis including mortality in children in ICU [24-26].

e mortality risk. As a result, there has been an increasing need to establish a clinical value to use in children in ICU easily. Several biomarkers such as delta neutrophil index, C-reactive protein, and thrombocytopenia have been suggested as predictors of poor prognosis including mortality in children in ICU [24-26]. We designed this study to investigate whether hypoalbuminemia at admission can predict prognosis including mortality in pediatric patients in ICU. Comparing hypoalbuminemia group of patients and normoalbuminemia group of patients, we showed that hypoalbuminemia at admission to ICU was associated with the higher mortality, regardless of the underlying disease. Furthermore, even though there was no statistical significance, incorporation of serum albumin at admission to other predictive indices showed a tendency to be more predictive in this study. This shows that serum albumin concentration, which could be obtained through a simple, sensitive, low cost assay, could help assessing the current status and predicting outcomes of critically ill pediatric patients in ICU. There are some limitations in the current study. First, this study was limited to the single center using the data collected retrospectively. Since serum albumin level below 2.5 g/dl were replaced and above 2.5 g/dl were not in this retrospective data, this study was not able to evaluate the effectiveness of albumin replacement. Therefore, further studies might be needed in a multicenter setting using prospective data.

ing the data collected retrospectively. Since serum albumin level below 2.5 g/dl were replaced and above 2.5 g/dl were not in this retrospective data, this study was not able to evaluate the effectiveness of albumin replacement. Therefore, further studies might be needed in a multicenter setting using prospective data. In conclusion, hypoalbuminemia at admission was associated with higher mortality in children in ICU. We suggest that hypoalbuminemia at admission could be used as a biomarker of poor prognosis of children in ICU, regardless of underlying etiology. No potential conflict of interest relevant to this article was reported. Supplementary Materials The online-only Supplement data are available with this article online: https://doi.org/10.4266/kjccm.2017.00437. Figure 1. Albumin level in the first 4 days in intensive care unit in (A) normoalbuminemia group and (B) in hypoalbuminemia group. The box and plots represent mean and standard deviation. Figure 2. Albumin level between non-survival and survival group. Serum albumin level in the survival group was higher than the non-survival group (3.4 g/dl [interquartile range, 3 to 3.8 g/dl] vs. 2.9 g/dl [interquartile range, 2.3 to 3.4 g/dl], P < 0.001). The error bar represents the median and interquartile range for each group.

umin level between non-survival and survival group. Serum albumin level in the survival group was higher than the non-survival group (3.4 g/dl [interquartile range, 3 to 3.8 g/dl] vs. 2.9 g/dl [interquartile range, 2.3 to 3.4 g/dl], P < 0.001). The error bar represents the median and interquartile range for each group. Figure 3. Survival probability according to albumin level. Kaplan-Meier survival estimate for patients according to serum albumin level (albumin <3.5 g/dl vs. albumin ≥3.5 g/dl). This survival curve showed that the hypoalbuminemia group had lower survival probability than the normoalbuminemia group. PICU: pediatric intensive care unit. Figure 4. Receiver operating characteristic curves for mortality between albumin and intensive care unit mortality scoring systems. Area under the receiver operating characteristic curve: PRISM III, 0.850 (95% CI, 0.801 to 0.899); PIM 3, 0.802 (95% CI, 0.747 to 0.858); albumin, 0.702 (95% CI, 0.633 to 0.772). PRISM: Pediatric Risk of Mortality; PIM: Pediatric Index of Mortality; CI: confidence interval. Table 1. Underlying disease and main reason for ICU admission of pediatric patients

Figure 4. Receiver operating characteristic curves for mortality between albumin and intensive care unit mortality scoring systems. Area under the receiver operating characteristic curve: PRISM III, 0.850 (95% CI, 0.801 to 0.899); PIM 3, 0.802 (95% CI, 0.747 to 0.858); albumin, 0.702 (95% CI, 0.633 to 0.772). PRISM: Pediatric Risk of Mortality; PIM: Pediatric Index of Mortality; CI: confidence interval. Table 1. Underlying disease and main reason for ICU admission of pediatric patients Variable No. (%) (n = 431) Underlying disease Neurologic disease 248 (57.54) Hemato-oncologic disease 78 (18.10) Pulmonary disease 23 (5.34) Gastro-intestinal disease 15 (3.48) Cardiologic disease 11 (2.55) Endocrinologic disease 6 (1.39) Nephrologic disease 6 (1.39) Metabolic disease 1 (0.23) No underlying disease 43 (10.00) Main reason for ICU admission Airway-lung problem 268 (62.18) Neurologic problem 64 (14.85) Postoperation and procedure care 37 (8.58) Hemato-oncologic emergency 15 (3.48) Nephrologic problem 14 (3.25) Gastrointestinal problem 9 (2.09) Metabolic disease 3 (0.70) Othersa 21 (4.87) ICU: intensive care unit. a Endocrinologic emergency, close monitoring, cardiovascular disease. Table 2. Subjects’ characteristics

Variable No. (%) (n = 431) Underlying disease Neurologic disease 248 (57.54) Hemato-oncologic disease 78 (18.10) Pulmonary disease 23 (5.34) Gastro-intestinal disease 15 (3.48) Cardiologic disease 11 (2.55) Endocrinologic disease 6 (1.39) Nephrologic disease 6 (1.39) Metabolic disease 1 (0.23) No underlying disease 43 (10.00) Main reason for ICU admission Airway-lung problem 268 (62.18) Neurologic problem 64 (14.85) Postoperation and procedure care 37 (8.58) Hemato-oncologic emergency 15 (3.48) Nephrologic problem 14 (3.25) Gastrointestinal problem 9 (2.09) Metabolic disease 3 (0.70) Othersa 21 (4.87) ICU: intensive care unit. a Endocrinologic emergency, close monitoring, cardiovascular disease. Table 2. Subjects’ characteristics Variable Value (n = 431) Age (yr) 3.1 (1.1–7.9) Male sex 261 (60.56) Nutritional state Well nourished 275 (63.81) Malnourished (weight for age <3%) 149 (34.57) Obese (weight for age ≥97%) 7 (1.62) Requirement for respiration support within 1 hour ICU admission 384 (89.10) Requirement for mechanical ventilation within 24 hours ICU admission 334 (77.49) Laboratory variable White blood cell (/µL) 10,930 (6,240–17,480) Absolute neutrophil count (/µL) 7,680 (3,830–13,253) Delta neutrophil index (%) 1.5 (0.0–5.2) Hemoglobin (g/dl) 10.7 (9.2–11.9) Platelets (103/µl) 271 (151–398) Erythrocyte sedimentation rate (mm/h) 21.0 (6.0–57.0) C-reactive protein (mg/L) 13.6 (2.6–57.9) Lactate (mmol/L) 1.6 (0.9–3.1) Serum albumin (g/dl) 3.4 (2.9–3.7) Hypoalbuminemia (<3.5 g/dl) 237 (54.99) Length of ICU stay (d) 9.0 (4.0–18.0) ICU scoring system PIM 3 8.46 (4.30–22.63) PRISM III 5.5 (2.0–12.0) Mortality 84 (19.49) 28-Day mortality 76 (17.63) Sepsis 82 (19.02) Septic shock 33 (7.65) Values are presented as median (interquartile range) or number (%).

l) 3.4 (2.9–3.7) Hypoalbuminemia (<3.5 g/dl) 237 (54.99) Length of ICU stay (d) 9.0 (4.0–18.0) ICU scoring system PIM 3 8.46 (4.30–22.63) PRISM III 5.5 (2.0–12.0) Mortality 84 (19.49) 28-Day mortality 76 (17.63) Sepsis 82 (19.02) Septic shock 33 (7.65) Values are presented as median (interquartile range) or number (%). ICU: intensive care unit; PIM: Pediatric Index of Mortality; PRISM: Pediatric Risk of Mortality. Table 3. Characteristics in normoalbuminemia and hypoalbuminemia group

l) 3.4 (2.9–3.7) Hypoalbuminemia (<3.5 g/dl) 237 (54.99) Length of ICU stay (d) 9.0 (4.0–18.0) ICU scoring system PIM 3 8.46 (4.30–22.63) PRISM III 5.5 (2.0–12.0) Mortality 84 (19.49) 28-Day mortality 76 (17.63) Sepsis 82 (19.02) Septic shock 33 (7.65) Values are presented as median (interquartile range) or number (%). ICU: intensive care unit; PIM: Pediatric Index of Mortality; PRISM: Pediatric Risk of Mortality. Table 3. Characteristics in normoalbuminemia and hypoalbuminemia group Variable Albumin <3.5 g/dl (n = 237) Albumin ≥3.5 g/dl (n = 194) P-value Clinical characteristic Age (yr) 4.0 (1.6–8.8) 2.0 (0.9–6.7) NS Male sex 141 (59.49) 120 (61.86) NS Nutritional state NS Well nourished 151 (63.71) 124 (63.92) Malnourished 83 (35.02) 66 (34.02) Obese (weight for age ≥97%) 3 (1.27) 4 (2.06) Main reason for ICU admission Airway-lung problem 151 (63.71) 117 (60.31) NS Neurologic problem 24 (10.13) 40 (20.62) NS Postoperation and procedure care 22 (9.28) 15 (7.73) NS Hemato-oncologic emergency 10 (4.22) 5 (2.58) NS Nephrologic problem 9 (3.80) 5 (2.58) NS Gastrointestinal problem 8 (3.38) 1 (0.52) NS Metabolic disease 1 (0.42) 2 (1.03) NS Othersa 12 (5.06) 9 (4.64) NS Laboratory variable White blood cell (/µL) 9,190 (5,580–15,960) 12,640 (7,685–18,603) NS Absolute neutrophil count (/µL) 6,700 (3,303–11,763) 8,922 (4,803–14,223) NS Delta neutrophil index (%) 2.0 (0.0–7.1) 0.6 (0.0–3.6) <0.001 Hemoglobin (g/dl) 10.2 (8.8–11.4) 11.0 (9.8–12.1) <0.001 Platelets (103/µl) 206 (96–335) 341 (232–475) <0.001 Erythrocyte sedimentation rate (mm/h) 23.0 (7.0–64.5) 19.5 (4.75–48.0) NS C-reactive protein (mg/L) 33.0 (8.5–97.5) 5.8 (0.9–24.2) <0.001 Lactate (mmol/L) 1.6 (0.9–3.1) 1.2 (0.7–1.9) <0.001 Length of ICU stay (d) 10.0 (4.0–19.0) 7.0 (4.0–17.0) NS ICU scoring system PIM 3 9.23 (4.52–29.48) 8.36 (3.80–18.80) <0.001 PRISM III 7.0 (3.0–13.0) 5.0 (0–10.0) <0.001 Mortality 64 (27.00) 20 (10.31) <0.001 28-Day mortality 59 (24.60) 18 (9.28) <0.001 Sepsis 53 (22.36) 29 (14.95) NS Septic shock 28 (11.81) 5 (2.58) <0.001 Values are presented as median (interquartile range) or number (%).

em PIM 3 9.23 (4.52–29.48) 8.36 (3.80–18.80) <0.001 PRISM III 7.0 (3.0–13.0) 5.0 (0–10.0) <0.001 Mortality 64 (27.00) 20 (10.31) <0.001 28-Day mortality 59 (24.60) 18 (9.28) <0.001 Sepsis 53 (22.36) 29 (14.95) NS Septic shock 28 (11.81) 5 (2.58) <0.001 Values are presented as median (interquartile range) or number (%). NS: non-specific; ICU: intensive care unit; PIM: Pediatric Index of Mortality; PRISM: Pediatric Risk of Mortality. a Endocrinologic emergency, close monitoring, cardiovascular disease. Table 4. Univariable logistic regression for independent factor in mortality prediction Variable Odds ratio 95% Confidence interval P-value Age 1.058 1.004–1.114 0.034 Sex 0.932 0.560–1.551 0.787 Pediatric Index of Mortality 3 1.031 1.023–1.038 <0.001 Pediatric Risk of Mortality III 1.185 1.139–1.232 <0.001 Albumin 0.34 0.233–0.496 <0.001 C-reactive protein 1.008 1.005–1.010 <0.001 Delta neutrophil index 1.079 1.054–1.105 <0.001 Lactate 1.245 1.167–1.328 <0.001 Platelet 0.995 0.993–0.997 <0.001

Introduction Critically ill patients in surgical intensive care units (ICUs) have been severely injured, present with acute surgical emergencies, require prolonged and complex elective surgical procedures, or have severe underlying medical conditions [1]. They have risks for developing all of the potential problems that afflict nonsurgical patients in ICU and also have risks for complications related to surgical procedures [2]. Pulmonary complications, such as atelectasis, pneumonia, and pulmonary edema, which are associated with in-hospital mortality and length of hospital stay, frequently develop in critically ill surgical patients [1,2]. Although portable chest radiography and physical examination of the respiratory system are routinely performed for most patients in surgical ICUs, the application of chest radiography in the ICU for the detection of pulmonary complications is limited because of low diagnostic yields [3,4].

ill surgical patients [1,2]. Although portable chest radiography and physical examination of the respiratory system are routinely performed for most patients in surgical ICUs, the application of chest radiography in the ICU for the detection of pulmonary complications is limited because of low diagnostic yields [3,4]. Lung ultrasound (LUS) has emerged in recent years as a bedside noninvasive test [5-8]. The use of bedside LUS in the ICU is increasing due to its ease of use, accessibility, safety profile, and immediate feedback [6-8]. Previous studies showed that LUS is useful for accurate diagnosis of various anatomical abnormalities including pleural effusion, diffuse interstitial syndrome, pneumothorax, pulmonary consolidation, and pulmonary abscess [5-12]. Furthermore, previous studies focused on comparing LUS with chest radiography and showed LUS had a high degree of diagnostic accuracy [8,12-14]. While many studies focused on diagnostic performance of LUS for patients in ICUs, none of the studies have provided insight into how to actually use LUS in a surgical ICU. The objective of this study was thus to investigate the clinical use of performing LUS in a surgical ICU.

h degree of diagnostic accuracy [8,12-14]. While many studies focused on diagnostic performance of LUS for patients in ICUs, none of the studies have provided insight into how to actually use LUS in a surgical ICU. The objective of this study was thus to investigate the clinical use of performing LUS in a surgical ICU. Materials and Methods 1) Patients We retrospectively reviewed the medical records of 262 patients who had previously undergone LUS at Inje University Ilsan Paik Hospital between May 2016 and December 2016. After excluding 98 patients who underwent LUS only at the posterolateral point for identification of pleural effusion, we performed LUS on 164 patients at standardized points according to bedside LUS in emergency (BLUE) protocol [7]. Among them, 67 patients who underwent LUS in the surgical ICU were enrolled in this study, excluding 97 patients who underwent LUS in the medical ICU (n = 81) and the ward (n = 16) (Figure 1). The Institutional Review Board of Inje University Ilsan Paik Hospital approved this study, including the review and publishing of information obtained from patient records (IRB No. 2016-12-023). The requirement for informed consent was waived for the use of patient medical data because all personally identifying information was removed before analysis.

Inje University Ilsan Paik Hospital approved this study, including the review and publishing of information obtained from patient records (IRB No. 2016-12-023). The requirement for informed consent was waived for the use of patient medical data because all personally identifying information was removed before analysis. 2) Measurement Patient medical records were reviewed to obtain data on demographic features, comorbidities, medical conditions, symptoms, laboratory data, radiologic findings, and LUS findings. Comorbidities including coronary artery disease, atrial fibrillation, chronic obstructive lung disease, rib fracture, chronic kidney disease, and underlying malignancy as well as whether patients underwent lung surgery previously were reviewed. Arterial blood gas analysis was performed for all patients undergoing LUS. To diagnose and manage pulmonary complications, all patients consulted with one physician (HKK), intensivist, and pulmonologist. They were selected for LUS based on clinical needs for diagnosis of pulmonary complications, which occurred during the intensive care period in surgical ICU. Indications for LUS included hypoxemia, abnormal chest radiographs without hypoxemia, fever without both hypoxemia and abnormal chest radiographs, and difficult weaning. Difficult weaning is identified for patients in two situations, i.e., if the patient fails initial weaning and requires up to three spontaneous breathing tests (SBTs) or if the patient requires up to 7 days from the first SBT to achieve successful weaning [15].

mia and abnormal chest radiographs, and difficult weaning. Difficult weaning is identified for patients in two situations, i.e., if the patient fails initial weaning and requires up to three spontaneous breathing tests (SBTs) or if the patient requires up to 7 days from the first SBT to achieve successful weaning [15]. Chest radiography was performed in all study patients in the morning prior to receiving LUS. All chest radiographs were obtained in the anteroposterior view using mobile equipment. Chest radiograph findings were divided into three classifications including unilateral air space opacification, bilateral air space opacification, and no significant abnormality. The final diagnosis was formulated by two pulmonologists, who reviewed the clinical manifestations, radiologic findings, laboratory data, LUS findings, and clinical progression.

Chest radiography was performed in all study patients in the morning prior to receiving LUS. All chest radiographs were obtained in the anteroposterior view using mobile equipment. Chest radiograph findings were divided into three classifications including unilateral air space opacification, bilateral air space opacification, and no significant abnormality. The final diagnosis was formulated by two pulmonologists, who reviewed the clinical manifestations, radiologic findings, laboratory data, LUS findings, and clinical progression. 3) Lung ultrasound measurements LUS was performed with the obtained clinical information including symptoms, laboratory data, and chest radiographs. All LUS examinations were performed by two well-trained experts (HKK and HJS) with a 13-MHz linear and 5-MHz curved array probe place over six standardized points for the BLUE protocol using an Acuson X300 ultrasound system (Siemens Healthineers, Erlangen, Germany) [7]. For patients who experienced difficult weaning during SBT [16], diaphragmatic movement for inspiration and expiration was additionally measured at one of the lower intercostal spaces in the right anterior axillary line for the right diaphragm and the left midaxillary line for the left diaphragm. Ultrasonographic diaphragmatic dysfunction was diagnosed if an excursion was <10 mm or negative.

atic movement for inspiration and expiration was additionally measured at one of the lower intercostal spaces in the right anterior axillary line for the right diaphragm and the left midaxillary line for the left diaphragm. Ultrasonographic diaphragmatic dysfunction was diagnosed if an excursion was <10 mm or negative. According to the BLUE protocol, at the four anterior chest walls, lung sliding with predominant A-lines, bilateral lung rockets, anterior lung rockets associated with abolished lung sliding, and unilateral lung rockets defined the A-profile, the B-profile, the B’-profile, and the A/B profile, respectively [7]. Anterior lung consolidation, regardless of number and size, defined the C-profile. At the two posterior chest walls, lung consolidations and pleural effusions are evaluated [7,11-13]. In LUS, pulmonary edema was defined as the presence of the B-profile with or without bilateral pleural effusion [7]. Pneumonia was defined as the presence of the B’- profile, the A/B profile, the C-profile, and/or lung consolidations regardless of unilateral effusion at posterior chest wall [7,10-12,17]. Diffuse interstitial pattern with anterior consolidation was defined as the presence of the B-profile and the C-profile at least one anterior chest wall (Table 1). Pneumothorax was approached in LUS using the sole abolition of lung sliding [18]. Pleural effusion was defined as the presence of an anechoic space between the parietal and visceral pleura [12].

rior consolidation was defined as the presence of the B-profile and the C-profile at least one anterior chest wall (Table 1). Pneumothorax was approached in LUS using the sole abolition of lung sliding [18]. Pleural effusion was defined as the presence of an anechoic space between the parietal and visceral pleura [12]. A total of 107 LUS examinations were performed in 67 patients. When subjects underwent more than one LUS examination during the study period, data from only the first examination were used in the analysis. 4) Statistical analysis Baseline characteristics are presented as medians and interquartile ranges (IQRs) for continuous variables and as numbers (%) for categorical variables. The data were compared using the Mann-Whitney U-test for continuous variables and Pearson’s chi-square test or Fisher exact test for categorical variables. All statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA).

ontinuous variables and as numbers (%) for categorical variables. The data were compared using the Mann-Whitney U-test for continuous variables and Pearson’s chi-square test or Fisher exact test for categorical variables. All statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA). Results 1) Baseline characteristics The baseline characteristics of 67 patients are summarized in Table 2. The study patients included 47 males (70.1%) and 20 females (29.9%), with a median age of 68 years (IQR, 55 to 78 years). The median number of examined LUS per person during the study period was 1 (IQR, 1 to 2). Of the 67 patients, 33 (49.3%), 15 (22.4%), 10 (15.0%), 6 (9.0%), 2 (3.0%), and 1 (1.5%) were neurosurgical, thoracic surgical, orthopedic surgical, abdominal surgical, obstetric surgical, and head and neck surgical patients, respectively. Twenty-three patients (34.3%) had trauma. Routine surgical procedures were performed for 49 patients (73.1%). 2) Indications and findings of performed LUS The indication for LUS included hypoxemia (n = 44, 65.7%), abnormal chest radiographs without hypoxemia (n = 17, 25.4%), fever without both hypoxemia and abnormal chest radiographs (n = 4, 6.0%), and difficult weaning (n = 2, 3.0%) (Figure 2).

Results 1) Baseline characteristics The baseline characteristics of 67 patients are summarized in Table 2. The study patients included 47 males (70.1%) and 20 females (29.9%), with a median age of 68 years (IQR, 55 to 78 years). The median number of examined LUS per person during the study period was 1 (IQR, 1 to 2). Of the 67 patients, 33 (49.3%), 15 (22.4%), 10 (15.0%), 6 (9.0%), 2 (3.0%), and 1 (1.5%) were neurosurgical, thoracic surgical, orthopedic surgical, abdominal surgical, obstetric surgical, and head and neck surgical patients, respectively. Twenty-three patients (34.3%) had trauma. Routine surgical procedures were performed for 49 patients (73.1%). 2) Indications and findings of performed LUS The indication for LUS included hypoxemia (n = 44, 65.7%), abnormal chest radiographs without hypoxemia (n = 17, 25.4%), fever without both hypoxemia and abnormal chest radiographs (n = 4, 6.0%), and difficult weaning (n = 2, 3.0%) (Figure 2). A total of 55 patients were diagnosed with pulmonary edema (n = 27, 41.8%), pneumonia (n = 20, 29.9%), diffuse interstitial pattern with anterior consolidation (n = 6, 10.9%), pneumothorax with effusion (n = 1, 1.5%), and diaphragm dysfunction (n = 1, 1.5%), respectively, via LUS. Among 27 patients who were diagnosed with pulmonary edema via LUS, 21 (77.8%) had bilateral pleural effusion. Among 20 patients who were diagnosed with pneumonia by LUS, both the A/B-profile and C-profile were observed in 16 patients (80%) (Table 3).

.5%), and diaphragm dysfunction (n = 1, 1.5%), respectively, via LUS. Among 27 patients who were diagnosed with pulmonary edema via LUS, 21 (77.8%) had bilateral pleural effusion. Among 20 patients who were diagnosed with pneumonia by LUS, both the A/B-profile and C-profile were observed in 16 patients (80%) (Table 3). One patient, who was confirmed to have a pneumothorax with effusion via LUS, was diagnosed with hemopneumothorax using pleural fluid analysis. For 12 patients, LUS did not identify any lung complications. Among six patients who had hypoxemia, three patients had difficulty in self-expectoration of secretion, and three patients were diagnosed with sepsis. One patient, who had an abnormal chest radiograph without hypoxemia, was not finally diagnosed with pulmonary complications. Among patients who had fever without both hypoxemia and abnormal chest radiographs, two patients had central fever after cerebral hemorrhage, and one patient had a deep neck infection. One patient with difficult weaning had cardiac load caused by systolic and diastolic dysfunction.

agnosed with pulmonary complications. Among patients who had fever without both hypoxemia and abnormal chest radiographs, two patients had central fever after cerebral hemorrhage, and one patient had a deep neck infection. One patient with difficult weaning had cardiac load caused by systolic and diastolic dysfunction. 3) Final diagnosis based on the location of space opacification on chest radiographs and LUS findings Based on the chest radiography findings, patients had unilateral air space opacification (n = 22, 32.8%), bilateral air space opacification (n = 33, 49.3%), and no abnormal findings (n = 12, 17.9%). Among 22 patients with unilateral air space opacification, one patient (4.5%) was finally diagnosed with pulmonary edema, and three patients (13.6%) were not diagnosed with any pulmonary complications. Among 33 patients with bilateral air space opacification, five patients (15.2%) were diagnosed with pneumonia regardless of pulmonary edema. Further, of 12 patients without chest radiograph abnormalities, three patients (25%) were diagnosed with pulmonary complications (Table 4). The image comparing chest radiograph and LUS in one case among three patients is shown in Figure 3. Based on LUS diagnosis, all 27 patients with pulmonary edema for LUS were finally diagnosed with pulmonary edema. Among 20 patients with pneumonia for LUS, 16 patients (80%) and four patients (20%) were finally diagnosed with pneumonia and atelectasis, respectively. The 12 patients whose LUS abnormal findings were inconclusive were not diagnosed with any pulmonary complications (Table 5).

re finally diagnosed with pulmonary edema. Among 20 patients with pneumonia for LUS, 16 patients (80%) and four patients (20%) were finally diagnosed with pneumonia and atelectasis, respectively. The 12 patients whose LUS abnormal findings were inconclusive were not diagnosed with any pulmonary complications (Table 5). Discussion In the present study, we showed LUS was used for evaluation of hypoxemia, abnormal chest radiograph findings without hypoxemia, fever, and difficult weaning in 67 surgically ill patients. LUS was helpful for diagnosis of pneumonia, atelectasis, pulmonary edema, or a combination of these diseases in 53 patients (79.2%). In addition, based on the location of space opacification on chest radiographs, LUS was used to diagnose pneumonia or atelectasis in four patients (17.4%) with bilateral abnormality and normal findings, and pulmonary edema in two patients (12.8%) with unilateral abnormality and normal finding.

s in 53 patients (79.2%). In addition, based on the location of space opacification on chest radiographs, LUS was used to diagnose pneumonia or atelectasis in four patients (17.4%) with bilateral abnormality and normal findings, and pulmonary edema in two patients (12.8%) with unilateral abnormality and normal finding. The presence of infiltrates on a chest radiograph is considered the definitive marker for diagnosing pneumonia when clinical and microbiologic features are supportive [19-21]. Lung infiltrates may be difficult to identify in critically ill patients for whom only portable chest radiography is available [20]. Furthermore, clinical symptoms including fever, cough, and sputum production may be difficult to identify in critically ill surgical patients. Decreased mental status in ICU patients makes it hard to express symptoms of pneumonia [1,2]. Moreover, both hypothermia and unreliable temperature due to intervention such as renal replacement therapy could mask fever [22,23]. Previous studies showed that the sensitivity and specificity for the diagnosis of pneumonia using LUS were 81%–97% and 88%–97%, respectively, suggesting that LUS is a useful test for diagnosing pneumonia when chest radiographic results are negative or inconclusive [13,14,24-28]. In this study, LUS detected lung parenchymal consolidation with air bronchogram in one patient without chest radiograph abnormalities.

ing LUS were 81%–97% and 88%–97%, respectively, suggesting that LUS is a useful test for diagnosing pneumonia when chest radiographic results are negative or inconclusive [13,14,24-28]. In this study, LUS detected lung parenchymal consolidation with air bronchogram in one patient without chest radiograph abnormalities. Pulmonary edema is secondary to accumulation of fluid in the lung interstitium or alveolar space [29]. Patients with extensive traumatic or surgical tissue injury, critical illness, or sepsis require replacement fluid therapy in addition to maintenance therapy, but fluid accumulation leading to a positive fluid balance could increase pulmonary edema in critically ill surgical patients [30]. Although chest radiographic findings of pulmonary edema include diffuse infiltrates regardless of pleural effusion and bilateral alveolar filling pattern [29], portable chest radiography in the critical care setting often yields inaccurate images [31,32]. Previous studies showed the positive relationship between diffuse lung interstitial involvement and bilateral sonographic B-lines [31,33-35]. In this study, two patients who showed no suspicious symptoms of pulmonary edema on chest radiographs were diagnosed with pulmonary edema during LUS. Additionally, noncardiogenic pulmonary edema is associated primarily with other clinical disorders, including pneumonia [29]. The final diagnosis for two patients in this study was pulmonary edema and pneumonia.

suspicious symptoms of pulmonary edema on chest radiographs were diagnosed with pulmonary edema during LUS. Additionally, noncardiogenic pulmonary edema is associated primarily with other clinical disorders, including pneumonia [29]. The final diagnosis for two patients in this study was pulmonary edema and pneumonia. The usefulness of LUS for diagnosis of pleural effusions has been previously demonstrated by several studies that performed LUS for ICU patients [32,36-38]. LUS helps distinguish between effusion and lung consolidations and is reliable for the identification of pleural effusion when compared to portable chest radiography [32,37]. The appearance of pleural effusion on LUS scans can suggest the nature of the fluid although a definitive diagnosis requires a thoracentesis to allow biochemical and microbiological analyses [39]. In addition, the ultrasound feature of multiple comet-tail artifacts can be helpful in the diagnosis of alveolar-interstitial syndrome [33]. Therefore, LUS could be useful to distinguish between pneumonia and pleural effusion.

ve diagnosis requires a thoracentesis to allow biochemical and microbiological analyses [39]. In addition, the ultrasound feature of multiple comet-tail artifacts can be helpful in the diagnosis of alveolar-interstitial syndrome [33]. Therefore, LUS could be useful to distinguish between pneumonia and pleural effusion. Difficult weaning from mechanical ventilation is associated with increased patient morbidity and mortality [15]. Various factors including respiratory load, cardiac load, neuromuscular abnormalities, and metabolic disorders could make it difficult to wean from mechanical ventilation [15]. LUS can be used to find diaphragmatic dysfunction caused by critical illness or neuromuscular abnormalities [16,40,41] and helps diagnose pneumonia or pulmonary edema related to respiratory load [12]. In our study, LUS was performed for detecting causes of difficult weaning in two patients. Since LUS can be used to assess lung parenchyma, pleural space, and diaphragm, its use may be helpful to detect causes of difficult weaning. In surgical ICUs, the detection of pulmonary complications, which are major causes of morbidity and mortality, is an important challenge [1]. In our study, LUS was useful for diagnosing pulmonary complications, such as pneumonia, pulmonary edema, and diaphragm dysfunction. Especially, all patients who had visible abnormal findings during LUS were not diagnosed with pulmonary complications, suggesting that LUS could be used to exclude pulmonary complications.

In our study, LUS was useful for diagnosing pulmonary complications, such as pneumonia, pulmonary edema, and diaphragm dysfunction. Especially, all patients who had visible abnormal findings during LUS were not diagnosed with pulmonary complications, suggesting that LUS could be used to exclude pulmonary complications. There are several limitations to the present study. First, this study was conducted as a retrospective design in a single center ICU. Patients admitted to surgical ICU could vary widely by institution. Second, there are operator- dependent limitations to LUS. LUS was performed by two LUS experts. However, to reduce operator-dependent limitations, all patients’ ultrasonographic findings were reviewed by a supervisor. Third, since color Doppler in lung consolidation was not performed, we might not distinguish between dependent atelectasis and consolidation of infectious nature using LUS in several mechanically ventilated patients. The vascular pattern indicators within the consolidation, as assessed by color Doppler, could determine the etiology of pulmonary consolidations [24]. However, clinical findings including patient history, physical examination, and laboratory analysis help distinguish between atelectasis and pneumonia in the present study. Fourth, many patients who underwent trauma and thoracic surgery were enrolled in this study. Lung contusion associated with severe trauma (Injury Severity score ≥16 and Glasgow Coma scale <8) is difficult to distinguish from pneumonia via LUS [42]. However, we did not use LUS as a tool for distinguishing lung contusion from pneumonia in this study because most patients with severe trauma had undergone chest computed tomography during ICU admission periods. Performing LUS in patients who underwent thoracic surgery could be negatively affected by various factors such as anatomical change due to manipulation, surgical dressings, and chest tube. Many patients were diagnosed with pulmonary complications on the opposite side of their thoracic surgery in this study. Nonetheless, LUS in patients with thoracic surgery may have many limitations.

negatively affected by various factors such as anatomical change due to manipulation, surgical dressings, and chest tube. Many patients were diagnosed with pulmonary complications on the opposite side of their thoracic surgery in this study. Nonetheless, LUS in patients with thoracic surgery may have many limitations. In conclusion, LUS is used to evaluate clinical needs including hypoxemia, fever, abnormal chest radiographic findings, and difficult weaning. Furthermore, it is useful for diagnosis of pulmonary complications, such as pneumonia, pulmonary edema, and diaphragm dysfunction. Therefore, it is a useful method for evaluating surgically ill patients who might have pulmonary complications. No potential conflict of interest relevant to this article was reported. Figure 1. Flow chart for study enrollment. LUS: lung ultrasound; BLUE: bedside LUS in emergency; ICU: intensive care unit. Figure 2. Overall flow diagram outlining the diagnoses of pulmonary complications. LUS: lung ultrasound; ARDS: acute respiratory distress syndrome. Figure 3. Comparison of a chest radiograph and lung ultrasound image in a representative case. (A) Chest anteroposterior view. Active lesions were not visible. (B) Lung ultrasound of the posterior chest wall. Hypoechoic consolidation with an air bronchogram (arrow) was noted. Table 1. Definition of diagnosis for LUS in this study

Figure 3. Comparison of a chest radiograph and lung ultrasound image in a representative case. (A) Chest anteroposterior view. Active lesions were not visible. (B) Lung ultrasound of the posterior chest wall. Hypoechoic consolidation with an air bronchogram (arrow) was noted. Table 1. Definition of diagnosis for LUS in this study Variable Right lung Left lung Both lung PLAPS Upper BLUE point Lower BLUE point Upper BLUE point Lower BLUE point Pneumonia B’ or A/B or C profile B’ or A/B or C profile B’ or A/B or C profile B’ or A/B or C profile + Pulmonary edema B-profile B-profile B-profile B-profile ± Diffuse interstitial pattern with anterior consolidation B-profile B-profile B-profile B-profile ± C profile at least one point of four BLUE points LUS: lung ultrasound; BLUE: bedside LUS in emergency; PLAPS: posterolateral alveolar and/or pleural syndrome. Table 2. Baseline characteristics of study patients (n = 67) Characteristic Value Age (yr) 68 (55–78) Male sex 47 (70.1) Body mass index (kg/m2) 23.9 (21.9–28.4) No. of examined LUS per person 1 (1–2) Comorbidity Coronary artery disease 14 (20.9) Atrial fibrillation 3 (4.5) Chronic obstructive lung disease 9 (13.4) Rib fracture 7 (10.4) Previous lung surgery 2 (3.0) Chronic kidney disease 3 (4.5) Underlying malignancy 3 (4.5) Department Neurosurgery 33 (49.3) Thoracic surgery 15 (22.4) Orthopedic Surgery 10 (15.0) Abdominal Surgery 6 (9.0) Obstetric surgery 2 (3.0) Head and neck surgery 1 (1.5) Trauma 23 (34.3) Surgical procedure 49 (73.1) Values are presented as mean (interquartile range) or number (%). LUS: lung ultrasound.

Characteristic Value Age (yr) 68 (55–78) Male sex 47 (70.1) Body mass index (kg/m2) 23.9 (21.9–28.4) No. of examined LUS per person 1 (1–2) Comorbidity Coronary artery disease 14 (20.9) Atrial fibrillation 3 (4.5) Chronic obstructive lung disease 9 (13.4) Rib fracture 7 (10.4) Previous lung surgery 2 (3.0) Chronic kidney disease 3 (4.5) Underlying malignancy 3 (4.5) Department Neurosurgery 33 (49.3) Thoracic surgery 15 (22.4) Orthopedic Surgery 10 (15.0) Abdominal Surgery 6 (9.0) Obstetric surgery 2 (3.0) Head and neck surgery 1 (1.5) Trauma 23 (34.3) Surgical procedure 49 (73.1) Values are presented as mean (interquartile range) or number (%). LUS: lung ultrasound. Table 3. Findings of LUS in 53 patients with pulmonary edema, pneumonia, and diffuse interstitial pattern with anterior consolidation LUS diagnosis Findings of LUS (no. of patients) Pulmonary edema (n = 27) B-profile with bilateral effusion (21) B-profile without bilateral effusion (6) Pneumonia (n = 20) A/B-profile and C-profile with unilateral effusion (8) A/B-profile and C-profile without unilateral effusion (8) A/B-profile with unilateral effusion (3) C-profile with unilateral effusion (1) Diffuse interstitial pattern with anterior consolidation (n = 6) B-profile and C-profile with bilateral effusion (6) LUS: lung ultrasound. Table 4. Final diagnosis according to abnormalities in chest radiography

LUS diagnosis Findings of LUS (no. of patients) Pulmonary edema (n = 27) B-profile with bilateral effusion (21) B-profile without bilateral effusion (6) Pneumonia (n = 20) A/B-profile and C-profile with unilateral effusion (8) A/B-profile and C-profile without unilateral effusion (8) A/B-profile with unilateral effusion (3) C-profile with unilateral effusion (1) Diffuse interstitial pattern with anterior consolidation (n = 6) B-profile and C-profile with bilateral effusion (6) LUS: lung ultrasound. Table 4. Final diagnosis according to abnormalities in chest radiography Chest radiographic findings Value Unilateral air space opacification (n = 22) Pneumonia 15 (68.2) Atelectasis 2 (9.1) Pulmonary edema 1 (4.5) Hemopneumothorax 1 (4.5) No evidence of pulmonary complications 3 (13.6) Bilateral air space opacification (n = 33) Pulmonary edema 26 (78.8) Pulmonary edema and pneumonia 2 (6.1) Pneumonia 3 (9.1) Atelectasis 1 (3.0) ARDS 1 (3.0) Within normal limit (n = 12) No evidence of pulmonary complications 9 Pulmonary edema 1 Atelectasis 1 Diaphragm dysfunction 1 Values are presented as number (%) or number. ARDS: acute respiratory distress syndrome. Table 5. The comparison between final diagnosis and LUS diagnosis

Chest radiographic findings Value Unilateral air space opacification (n = 22) Pneumonia 15 (68.2) Atelectasis 2 (9.1) Pulmonary edema 1 (4.5) Hemopneumothorax 1 (4.5) No evidence of pulmonary complications 3 (13.6) Bilateral air space opacification (n = 33) Pulmonary edema 26 (78.8) Pulmonary edema and pneumonia 2 (6.1) Pneumonia 3 (9.1) Atelectasis 1 (3.0) ARDS 1 (3.0) Within normal limit (n = 12) No evidence of pulmonary complications 9 Pulmonary edema 1 Atelectasis 1 Diaphragm dysfunction 1 Values are presented as number (%) or number. ARDS: acute respiratory distress syndrome. Table 5. The comparison between final diagnosis and LUS diagnosis Diagnosis for LUS Final diagnosis (no. of patients) Pulmonary edema (n = 27) Pulmonary edema (27) Pneumonia (n = 20) Pneumonia (16), atelectasis (4) Diffuse interstitial pattern with anterior consolidation (n = 6) Pneumonia (2), pulmonary edema and pneumonia (2), pulmonary edema (1), ARDS (1) Pneumothorax with effusion (n = 1) Hemopneumothorax (1) Diaphragm dysfunction (n = 1) Diaphragm dysfunction (1) Nonvisible abnormal findings (n=12) No evidence of pulmonary complications (12) LUS: lung ultrasound; ARDS: acute respiratory distress syndrome.

Introduction Mechanical ventilation (MV) can be distinguished in two major categories: controlled MV, where the act of breathing is entirely controlled by the ventilator and assisted MV, where the patients’ respiratory system and the ventilator work together. Controlled ventilation is mainly applied for short periods, especially during the acute phase of illness. As soon as there are no contraindications to sedation reduction and the patient is able to breathe spontaneously, an assisted ventilator mode is applied. This practice has been dictated by a plethora of evidence emphasizing the detrimental effects of controlled MV and associated respiratory muscles inactivation, most notably the development of ventilator-induced diaphragmatic dysfunction [1,2]. Contrariwise, assisted ventilation is associated with reduced dose and duration of sedation and decreased used of neuromuscular blockade which have been shown to promote hemodynamic stability, diminish the risk of critical illness polyneuromyopathy, improve gas exchange and shorten the duration of MV [3-6].

tion [1,2]. Contrariwise, assisted ventilation is associated with reduced dose and duration of sedation and decreased used of neuromuscular blockade which have been shown to promote hemodynamic stability, diminish the risk of critical illness polyneuromyopathy, improve gas exchange and shorten the duration of MV [3-6]. With the onset of assisted MV, two entirely different systems are called to collaborate: the respiratory system and the ventilator. The quality of their interaction, temporal and quantitative, is imperative for MV to be beneficial and is described by the Greek word “synchrony.” Patient-ventilator synchrony occurs when (1) the ventilator provides flow and pressure as soon as patient effort begins; (2) the magnitude of this pressure and flow meets patient respiratory demand; and (3) the ventilator assistance is terminated when patient effort ends. Whenever any of the above is not fulfilled, patient-ventilator dyssynchrony occurs. Adverse effects of patient-ventilator dyssynchrony include: increased work of breathing, patient discomfort, alveolar overdistention and lung injury, sleep disturbances, periodic breathing, unnecessary use of sedation and excessive unloading of the diaphragm leading to ventilator-induced diaphragmatic dysfunction [7-10]. To detect patient-ventilator dyssynchronies, the physician should assess patient comfort and carefully inspect the pressure- and flow-time waveforms, available on the ventilator screen of all modern ventilators.

ion and excessive unloading of the diaphragm leading to ventilator-induced diaphragmatic dysfunction [7-10]. To detect patient-ventilator dyssynchronies, the physician should assess patient comfort and carefully inspect the pressure- and flow-time waveforms, available on the ventilator screen of all modern ventilators. This review aims to describe the various forms of patient-ventilator dyssynchrony, focusing on how to identify them at the bedside, to analyze their causes, the actions required to reduce or eliminate them and their clinical impact. Before that, a brief revision on the pathophysiology of breathing and the basic principles dictating the operation of assisted ventilatory modes will be provided. Pathophysiology To understand patient-ventilator dyssynchrony during assisted MV, one must first understand the pressures developed in the respiratory system during a spontaneous breath. To initiate a breath, the respiratory muscles contract, obeying nerve stimuli coming from the respiratory control center. Inspiratory muscle contraction expands the alveoli and decreases alveolar pressure below atmospheric pressure, driving gas into the lungs. Respiratory muscle contraction generates a pressure (Pmus), which at any time during inspiration (t), is dissipated to overcome two pressures opposing respiratory system inflation, the resistive (Pres) and elastic (Pel) pressure of the respiratory system (inertia is assumed to be negligible) [11]. This is accurately described by the equation of motion of the respiratory system:

which at any time during inspiration (t), is dissipated to overcome two pressures opposing respiratory system inflation, the resistive (Pres) and elastic (Pel) pressure of the respiratory system (inertia is assumed to be negligible) [11]. This is accurately described by the equation of motion of the respiratory system: (1) Pmus(t)=Pres(t)+Pel(t)=Rrs×V’(t)+Ers×∆VFRC(t) Where Rrs and Ers are resistance and elastance of the respiratory system respectively, ΔVFRC(t) is instantaneous volume above the passive functional residual capacity (FRC) and V’(t) is instantaneous flow. If ΔV is related to end-expiratory lung volume (EE), (2) Pmus(t)=Pres(t)+Pel(t)=Rrs×V’(t)+Ers×∆VEE(t)+PEEPi where PEEPi, is the elastic recoil pressure of the respiratory system at end-expiration. During control MV, the respiratory muscles do not contract, Pmus is 0 and the pressure needed to overcome the elastance and resistance of the respiratory system is entirely provided by the ventilator. The equation of motion in this case is modified as follows: (3) Paw=Pres(t)+Pel(t)=Rrs×V’(t)+Ers×∆VEE(t)+PEEPi where Paw is the pressure provided by the ventilator. In assisted MV, the total pressure applied to the respiratory systems comes both from the ventilator and the respiratory muscles and the equation of motion is modified as follows: (4) PTOT=Paw+Pmus=Pres(t)+Pel(t)=Rrs×V’(t)+Ers×∆VEE(t)+PEEPi where PTOT is the total pressure. It is clear that the variables of the above equation are tightly interdependent and each change of one may affect the others. This tight interrelation is schematically described in Figure 1 [11].

In assisted MV, the total pressure applied to the respiratory systems comes both from the ventilator and the respiratory muscles and the equation of motion is modified as follows: (4) PTOT=Paw+Pmus=Pres(t)+Pel(t)=Rrs×V’(t)+Ers×∆VEE(t)+PEEPi where PTOT is the total pressure. It is clear that the variables of the above equation are tightly interdependent and each change of one may affect the others. This tight interrelation is schematically described in Figure 1 [11]. Basic Principles of Assisted Ventilation To interpret the basic waveforms and subsequently to recognize patient-ventilator dyssynchronies, it is useful to discuss the basic principles dictating the operation of assisted ventilator modes. Three variables determine the function of a positive pressure ventilator: (1) the triggering variable, (2) the variable that controls the delivered pressure or flow during the mechanical inspiration, and (3) the cycling off variable.

Basic Principles of Assisted Ventilation To interpret the basic waveforms and subsequently to recognize patient-ventilator dyssynchronies, it is useful to discuss the basic principles dictating the operation of assisted ventilator modes. Three variables determine the function of a positive pressure ventilator: (1) the triggering variable, (2) the variable that controls the delivered pressure or flow during the mechanical inspiration, and (3) the cycling off variable. The triggering variable is the signal that initiates the mechanical breath. In assisted ventilation, the most commonly used triggering variables are flow and pressure. Mechanical inspiration starts when patient inspiratory effort decreases either the flow (flow triggering) or the pressure (pressure triggering) in the ventilator circuit to a preset level. It is generally believed that flow triggering is more friendly to the patient regarding the work of breathing, because the pressure triggering necessitates isometric contraction of the respiratory muscles [12,13]. However, studies have shown that, in the modern ventilators, there are minimal differences, if any, concerning the work of breathing between the two ways of triggering [12-14]. Other triggering variables include flow waveform, volume, transdiaphragmatic pressure (Pdi)-driven servoventilation and the electromyographic activity of the diaphragm (EAdi; neurally adjusted ventilator assist [NAVA]) but their use in clinical practice is limited [12,15-21].

between the two ways of triggering [12-14]. Other triggering variables include flow waveform, volume, transdiaphragmatic pressure (Pdi)-driven servoventilation and the electromyographic activity of the diaphragm (EAdi; neurally adjusted ventilator assist [NAVA]) but their use in clinical practice is limited [12,15-21]. The variable that controls the pressure and flow delivery distinguishes the various assisted MV modes [22,23]. In assist volume control, the ventilator delivers a preset tidal volume with a preset flow-time profile. These are the independent variables, while the airway pressure (Paw) needed to deliver the preset flow and volume depends on the mechanical properties of the respiratory system and is the dependent variable. In pressure control or pressure support, the ventilator delivers a preset pressure (independent variable). The dependent variables on pressure-preset modes are the volume and flow. These variables change according to the mechanical properties of the respiratory system and the pressure delivered. In proportional modes of ventilation, no variable is preset: the ventilator delivers support which is proportional to patient inspiratory effort. The later is expressed either through changes in instantaneous flow and volume (proportional assist ventilation [PAV]), or through changes in the neural activity of the diaphragm (NAVA) [11,22,24].

f ventilation, no variable is preset: the ventilator delivers support which is proportional to patient inspiratory effort. The later is expressed either through changes in instantaneous flow and volume (proportional assist ventilation [PAV]), or through changes in the neural activity of the diaphragm (NAVA) [11,22,24]. The cycling off variable is defined as the signal of terminating the delivery phase. The usual cycling off criterion in assisted pressure-preset modes of MV is airway pressure or flow: the ventilator terminates assist and opens the expiratory valve when Paw increases or flow decreases to a preset criterion. Other cycling off criteria are time, volume, the flow-waveform method and the EAdi in NAVA [15]. Ideally, the cycling off should occur simultaneously with the end of neural inspiration. However, this fact rarely, if ever, happens and the expiratory asynchrony is a common event with all conventional modes of assisted ventilation.

ing off criteria are time, volume, the flow-waveform method and the EAdi in NAVA [15]. Ideally, the cycling off should occur simultaneously with the end of neural inspiration. However, this fact rarely, if ever, happens and the expiratory asynchrony is a common event with all conventional modes of assisted ventilation. With pressure-preset modes rise time is a setting that determines how fast the ventilator will reach the selected Paw. It is available in the new generation ventilators during pressure support and pressure control ventilation. Fast rising time is associated with a sharp increase in inspiratory flow and may cause pressure overshoot and promote dyspnoea. Very slow rising time may be recognized by a rounded shaped inspiratory flow [25-27]. Both the low and high rise time may alter the mechanical inspiratory time and the volume delivered and may provoke dissociation between mechanical and neural inspiration [27,28]. Since, there are no rules for setting an optimal rise time, both very rapid and slow rise time should be avoided and rising time should be adjusted according to patient respiratory drive [25-28].

al inspiratory time and the volume delivered and may provoke dissociation between mechanical and neural inspiration [27,28]. Since, there are no rules for setting an optimal rise time, both very rapid and slow rise time should be avoided and rising time should be adjusted according to patient respiratory drive [25-28]. Patient-Ventilator Dyssynchronies Patient-ventilator dyssynchronies can be distinguished in two major categories: (1) dyssynchronies that occur because neural breath is not in phase with mechanical breath. This group includes: triggering delay, ineffective efforts, autotriggering, reverse triggering, delayed opening of the expiratory valve, premature opening of the expiratory valve, double triggering, and breath stacking and (2) dyssynchronies related to a discrepancy between the level of assist that the patient needs and the actual assist that the ventilator provides. 1) Delay of triggering and ineffective efforts Triggering delay is the time interval between the initiation of the neural and mechanical inspiration [11,18]. If esophageal pressure (Pes) or EAdi monitoring is available, it is observed as the time elapsed between the reduction in Pes or increase in EAdi (start of neural inspiration) and the abrupt increase of flow or Paw (start of mechanical inspiration).

tween the initiation of the neural and mechanical inspiration [11,18]. If esophageal pressure (Pes) or EAdi monitoring is available, it is observed as the time elapsed between the reduction in Pes or increase in EAdi (start of neural inspiration) and the abrupt increase of flow or Paw (start of mechanical inspiration). Ineffective efforts are patient’s efforts that fail to trigger the ventilator. Although usually described as a triggeringassociated dyssynchrony, ineffective efforts can happen at any time during the mechanical breath, during mechanical inspiration, expiration or at the transition between these two phases. The gold standard for their recognition is the simultaneous observation of the patient inspiratory activity. This is achieved either through recording of Pes or EAdi (Figure 2). Both methods are invasive and require the insertion of specified catheters: esophageal catheter, for Pes and NAVA catheter, for EAdi. In most cases, ineffective efforts can be identified non-invasively at the bedside, by observing the flow-time and Paw-time waveforms at the ventilator screen. Ineffective efforts cause distortions in the Paw and, more obviously, in the flow curve: an abrupt increase in inspiratory flow and a decrease in expiratory flow indicate the presence of ineffective efforts during mechanical inspiration and expiration, respectively (Figures 2 and 3) [11,18].

s at the ventilator screen. Ineffective efforts cause distortions in the Paw and, more obviously, in the flow curve: an abrupt increase in inspiratory flow and a decrease in expiratory flow indicate the presence of ineffective efforts during mechanical inspiration and expiration, respectively (Figures 2 and 3) [11,18]. Triggering delay and ineffective efforts share common pathophysiological mechanisms. Their causes can be classified into two main categories: ventilator settings and patients’ characteristics [8,29-31]. Ventilator settings that predispose to triggering delay and ineffective efforts are the high assist level, the delayed opening of the expiratory valve and the low triggering sensitivity. These factors are determined by the physician. The main patients’ characteristics include low respiratory drive (i.e., sedation, central nervous diseases, etc.), weak inspiratory muscles (i.e., critical illness polyneuromyopathy, myasthenia, etc.), and high resistance and compliance that increase the time constant of the respiratory system.

rmined by the physician. The main patients’ characteristics include low respiratory drive (i.e., sedation, central nervous diseases, etc.), weak inspiratory muscles (i.e., critical illness polyneuromyopathy, myasthenia, etc.), and high resistance and compliance that increase the time constant of the respiratory system. The incidence and the magnitude of the delay of triggering and ineffective efforts are not minor in clinical practice. Ineffective efforts are the commonest form of patient-ventilator dyssynchrony. Vaporidi et al. [32] evaluated the role of ineffective efforts, specifically clusters of them, during MV on the outcome of critically ill patients. Events of ineffective efforts were identified in 38% of patients with prolonged MV and these events were associated with prolonged MV and increased mortality. Their incidence is higher in patients with obstructive lung disease [29,33]. In these patients, the low elastic recoil pressure and/or increased resistance increase the time required for the patient to exhale (high time constant), predisposing to air trapping and dynamic hyperinflation. Dynamic hyperinflation is the most important cause of triggering delay and ineffective efforts. In the presence of dynamic hyperinflation, an elastic threshold load (PEEPi) is imposed on the inspiratory muscles at the beginning of inspiration. The inspiratory muscles must first counterbalance PEEPi in order to decrease alveolar pressure below external positive end-expiratory pressure (PEEPe) and trigger the ventilator [34-36]. This creates a delay between the beginning of inspiratory effort and ventilator triggering. In severe cases of dynamic hyperinflation, especially if the effort is weak, the patient may fail to trigger the ventilator (ineffective effort). High airway resistance can be recognized in all ventilator modes by observing the expiratory flow-time curve. Early in expiration, a spike in expiratory flow signifies the dynamic compression of central airways [11]. Thereafter, the expiratory flow decreases but very slowly, if ever, returns to zero line before the following inspiration (Figures 2 and 3).

zed in all ventilator modes by observing the expiratory flow-time curve. Early in expiration, a spike in expiratory flow signifies the dynamic compression of central airways [11]. Thereafter, the expiratory flow decreases but very slowly, if ever, returns to zero line before the following inspiration (Figures 2 and 3). Delayed or missed triggers represent eccentric contractions of the diaphragm which may be deleterious for its function. They may cause distress and in their presence, the respiratory rate on the ventilator screen does not reflect the true respiratory rate of the patient. Strategies in order to reduce the triggering delay and the number of ineffective efforts should be investigated [8,11,22,29,31]. Firstly, ventilator settings should be revised. Setting the most sensitive triggering that does not cause autotriggering is an option. With regard to triggering sensitivity, it has been shown that the flow-waveform method [15], compared to flow triggering, enhances the sensitivity and thus reduces the incidence of these events. Furthermore, the newer methods for triggering in the context of Pdidriven servoventilation [20] and NAVA [21] eliminate these events. Ventilator settings that decrease dynamic hyperinflation and, hence, triggering delay and ineffective efforts are: (1) reduction of minute ventilation by lowering assist (decrease set pressure, set tidal volume) and respiratory rate decrease, (2) lengthening expiratory time through a higher flow threshold for cycling off and a faster rising time in pressure support or a higher inspiratory flow and a shorter plateau inspiratory pressure in assist volume control mode and (3) application of PEEPe. The PEEPe narrows the difference between the alveolar pressure and the threshold of pressure for the initiation of the mechanical inspiration, hence helping the patient to trigger the ventilator. This is especially helpful in some patients with severe obstructive lung disease [29,33]. Secondly, patient-related causes should be reviewed. Attention should be given in respiratory drive and the factors that may reduce it, such as excessive sedation and alkalemia. Efforts in order to decrease the magnitude of dynamic hyperinflation also should be done, in instance the use of corticosteroids, bronchodilating therapy and aspiration of secretions may result in reduced expiratory resistance.

piratory drive and the factors that may reduce it, such as excessive sedation and alkalemia. Efforts in order to decrease the magnitude of dynamic hyperinflation also should be done, in instance the use of corticosteroids, bronchodilating therapy and aspiration of secretions may result in reduced expiratory resistance. 2) Autotriggering Autotriggering occurs when the ventilator is triggered in the absence of patient effort [37,38]. This phenomenon can also be recognized by inspection of the ventilator waveforms. The absence of decrease in Pes (if it is monitored) or in Paw before the delivery phase, especially in the presence of a zero flow long enough before the mechanical breath, are signs that the breath is not triggered by the patient. The flow-time profile of the autotriggered breaths is, often, different from the corresponding of patient- triggered breaths (Figure 4). If secretions or cardiac oscillations are the cause, one might notice the associate flow distortion and suspect the phenomenon.

signs that the breath is not triggered by the patient. The flow-time profile of the autotriggered breaths is, often, different from the corresponding of patient- triggered breaths (Figure 4). If secretions or cardiac oscillations are the cause, one might notice the associate flow distortion and suspect the phenomenon. Autotriggering may be caused by a low threshold for triggering and/or artifacts that may cause a drop of Paw or flow, which is misinterpreted by the ventilator as patient effort. Artifacts that frequently cause autotriggering are circuit leaks, presence of water in the circuit, hiccups and strong cardiogenic oscillators [37,38]. Factors that predispose a patient to increased risk of autotriggering are the low respiratory drive and breathing frequency, the low time constant of the respiratory system (increased elastic recoil, low resistance) and the presence of hyperdynamic circulation (larger cardiac output and higher ventricular filling pressures). Autotriggering is a common phenomenon during assisted modes of MV (invasive and non-invasive) [18,39]. When apparent, the respiratory rate is falsely elevated. These additional breaths may lead to hyperventilation, respiratory alkalosis, hyperinflation, and diaphragmatic dysfunction [2,40].

Autotriggering may be caused by a low threshold for triggering and/or artifacts that may cause a drop of Paw or flow, which is misinterpreted by the ventilator as patient effort. Artifacts that frequently cause autotriggering are circuit leaks, presence of water in the circuit, hiccups and strong cardiogenic oscillators [37,38]. Factors that predispose a patient to increased risk of autotriggering are the low respiratory drive and breathing frequency, the low time constant of the respiratory system (increased elastic recoil, low resistance) and the presence of hyperdynamic circulation (larger cardiac output and higher ventricular filling pressures). Autotriggering is a common phenomenon during assisted modes of MV (invasive and non-invasive) [18,39]. When apparent, the respiratory rate is falsely elevated. These additional breaths may lead to hyperventilation, respiratory alkalosis, hyperinflation, and diaphragmatic dysfunction [2,40]. To eliminate autotriggering, the physician must try to correct the underlying cause: enhance the respiratory drive by reducing sedation, correct alkalosis by reducing the level of assist, aspirate secretions, and minimize circuit leaks. Moreover, a higher triggering threshold, or change from flow to pressure triggering may abolish autotriggered breaths [37,38].

must try to correct the underlying cause: enhance the respiratory drive by reducing sedation, correct alkalosis by reducing the level of assist, aspirate secretions, and minimize circuit leaks. Moreover, a higher triggering threshold, or change from flow to pressure triggering may abolish autotriggered breaths [37,38]. 3) Premature cycling off When the exhalation valve of the ventilator opens too early, mechanical inspiration lasts shorter than neural inspiration. This form of dyssynchrony is often associated with insufficient support. Inspiratory muscle effort after the end of ventilator insufflation enhances the work of breathing and puts the patient at risk for double triggering. This asynchrony can present with the following ways [11]. (1) Zero or small inspiratory flow for some time after opening of the exhalation valve (Paw decreases to zero or PEEP level) indicates that inspiratory muscles continue to contract after the end of mechanical inspiration [41]. (2) A sharp decrease from the peak expiratory flow which lasts few milliseconds followed by an increase and then decreases gradually to zero toward the end of expiration: this pattern in the flow-time waveform is also a sign that considerable inspiratory muscle activity is present after opening of exhalation valve. When mechanical inspiration ends, end-inspiratory elastic recoil pressure is greater than inspiratory muscle pressure, creating positive alveolar pressure and, therefore, expiratory flow. Elastic recoil pressure decreases while lung volume declines due to inspiratory muscle contraction. An increasing opposing pressure to expiratory flow develops, causing an abrupt decrease in expiratory flow. This decrease is interrupted by the relaxation of the inspiratory muscles and expiratory flow increases and, therefore, follows the route as determined by the elastic recoil pressure and resistance of the patient and expiratory circuit [11]. (3) Double (or multiple) triggering: this refers to the delivery of two (or even more) ventilator insufflations during one single inspiratory effort. The inspiratory muscle contraction is great enough to overwhelm the elastic recoil pressure of the respiratory system and trigger the ventilator multiple times.

uit [11]. (3) Double (or multiple) triggering: this refers to the delivery of two (or even more) ventilator insufflations during one single inspiratory effort. The inspiratory muscle contraction is great enough to overwhelm the elastic recoil pressure of the respiratory system and trigger the ventilator multiple times. Premature cycling is associated with particular ventilator settings, such as low level of assist, relatively high threshold for cycling off and short inflation time. Short time constant of the respiratory system and long neural inspiration time are predisposing factors for this dyssynchrony. Furthermore, risk factors for multiple triggering are the low elastic recoil at the end of mechanical inflation and the intense inspiratory muscle activity [11,22]. As already mentioned, premature opening of the exhalation valve is related with increased work of breathing. Moreover, double triggering may provoke high tidal volumes (up to twice of the predetermined value) putting the patient at risk of ventilation-induced lung injury and ventilation-induced diaphragmatic dysfunction.

As already mentioned, premature opening of the exhalation valve is related with increased work of breathing. Moreover, double triggering may provoke high tidal volumes (up to twice of the predetermined value) putting the patient at risk of ventilation-induced lung injury and ventilation-induced diaphragmatic dysfunction. Premature opening of the expiratory valve and related multiple triggering can be minimized either by increasing the mechanical inspiration time and/or, if possible, by decreasing the neural inspiration time. The latter should be considered in patients with prolonged inspiratory efforts due to excessive administration of opioids. Actions that increase the mechanical inflation time include the lower flow threshold for cycling, the higher support and the slower rising time with pressure support mode [11,28]. With control modes, higher duration of mechanical inspiration should be easily achieved by proper adjustment of ventilator settings, such as inspiratory time, inspiratory flow or application of end-inspiratory pause [11,22].

for cycling, the higher support and the slower rising time with pressure support mode [11,28]. With control modes, higher duration of mechanical inspiration should be easily achieved by proper adjustment of ventilator settings, such as inspiratory time, inspiratory flow or application of end-inspiratory pause [11,22]. 4) Delayed cycling off When the exhalation valve opens too late, mechanical inspiration extends into neural expiration. This may promote rigorous expiratory muscle activity and increase the work of breathing as the patient struggles to terminate inspiration. Furthermore, it may unnecessarily increase the delivered tidal volume and shorten the available expiratory time. These effects may promote dynamic hyperinflation and associated dyssynchronies (delayed triggering, ineffective efforts), especially in patients with obstructive lung disease or provoke lung overdistention. Recognition of the delayed opening of the exhalation valve is, often, challenging. During pressure support mode, it can be indicated by the fast decrease of the inspiratory flow followed by an exponential decline towards the end of mechanical inspiration [11]. An abrupt increase (spike) of the Paw near the end of mechanical inspiration is either a sign of inspiratory muscle relaxation or expiratory muscle contraction [42]. Whichever is the cause, this spike is an evidence of delayed cycling off (Figure 5). With assist-volume control ventilation, the sharp increase of Paw towards the end of mechanical breath indicates that mechanical inspiration is longer than neural.

of inspiratory muscle relaxation or expiratory muscle contraction [42]. Whichever is the cause, this spike is an evidence of delayed cycling off (Figure 5). With assist-volume control ventilation, the sharp increase of Paw towards the end of mechanical breath indicates that mechanical inspiration is longer than neural. Low flow threshold for cycling off, high support and low rise time are predisposing factors for this asynchrony during pressure support ventilation [43]. With assisted volume control mode, high tidal volume, increased inflation time, low inspiratory flow and application of end-inspiratory pause may causes delayed opening of the exhalation valve [43]. With pressure support mode, delayed opening of the expiratory valve is commonly observed in patients with long time constant of respiratory system, such as in patients with chronic obstructive lung diseases [44]. Independent of the mode of the ventilation, changes of the ventilator settings should be combined by measures that minimize the airway resistance and dynamic hyperinflation (i.e., bronchodilation, steroid therapy, aspiration of secretions). 5) Reverse triggering Reverse triggering was first described by Akoumianaki et al. [45] who observed, in a group of heavily sedated

Independent of the mode of the ventilation, changes of the ventilator settings should be combined by measures that minimize the airway resistance and dynamic hyperinflation (i.e., bronchodilation, steroid therapy, aspiration of secretions). 5) Reverse triggering Reverse triggering was first described by Akoumianaki et al. [45] who observed, in a group of heavily sedated patients, the occurrence of patient inspiratory efforts that were triggered by the ventilator. This phenomenon, known as respiratory entrainment, has been previously described in animals, healthy subjects and preterm infants [40,46,47]. It differs from the other types of dyssynchronies in that the patient completely loses its normal breathing variability and actually breaths like a machine, exhibiting a fixed temporal relationship (1:1 or, less commonly, 1:2 or 1:3) between the onset of his inspiratory efforts and the onset of mechanical breaths. Vagal feedback and cortical influences are involved in the pathophysiological mechanism of these events [46,47]. Reverse triggered breaths may occur at any phase of the respiratory cycle and for variable periods. Their recognition, when Pes or EAdi recording is unavailable, is often difficult and requires careful inspection of the Paw- and flow-time waveforms (Figures 6 and 7). Depending on the phase of the mechanical breath that they occur, they may induce isometric or eccentric contractions of the respiratory muscles, augment the inflated tidal volume in pressure targeted modes or even trigger second breaths (breath stacking) causing hyperventilation and increasing the risk for lung overdistention (Figure 7) .

se of the mechanical breath that they occur, they may induce isometric or eccentric contractions of the respiratory muscles, augment the inflated tidal volume in pressure targeted modes or even trigger second breaths (breath stacking) causing hyperventilation and increasing the risk for lung overdistention (Figure 7) . 6) Dyssynchronies related to a discrepancy between the needs of the patient and the ventilator assist Inappropriate assist level (inadequate or excessive assist) with regard to patients’ ventilator demands is another cause of patient-ventilator dyssynchrony [11,22]. Inadequate assist level is usually observed in patients with high demands and increased respiratory drive (i.e., sepsis and metabolic acidosis). This discrepancy is combined with increased work of breathing and often the clinical status of the patient is indicative (i.e., use of accessory respiratory muscles, high respiratory rate). Excessive assist level is often an outcome of low respiratory drive and/or inappropriate ventilator settings, but the adverse effects of this incompatibility are important. Dynamic hyperinflation, respiratory alkalosis, periodic breathing pattern due to lower apnoeic threshold, sleep disturbances, ventilator-induced diaphragmatic dysfunction and/or lung injury are potential consequences of high assist [44,48].

ilator settings, but the adverse effects of this incompatibility are important. Dynamic hyperinflation, respiratory alkalosis, periodic breathing pattern due to lower apnoeic threshold, sleep disturbances, ventilator-induced diaphragmatic dysfunction and/or lung injury are potential consequences of high assist [44,48]. By inspecting the waveforms of Paw or flow, the respiratory effort in relation to assist level can be estimated [11,22]. Paw decreases when the inspiratory muscles contract and increases when the expiratory muscles contract. Paw is more sensitive in assist-volume control ventilation since it is the dependent variable. Although in pressure-targeted modes Paw should remain relatively constant, its shape may alter when the respiratory muscles contract rigorously (Figure 8). Furthermore, change in the flow pattern beyond the typical declining pattern is a sign of muscle effort [49]. Rapid decrease in inspiratory flow to flow threshold for cycling off in a patient with relatively long time constant is a sign of expiratory muscle contraction and thus a sign of high assist (Figure 9) [50]. Delayed cycling off might also indicate excessive assist. On the other hand, rounded or constant inspiratory flow represents significant inspiratory effort and might be due to insufficient assist (Figure 6). Supportive of insufficient assist is also the appearance of signs of premature cycling off or double triggering.

Delayed cycling off might also indicate excessive assist. On the other hand, rounded or constant inspiratory flow represents significant inspiratory effort and might be due to insufficient assist (Figure 6). Supportive of insufficient assist is also the appearance of signs of premature cycling off or double triggering. New Ventilator Modes Apart from ventilator settings adjustments, the introduction of new modes of MV, the so-called proportional ventilation modes, was a major step towards a better patient-ventilator interaction. These modes share similar principles: they provide assistance in proportion to patient effort. Patient effort is expressed either as a change in instantaneous flow and volume (PAV) or as a change in the EAdi (NAVA). The clinician sets the “gain” to augment patient effort and pressure and flow delivery change breath by breath following patient ventilator demand. Triggering in PAV is similar to conventional assisted modes but in NAVA the EAdi triggers the ventilator. PAV measures also the respiratory mechanics through dedicated software. Several clinical studies have shown the PAV and NAVA greatly improve the synchrony between the patient and the ventilator. Specifically ineffective efforts, delayed or premature opening of the expiratory valve and excessive ventilator assist are greatly minimized [10,51-55]. The reason for improved patient-ventilator synchrony is the tight link between the patient’s effort and the ventilator support: the patient retains considerable control of his breathing pattern and of the tidal volume. Even at high levels of assist, negative feedbacks from the respiratory controller minimize the risk for excessive tidal volume, protecting from overdistention and overventilation [53,56].

ient’s effort and the ventilator support: the patient retains considerable control of his breathing pattern and of the tidal volume. Even at high levels of assist, negative feedbacks from the respiratory controller minimize the risk for excessive tidal volume, protecting from overdistention and overventilation [53,56]. Conclusions Assisted MV offers significant advantages over controlled MV but attention should be paid on the presence of patient-ventilator dyssynchrony. Patient-ventilator dyssynchrony is common; it may occur at any phase of the respiratory cycle and can be due to time discrepancies or ventilator assist/respiratory demand discrepancies between the patient and the ventilator. It is associated with numerous unwanted effects and, may adversely affect the outcome. The gold standard for their recognition is the documentation of patient effort. However, in several cases, careful inspection of the ventilator screen along with the clinical picture of the patient is sufficient for the physician to recognize patient-ventilator dyssynchronies. Modern ventilators offer several modifiable settings to improve patient-ventilator interaction. New proportional modes of ventilation are also very helpful in improving patient-ventilator interaction. No potential conflict of interest relevant to this article was reported.

Conclusions Assisted MV offers significant advantages over controlled MV but attention should be paid on the presence of patient-ventilator dyssynchrony. Patient-ventilator dyssynchrony is common; it may occur at any phase of the respiratory cycle and can be due to time discrepancies or ventilator assist/respiratory demand discrepancies between the patient and the ventilator. It is associated with numerous unwanted effects and, may adversely affect the outcome. The gold standard for their recognition is the documentation of patient effort. However, in several cases, careful inspection of the ventilator screen along with the clinical picture of the patient is sufficient for the physician to recognize patient-ventilator dyssynchronies. Modern ventilators offer several modifiable settings to improve patient-ventilator interaction. New proportional modes of ventilation are also very helpful in improving patient-ventilator interaction. No potential conflict of interest relevant to this article was reported. Figure 1. Schematic illustration of the respiratory system and the applied pressures. The respiratory system is presented as a balloon at passive functional residual capacity (FRC) (continuous line). Two dashed lines indicate volumes above (ΔV1) and below (ΔV2) FRC. In mechanically ventilated patients during inspiration ventilator pressure (Paw) and pressure developed by inspiratory muscles (PmusI) generate flow and the volume increases above passive FRC. The sum of these two pressures is dissipated to overcome elastic pressure (Pel), and resistive pressure (Pres). All these pressures have positive values in the equation of motion. Pressure developed by contraction of expiratory muscles (PmusE), elastic recoil pressure due to volume below passive FRC and resistive pressure due to V’E have negative values in the equation. V’I : inspiratory flow; Rrs: resistance of the respiratory system; V’E: expiratory flow; Ers: elastance of the respiratory system. Modified from Kondili et al. Br J Anaesth 2003;91:106-19, with permission of Oxford University Press [43].

assive FRC and resistive pressure due to V’E have negative values in the equation. V’I : inspiratory flow; Rrs: resistance of the respiratory system; V’E: expiratory flow; Ers: elastance of the respiratory system. Modified from Kondili et al. Br J Anaesth 2003;91:106-19, with permission of Oxford University Press [43]. Figure 2. Airway pressure (Paw), flow and esophageal pressure (Pes) time curves in a patient ventilated with pressure support ventilation. Observe that the second decrease in Pes, which represents inspiratory effort of the patient, is not followed by a mechanical breath. This is ineffective effort (IE) during expiration and is manifested by a slight decrease in Paw associated with a simultaneous decrease in expiratory flow (red arrows). Notice that the signal of flow distortion is much clearer than the corresponding Paw change. In every mechanical breath, there is a time lag between the start of neural inspiration (first dotted line) and the start of mechanical inspiration (second dotted line). This time lag is the triggering delay. Observe the spike early in expiratory flow (black arrows) after each breath that suggests high airway resistance and long-time constant causing incomplete exhalation (flow is not zero before the next breath). Dynamic hyperinflation causes triggering delay and, combined with a relatively weaker patient effort (second Pes deflection smaller than the others) leads to ineffective triggering.

reath that suggests high airway resistance and long-time constant causing incomplete exhalation (flow is not zero before the next breath). Dynamic hyperinflation causes triggering delay and, combined with a relatively weaker patient effort (second Pes deflection smaller than the others) leads to ineffective triggering. Figure 3. Flow and esophageal pressure (Pes) time curves in a patient ventilated with pressure support ventilation. The start of neural inspiration (dotted line) is indicated by a rapid decrease in Pes associated with a rapid decrease in expiratory flow (expiratory flow returns rapidly to zero line). The two subsequent patient efforts are not accompanied by a mechanical breath and represent ineffective efforts (IE, red arrows). Both can be identified by the associated flow distortion. The first IE takes place during mechanical inspiration and causes an increase in inspiratory flow waveform. The second IE happens during expiration and is manifested by a decrease in expiratory flow. The spike early in expiratory flow (black arrow) due to high airway resistance and the incomplete exhalation (flow is not zero before the next breath) are signs of dynamic hyperinflation.

n increase in inspiratory flow waveform. The second IE happens during expiration and is manifested by a decrease in expiratory flow. The spike early in expiratory flow (black arrow) due to high airway resistance and the incomplete exhalation (flow is not zero before the next breath) are signs of dynamic hyperinflation. Figure 4. Airway pressure (Paw), flow and transdiaphragmatic pressure (Pdi) time curves of a patient ventilated on pressure support ventilation are illustrated. As indicated by the absence of Pdi increase, there is no inspiratory effort before the second mechanical breath (autotriggered breath, see blue shaded area). We can observe that, in comparison to patient-triggered breaths, where a decrease in Paw is observed before the start of mechanical inflation (grey shaded areas), there is no distortion in the Paw- (no decrease in Paw) and flow-time curve in the autotriggered breath. Moreover, the shape of the inspiratory flow-time curve is different compared to that of patient-triggered breaths. Notice the absence of dynamic hyperinflation in this patient (expiratory flow returns to zero after each breath).

istortion in the Paw- (no decrease in Paw) and flow-time curve in the autotriggered breath. Moreover, the shape of the inspiratory flow-time curve is different compared to that of patient-triggered breaths. Notice the absence of dynamic hyperinflation in this patient (expiratory flow returns to zero after each breath). Figure 5. Delayed opening of the expiratory valve. Flow, airway pressure (Paw), gastric pressure (Pgas) and esophageal pressure (Pes) time waveforms in a patient ventilated with pressure support ventilation. There is a significant time delay (blue shaded area) between the end of neural inspiration, recognized by a rapid increase in Pes, and the end of mechanical inspiration, signified by the termination of inspiratory flow (inspiratory flow equals zero). Observe the rapid increase of Paw towards the end of mechanical inspiration, indicating inspiratory muscle relaxation. Figure 6. Reverse triggering in a patient ventilated with assist pressure control ventilation. There is an inspiratory effort of the patient (dotted lines), as evidenced by the rapid increase in electromyographic activity of the diaphragm (EAdi) after every mechanical inflation (1:1 relationship). The time interval between the initiation of mechanical and neural inspiration is fixed. Indirect evidence of patient inspiratory activity during mechanical inflation is the notch in Paw (grey shaded area). Paw: airway pressure; VT: tidal volume.

the diaphragm (EAdi) after every mechanical inflation (1:1 relationship). The time interval between the initiation of mechanical and neural inspiration is fixed. Indirect evidence of patient inspiratory activity during mechanical inflation is the notch in Paw (grey shaded area). Paw: airway pressure; VT: tidal volume. Figure 7. Reverse triggering in a patient ventilated with assist volume control ventilation. Esophageal pressure (Pes) decrease reveals patient inspiratory efforts (blue line) after every mechanical inflation in 1:1 relationship. Indirect evidence of patient inspiratory activity during mechanical inflation is the flow distortion (grey shaded area) and the disappearance (blue arrows) of plateau airway pressure (Paw) in the flow-time and Paw-time waveform, respectively. In this patient, a reverse triggered breath was strong enough to trigger the ventilator at the end of the mechanical inspiration, causing breath stacking (red shaded area). Inflated tidal volume (VT) during breath stacking increased from 444 ml to 800 ml (double arrow).

he flow-time and Paw-time waveform, respectively. In this patient, a reverse triggered breath was strong enough to trigger the ventilator at the end of the mechanical inspiration, causing breath stacking (red shaded area). Inflated tidal volume (VT) during breath stacking increased from 444 ml to 800 ml (double arrow). Figure 8. Flow, airway pressure (Paw), esophageal pressure (Pes) and transdiaphragmatic pressure (Pdi) time waveforms in a patient ventilated with pressure support ventilation. Observe the vigorous contraction of inspiratory muscles (Pdi increase) during the mechanical inspiration. The magnitude of this contraction causes a rounded inspiratory flow and a large decrease of Paw (gray shaded area) from the expected square-shaped form during inspiration. Rounded flow and Paw decrease are signs of low ventilator assist with respect to patients ventilator demands. Figure 9. High assist in a patient ventilated with pressure support ventilation. Observe the square shaped airway pressure (Paw) and the abrupt decrease in inspiratory flow to flow threshold for cycling off towards the end of inspiration (arrows). There is also a significant cycling off delay (blue shaded area), seen often at high assist levels. Esophageal pressure (Pes) and electromyographic activity of the diaphragm (EAdi) decrease rapidly but mechanical inflation continues. Importantly, expiratory muscles contract during the whole expiration.

We appreciate your interest in our paper and are thankful for taking the time to express your opinions. We would also like to thank you for the opportunity to clarify aspects of our methodology in relation to concerns on diagnostic criteria of the study patients and further express our opinions on the issue of elderly patients’ intensive care unit (ICU) care. As for the low proportion of sepsis, the article has a limitation. The Materials and Method section states that “The ICU patients with diagnoses associated with ICU mortality were classified into 10 subcategories. Diagnoses were sorted according to main 10th revision of the International Statistical Classification of Diseases (ICD-10) codes of the patients.” Due to large number of study patients (10,366), checking whether each patient had been diagnosed of sepsis according to strict clinical definitions was impossible, and it is possible that the patients in other disease categories would have been treated for sepsis; for example, sepsis originated from pneumonia, hepatobiliary infection, etc. For future studies, data collection by reviewing individual patient’s medical charts or well-designed prospective studies could overcome the limitation mentioned above.

e patients in other disease categories would have been treated for sepsis; for example, sepsis originated from pneumonia, hepatobiliary infection, etc. For future studies, data collection by reviewing individual patient’s medical charts or well-designed prospective studies could overcome the limitation mentioned above. The very elderly patients have high risk of death for critical illness, when compared to younger age groups [1]. The elderly patients have a higher prevalence of chronic illnesses and an age-related decrease of physical ability [2]. Age of the patients is a significant factor when deciding whether patients should undergo active or palliative treatment. The author has well pointed out that aggressive ICU care for very elderly patients could not be clinically beneficial and concurs heavy economic burdens for patients’ families. A study in Korea showed that for the elderly, the proportion of patients who had specified “do not resuscitate” is higher than younger age groups [3]. In recent publications from JAMA, Guidet et al. [4] report the results of ICE-CUB 2 study. Patients aged 75 years or more were randomized to usual care hospitals and intervention hospital group, in which the study patients were more actively admitted to ICUs when compared to the counterparts. Paradoxically, even though admission rate was two times higher, intervention group showed no clinical benefit and in-hospital mortality was even higher.

randomized to usual care hospitals and intervention hospital group, in which the study patients were more actively admitted to ICUs when compared to the counterparts. Paradoxically, even though admission rate was two times higher, intervention group showed no clinical benefit and in-hospital mortality was even higher. On the other hand, many studies also support ICU care for the elderly patients. A study from the Netherlands showed that both short-term and long-term risk-adjusted mortality decreased significantly in both very elderly ICU patients and patients aged less than 80 years during the period of 2008–2014 [5]. Another study in Korea showed that ICU and in-hospital mortalities were not significantly different for very elderly critically ill patients compared to the younger patients [6]. For these reasons, setting age of 80 years as a cutoff for receiving active ICU care could create other problems such as a considerable number of very elderly patients who could be recovered by active ICU care, missing opportunity of treatment. When deciding whether patients should undergo active ICU care, the age of patients should be considered in conjunction to other important factors such as wills of the patient and family to continue intensive treatment, reversibility of the disease and underlying comorbidities.

CU care, missing opportunity of treatment. When deciding whether patients should undergo active ICU care, the age of patients should be considered in conjunction to other important factors such as wills of the patient and family to continue intensive treatment, reversibility of the disease and underlying comorbidities. In conclusion, we agree that medical care for very elderly patients requires different clinical approaches compared to their younger counterparts. Before ICU care of very elderly patients, physicians should carefully consider various patient-related factors in the decision of aggressive versus palliative care, for optimal results. No potential conflict of interest relevant to this article was reported.

Dear Editor: Spontaneous echo contrast (SEC) is often observed in patients with mitral stenosis, atrial fibrillation, cardiomyopathy, or a ventricular aneurysm [1]. SEC is a smoke-like echo density observed on echocardiograms, and is caused by increased red blood cell aggregation during low-flow states. It is also a risk factor of thromboembolism [2]. SEC can be observed in patients with severe ventricular dysfunction receiving venoarterial extracorporeal membrane oxygenation (VA-ECMO). We present a case in which left ventricular-SEC (LV-SEC) was mistaken for a LV thrombus during VA-ECMO for severe LV dysfunction. A 36-year-old female patient diagnosed with acute fulminant myocarditis was provided VA-ECMO support on hospital day (HD) 1. Briefly, VA-ECMO (Rota-Flow; Maquet Inc., Hirrlingen, Germany) was implanted in the right femoral artery (15-French arterial cannula) and the left femoral vein (20-French venous cannula). Her height and body weight are 163 cm and 52 kg (body surface area, 1.53 m2). VA-ECMO was initiated with a circuit flow of 3.5 L/min (cardiac index, 2.3 2L/min/m2). Her creatine kinase-myocardial band and troponin-I levels at admission were 188.03 ng/ml (normal range, 0 to 5 ng/ml) and >50.0 ng/ml (normal range, 0 to 0.78 ng/ml), respectively. Impaired ventricular function (ejection fraction, 22%) suspected as acute fulminant myocarditis was detected by transthoracic echocardiography (TTE) at admission.

cardial band and troponin-I levels at admission were 188.03 ng/ml (normal range, 0 to 5 ng/ml) and >50.0 ng/ml (normal range, 0 to 0.78 ng/ml), respectively. Impaired ventricular function (ejection fraction, 22%) suspected as acute fulminant myocarditis was detected by transthoracic echocardiography (TTE) at admission. TTE revealed decreased LV function (ejection fraction, 10%) with mild mitral regurgitation (grade II) immediately after VA-ECMO. Opening of the aortic valve and arterial pulsatility were not observed. Pulmonary edema was aggravated on HD 4. Left atrial (LA) decompression was achieved using a LA catheter (20-French femoral venous cannula) by balloon atrial septostomy through the right femoral vein. Pulmonary edema and cardiomegaly improved after LA decompression (Figure 1) and cardiac enzymes levels were reduced (Figure 2). Input and output were controlled to improve pulmonary edema; about 5,000 ml volume was removed from HD 5 to 10. On HD 11, ECMO flow was abruptly reduced from 3.2 to 1.6 L/min, and hyperechogenic material was detected in the LV using a portable TTE (Figure 3A). In view of the abrupt ECMO flow reduction, we considered LV thrombus, but after infusing normal saline (500 ml) ECMO flow recovered and the hyperechogenic material disappeared (Figure 3B). We realized that LV-SEC was misdiagnosed as LV thrombus. At that time, her activated prothrombin time was 100 seconds, and fortunately there was no evidence of thromboembolic complications. Her cardiac function then recovered and on HD 14 VA-ECMO was weaned without embolic or bleeding complications.

peared (Figure 3B). We realized that LV-SEC was misdiagnosed as LV thrombus. At that time, her activated prothrombin time was 100 seconds, and fortunately there was no evidence of thromboembolic complications. Her cardiac function then recovered and on HD 14 VA-ECMO was weaned without embolic or bleeding complications. VA-ECMO administered using peripheral cannulation induces retrograde blood flow and increases afterload. If LV dysfunction is severe, the aortic valve may not open against a high afterload and pulsatility may disappear. In such cases, blood stasis and SEC are observed [3]. In the absence of pulsatility, TTE is important to ensure the presence of swirling. Excessive volume removal and LA venting can aggravate SEC because preload is lowered and the LV cavity is emptied [4]. Furthermore, if ECMO flow decreases abruptly, SEC may be mistaken for thrombus. In this context, cautious interpretation of echocardiographic finding based on the clinical situation is important to avoid invasive procedures, such as exploratory cardiac surgery. It is also important to reduce the afterload for improving the pulsatility. Combining ECMO with intra-aortic balloon pump may have benefit in term of decreasing the afterload [5]. Vasodilator medications, such as nitroprusside, nitroglycerin, calcium channel blockers can reduce the afterload as well. The flow of VA-ECMO should also be adjusted while confirming the perfusion status and mean arterial pressure [6].

intra-aortic balloon pump may have benefit in term of decreasing the afterload [5]. Vasodilator medications, such as nitroprusside, nitroglycerin, calcium channel blockers can reduce the afterload as well. The flow of VA-ECMO should also be adjusted while confirming the perfusion status and mean arterial pressure [6]. Myocarditis is a hypercoagulable state induced by a systemic inflammatory process [7], and thus adequate anticoagulation therapy is required during VA-ECMO to prevent thrombotic complications. In particular, when SEC is observed in a patient with severe LV dysfunction, the risk of these complications increases. In the described case, we performed VA-ECMO for acute fulminant myocarditis with severe LV dysfunction, and followed this with daily echocardiography to identify swirling or SEC. Target activated prothrombin time of heparin therapy was maintained at a high level (>70 seconds) because the aortic valve did not open and swirling was observed. Furthermore, if our patient had not been given inappropriate anticoagulation therapy when SEC was present, complications of thrombus would certainly have occurred.

activated prothrombin time of heparin therapy was maintained at a high level (>70 seconds) because the aortic valve did not open and swirling was observed. Furthermore, if our patient had not been given inappropriate anticoagulation therapy when SEC was present, complications of thrombus would certainly have occurred. In conclusion, VA-ECMO in patients with acute fulminant myocarditis and severe LV compromise provides meaningful life support and facilitates the recovery of cardiac function. However, SEC can occur in the absence of pulsatility or aortic valve opening, and severe volume restriction and LA venting might increase the risk of LV-SEC by emptied the LV cavity. LV-SEC may be mistaken for thrombus when ECMO flow abruptly reduced, and thus, repeat evaluation using TTE and cautious interpretation of echocardiographic finding are required. On the other hand, in patients with SEC, proper anticoagulation therapy is important to prevent complications of thrombus. No potential conflict of interest relevant to this article was reported. Figure 1. Chest radiographs obtained on hospital days 4 to 9. (A) 4, (B) 5, (C) 6, (D) 7, (E) 8, (F) 9 hospital days. Figure 2. Cardiac enzyme trends. Figure 3. Trans-thoracic echocardiographs obtained on hospital day 11. (A) Spontaneous echo contrast mistaken for left ventricular thrombus (parasternal short and long axis views). (B) Disappeared spontaneous echo contrast after volume replacement (parasternal short and long axis views).

Malignant syndrome (MS) in Parkinson disease (PD) is rare but potentially fatal. Takubo et al. [1] reported that MS in PD who did not recover to the pre-MS state was 33%. Clinical features of MS in PD include muscle rigidity, hyperpyrexia, and altered consciousness, which are similar to the clinical symptoms of neuroleptic MS [1-3]. Whereas neuroleptic MS has been reported in patients receiving neuroleptic drugs, MS in PD has been reported in patients following withdrawal of antiparkinsonian drugs [1,2]. If MS in PD is not caused by withdrawal of antiparkinsonian drugs, the typical symptoms of MS in PD, such as fever and autonomic instability, are difficult to distinguish from systemic inflammatory response syndrome (SIRS), resulting in delayed diagnosis and treatment. We present a case of an elderly woman who was initially misdiagnosed with SIRS instead of MS because of no history of withdrawal of antiparkinsonian drugs.

as fever and autonomic instability, are difficult to distinguish from systemic inflammatory response syndrome (SIRS), resulting in delayed diagnosis and treatment. We present a case of an elderly woman who was initially misdiagnosed with SIRS instead of MS because of no history of withdrawal of antiparkinsonian drugs. Case Report A 70-year-old woman with a 9-year history of PD was admitted to the emergency department (ED) with a 1-day history of altered consciousness, fever, and convulsive movements. She was already bedridden (Hoehn and Yahr scale 5) and had left hand tremor. Her medications included carbidopa/levodopa (25 mg/100 mg per oral [PO] three times a day) and memantine (10 mg PO twice a day). Her family member informed us that she had taken her antiparkinsonian medications regularly before admission to the ED and that she had recurrent diarrhea with poor oral intake 1 week before the ED visit. After the recurrent diarrhea 1 week before the ED visit, the left hand tremor was aggravated. At admission, her vital signs included a blood pressure of 100/60 mmHg, temperature of 39.0°C, respiratory rate of 24 breaths/min, and heart rate of 138 beats/min. A physical examination showed hot and wet skin, rigidity of extremities, and left hand tremor. The patient showed seizure-like movements accompanied by leftsided eye deviation and muscle contraction 30 minutes after admission to the ED. The results of an arterial blood gas analysis showed respiratory alkalosis (pH 7.51, PaCO2 22 mmHg, PaO2 148 mmHg, and HCO3- 18.3 mEq/L) via O2 face mask 10 L. The initial laboratory results showed a white blood cell count of 16.2 × 103/L, sodium at 146 mEq/L, blood urea nitrogen (BUN) at 34.4 mg/dl, creatinine at 0.6 mg/dl, creatinine kinase(CK) at 41 U/L, and C-reactive protein (CRP) < 0.5 mg/dl. The computed tomography (CT) scan of the chest and abdomen did not show any abnormal findings. The brain CT showed severe brain atrophy with communicating hydrocephalus, but the cerebrospinal fluid examination result was normal. An electroencephalogram showed no epileptiform abnormalities. There were no cultured bacteria in the patient’s blood or urine. Six hours after admission, the patient showed high fever (> 40°C). The blood pressure of this patient was 90/50 mm/Hg despite administering 1.5 L of 0.9% NaCl over 4 hours. At 12 hours after admission, blood pressure was decreased (systolic blood pressure of 40 mmHg) and her respiration was shallow.

t’s blood or urine. Six hours after admission, the patient showed high fever (> 40°C). The blood pressure of this patient was 90/50 mm/Hg despite administering 1.5 L of 0.9% NaCl over 4 hours. At 12 hours after admission, blood pressure was decreased (systolic blood pressure of 40 mmHg) and her respiration was shallow. Then, we decided intubation and mechanical ventilation support, and the patient was given a vasopressor (norepinephrine 0.6 μg/kg/h IV). Central venous catheterization was performed via the right subclavian vein and central venous pressure was 3 mmHg. To control the fever, dantrolene sodium (40 mg IV) and a cooling blanket were used; as a result, her temperature decreased to 36°C 15 hours after admission. At 14 hours after admission, the laboratory results indicated sodium at 149 mEq/L, BUN at 63.3 mg/dl, creatinine at 1.6 mg/dl, serum procalcitonin (PCT) at 25.63 ng/ml, and myoglobin at 970.8 ng/ml but showed no elevated CRP (Figure 1). Based on a presumptive diagnosis of severe pseudomembranous colitis due to a history of antibiotic use (ceftriaxone 2 g for 5 days) 1 month before and elevated PCT, oral vancomycin was given once. Urine output was decreased (< 0.4 ml/kg/h), and hence, continuous venovenous hemodiafiltration (CVVHDF) with Multifiltrate® (Fresenius Medical Care, Bad Homburg, Germany) for acute kidney injury and oligouria was applied at 20 hours after admission. At 43 hours after admission, the laboratory results showed elevated CK (274 U/L) and decreased PCT levels (9.79 ng/ml). Clostridium difficile and its toxins were not detected in the stool and results of polymerase chain reaction assays were negative. There was improvement in the renal function tests and oligouria, so CVVHDF was stopped at 56 hours after admission. Based on a diagnosis of MS in PD on the 4th hospital day (HD), the patient received antiparkinsonian drugs that had been taken previously via a Levin tube. The frequency of the left hand tremor decreased, and her consciousness recovered to its previous condition. Weaning and extubation were performed on the 12th HD. After conservative treatment, she was discharged on the 22nd HD after recovery to her previous condition before admission.

taken previously via a Levin tube. The frequency of the left hand tremor decreased, and her consciousness recovered to its previous condition. Weaning and extubation were performed on the 12th HD. After conservative treatment, she was discharged on the 22nd HD after recovery to her previous condition before admission. Discussion In our case, the patient was diagnosed with MS based on the modified diagnostic criteria of Levenson [1,3]. Our patient showed delayed mild elevation of CK, but this could be explained by focal subclinical rhabdomyolysis or due to an increase in the permeability of the muscle membrane [2]. Myoglobinemia may also serve as an evidence of rhabdomyolysis.

nosed with MS based on the modified diagnostic criteria of Levenson [1,3]. Our patient showed delayed mild elevation of CK, but this could be explained by focal subclinical rhabdomyolysis or due to an increase in the permeability of the muscle membrane [2]. Myoglobinemia may also serve as an evidence of rhabdomyolysis. The most common cause of MS in PD is withdrawal from antiparkinsonian drugs (29%), followed by infection (19%), and poor oral intake (13%) [1]. Additionally, sodium imbalance can cause MS in PD [4-6]. The dehydration caused by persistent diarrhea and hypernatremia may have been the cause of MS in our patient. Old age, high Hoehn and Yahr stage during the symptomatic phase of MS, high akinesia score, and an absence of withdrawal symptoms from medication prior to developing MS have been associated with poor outcomes [1]. In our case, the patient had a high Hoehn and Yahrscore, high akinesia score, and no withdrawal symptoms from dopaminergic medication. Furthermore, as it was difficult to distinguish between MS and SIRS, the diagnosis of MS in PD was delayed. Therefore, our patient was expected to have a poor outcome. However, she recovered to her previous condition, perhaps because her high fever was quickly corrected by dantrolene sodium and cooling blankets. One previous study found that the outcome of malignant hyperthermia depends on the time interval from the start of the reaction to the establishment of correct treatment, which includes terminating the trigger agent exposure and administering dantrolene [7]. Therefore, this intensive therapy reduced the duration of MS and helped the recovery of our patient.

me of malignant hyperthermia depends on the time interval from the start of the reaction to the establishment of correct treatment, which includes terminating the trigger agent exposure and administering dantrolene [7]. Therefore, this intensive therapy reduced the duration of MS and helped the recovery of our patient. This patient showed high serum PCT levels at an early stage, but at that time, we found no evidence of severe infection; serum PCT levels may also be elevated by non-infectious causes, such as inhalation injury, burn injury, mechanical trauma, extensive surgery, or heatstroke tissue injury [8]. Because PCT can be produced by tissue injury, we think that hyperpyrexia and rhabdomyolysis of MS in PD may cause damage to muscle tissue, resulting in elevated PCT. Lovas et al. [9] reported that a young male patient presenting with high fever and neurological impairment caused by amphetamine intoxication showed extremely high serum PCT levels without evidence of a bacterial infection. Because serum PCT decreases daily by approximately 50% due to its half-life, high serum PCT levels may not be associated with a bacterial infection. However, a follow-up blood culture showed methicillin-resistant Staphylococcus aureus, which was considered to be a catheter related infection for CVVHDF. In summary, emergency physicians should consider MS in patients who have a history of PD and who present with high fever without evidence of an infection. Additionally, physicians should know that PCT in patients with MS can be elevated without a concomitant bacterial infection.

This patient showed high serum PCT levels at an early stage, but at that time, we found no evidence of severe infection; serum PCT levels may also be elevated by non-infectious causes, such as inhalation injury, burn injury, mechanical trauma, extensive surgery, or heatstroke tissue injury [8]. Because PCT can be produced by tissue injury, we think that hyperpyrexia and rhabdomyolysis of MS in PD may cause damage to muscle tissue, resulting in elevated PCT. Lovas et al. [9] reported that a young male patient presenting with high fever and neurological impairment caused by amphetamine intoxication showed extremely high serum PCT levels without evidence of a bacterial infection. Because serum PCT decreases daily by approximately 50% due to its half-life, high serum PCT levels may not be associated with a bacterial infection. However, a follow-up blood culture showed methicillin-resistant Staphylococcus aureus, which was considered to be a catheter related infection for CVVHDF. In summary, emergency physicians should consider MS in patients who have a history of PD and who present with high fever without evidence of an infection. Additionally, physicians should know that PCT in patients with MS can be elevated without a concomitant bacterial infection. No potential conflict of interest relevant to this article was reported. Figure 1. Change in procalcitonin (PCT) and C-reactive protein (CRP) during the first 3 days. At admission, the level of PCT was extremely high (reference range, 0-0.5 ng/ml), while the level of CRP was within the normal range (reference range, 0.1-1 mg/dl).

Dear Editor: I read with great interest the article “Demographic Changes in Intensive Care Units in Korea over the Last Decade and Outcomes of Elderly Patients: A Single-Center Retrospective Study” published in the Korean Journal of Critical Care Medicine in May 2017 [1]. The results indicated that the proportion of inpatients aged 65–79 years admitted to an intensive care unit (ICU) increased from 47.9% in 2005 to 63.7% in 2014, and the proportion of ICU-hospitalized patients older than 80 years increased from 12.8% in 2005 to 20.7% in 2014. However, the overall mortality rate did not increase despite a higher mortality rate in the elderly than in the younger patients. These results are worthy and impressively demonstrate the recent changing trends in demographic data of ICU patients in Korea. However, I would like to comment on the following two points.

Figure 1. Patients included in the study analysis. ICU: intensive care unit; CAG: coronary angiography. Figure 2. Graph showing annual changes in proportions of two age groups (65–79 and ≥80 yr). P-value indicates statistical difference between two age groups (P-value <0.05). Table 1. Annual percentages of the four different age groups of intensive care unit patients Year <50 yr 50–64 yr 65–79 yr ≥80 yr 2005 248 (22.7) 320 (29.3) 383 (35.1) 140 (12.8) 2006 230 (19.9) 303 (26.2) 462 (40.0) 160 (13.9) 2007 223 (20.9) 297 (27.9) 407 (38.2) 138 (13.0) 2008 191 (18.6) 250 (24.4) 435 (42.4) 150 (14.6) 2009 173 (17.9) 241 (24.9) 397 (41.0) 158 (16.3) 2010 194 (17.4) 285 (25.5) 479 (42.8) 160 (14.3) 2011 165 (15.3) 293 (27.1) 451 (41.8) 171 (15.8) 2012 162 (15.1) 278 (25.9) 435 (40.6) 197 (18.4) 2013 111 (12.1) 277 (30.2) 382 (41.7) 146 (15.9) 2014 95 (10.9) 222 (25.4) 376 (43.0) 181 (20.7) P-value <0.0001a Value are presented as number (%). a Result of modified chi-square test for linear by linear association. Table 2. Patients characteristics by year

Until recently, the lung was considered “forbidden territory” for ultrasound. With lung ultrasound, however, the amount of lung consolidation and pleural effusion can be assessed semiquantitatively. Lung ultrasound consists of the identification of 10 signs, and there are several well-established protocols such as the BLUE (Bedside Lung Ultrasonography in Emergency) protocol for diagnosing acute respiratory failure and the FALLS (Fluid Administration Limited by Lung Sonography) protocol for managing acute circulatory failure. The BLUE protocol is a fast protocol that defines eight profiles, correlated with six diseases seen in 97% of patients admitted to the intensive care unit (ICU). With this protocol, it becomes possible to differentiate between pulmonary edema, pulmonary embolism, pneumonia, chronic obstructive pulmonary disease, asthma, and pneumothorax [1]. The FALLS protocol uses the potential of lung ultrasound for the early demonstration of fluid overload at an infra-clinical level [2]. It is used in patients with acute respiratory failure, allowing a sequential search for obstructive, cardiogenic, hypovolemic, and distributive shock using simple real-time echocardiography in combination with lung ultrasound, with the appearance of B lines considered to be the endpoint of fluid therapy. In addition, ultrasound can help to guide airway management in a patient with acute respiratory distress who needs to be intubated and mechanically ventilated (PINK protocol). In a patient with acute respiratory distress who is often ventilated and difficult to transport, computed tomography (CT) is not an easy option, and lung ultrasound can help to predict difficult airway and proper endotracheal tube size, or to confirm proper endotracheal tube placement with avoidance of desaturation during CT [3]. In addition, lung ultrasound can be used to determine the cause of fever distinguishing pneumonia from atelectasis [4], and to rule out pneumothorax, hypovolemia, pulmonary embolism and pericardial tamponade in cardiac arrest (SESAME protocol) [5].

dotracheal tube placement with avoidance of desaturation during CT [3]. In addition, lung ultrasound can be used to determine the cause of fever distinguishing pneumonia from atelectasis [4], and to rule out pneumothorax, hypovolemia, pulmonary embolism and pericardial tamponade in cardiac arrest (SESAME protocol) [5]. In the critical care setting, lung ultrasound is increasingly used, as it allows bedside visualization of the lungs. Critical care ultrasound is a combination of simple protocols, with lung ultrasound being a basic application, allowing the assessment of urgent diagnoses and therapeutic decisions. Although chest radiographs (CXR) and CT are mostly used for daily or prompt evaluation of lung in the ICU, there are significant drawbacks such as the huge radiation hazard, need for transportation, and risk of contrast use. On the other hand, lung ultrasound has advantages of absence of radiation, bedside availability, good reproducibility, and cost efficiency [6]. Lung ultrasound has more accuracy than bedside CXR and roughly the same accuracy as CT. Ultrasound is far superior for the detection of pneumothorax and pleural effusion compared with CXR and provides accurate quantitative data regarding the volume of pleural effusions, lung consolidations, and pneumothorax [7,8]. Although supine portable CXR is notoriously unreliable in the evaluation of pneumothorax, in the absence of tube thoracostomy or subcutaneous emphysema, ultrasound has sensitivity and specificity superior to CXR for the detection of pneumothorax [9]. Plain CXR is most sensitive for pleural effusion when the patient is in the upright or lateral decubitus position, but optimal positioning is difficult in the ICU. Ultrasound can detect pleural effusions with a sensitivity and specificity of 93% when CT is used as a gold standard [10].

he detection of pneumothorax [9]. Plain CXR is most sensitive for pleural effusion when the patient is in the upright or lateral decubitus position, but optimal positioning is difficult in the ICU. Ultrasound can detect pleural effusions with a sensitivity and specificity of 93% when CT is used as a gold standard [10]. In an article of the Korean Journal of Critical Care Medicine, Kang et al. [11] reported the usefulness of lung ultrasound to detect pulmonary complication including pulmonary edema and pneumonia especially in the surgical ICUs, where pulmonary complications are major causes of morbidity and mortality. While previous studies on lung ultrasound were mostly for use in medical ICUs, the authors demonstrated the usefulness of lung ultrasound in the surgical ICU. Their indications for lung ultrasound included hypoxemia, abnormal CXR without hypoxemia, fever, and difficult weaning. Lung ultrasound was helpful for diagnosis of pneumonia, atelectasis, pulmonary edema, or a combination of these diseases. In addition, lung ultrasound detected lung parenchymal consolidation with air bronchogram, pulmonary edema, and pneumonia even in patients without CXR abnormalities. In the surgically ill and injured patients, combined with venous, cardiac, and abdominal examination, ultrasound investigation of lung can provide an overview of cardiac performance and intravascular volume and practically guide management for hemodynamic optimization.

umonia even in patients without CXR abnormalities. In the surgically ill and injured patients, combined with venous, cardiac, and abdominal examination, ultrasound investigation of lung can provide an overview of cardiac performance and intravascular volume and practically guide management for hemodynamic optimization. Lung ultrasound can be extended from neonates to adults, and from medical to surgical and several other disciplines (anesthesiology, emergency medicine, etc.). The higher rate of detection of ultrasound, combined with its ease and increasing accessibility, makes for a powerful diagnosis in the ICU. Although lung ultrasound requires acquisition of an ultrasound machine and training of physicians, it allows a critical care provider to quickly respond to a majority of critical situations. Therefore, lung ultrasound could be a reasonable, fully operational, bedside gold standard in the ICU. No potential conflict of interest relevant to this article was reported.

. However, the overall mortality rate did not increase despite a higher mortality rate in the elderly than in the younger patients. These results are worthy and impressively demonstrate the recent changing trends in demographic data of ICU patients in Korea. However, I would like to comment on the following two points. First, the authors might have shown a slightly lower sepsis prevalence, not greater than 1% for each subgroup in the ICU. The study conducted by Oh et al. [2], using the Health Insurance Review & Assessment Service database, revealed that the inhospital mortality of patients with sepsis was as high as 38.9%, and the proportion of sepsis increased with age. In addition, in the United States, sepsis was ranked 11 in the top primary diagnoses in 1996 among patients older than 65 years who were admitted to the ICU, but in 2010, sepsis was ranked 1 among the primary diagnoses in older patients admitted to the ICU [3]. In Korea, where the proportion of elderly population is rapidly increasing, it is expected that the rate of primary diagnosis of sepsis in the elderly patients who are admitted to the ICU would be higher. It is considered to be a limitation due to single-center studies, and nationwide demographic studies of ICU patients are required.

the proportion of elderly population is rapidly increasing, it is expected that the rate of primary diagnosis of sepsis in the elderly patients who are admitted to the ICU would be higher. It is considered to be a limitation due to single-center studies, and nationwide demographic studies of ICU patients are required. Second, there had been questions regarding the appropriateness of ICU hospitalization of very old patients (VOPs) aged greater than 80 years with chronic illnesses. Roch et al. [4] reported that among patients older than 80 years, the ICU mortality rate was 46%, 1-year mortality rate was 72%, and 2-year mortality rate was 79%. The purpose of care for VOPs admitted to ICU might be just life-sustaining therapy rather than survival with highly qualified life. Increased long life-sustaining therapy might result in increased economic and emotional stress for patients, families, and medical costs. In addition, it will be difficult to properly operate in the ICU because of the prolonged length of stay of VOPs with chronic illnesses and the shortage of human resources for the other patients.

g life-sustaining therapy might result in increased economic and emotional stress for patients, families, and medical costs. In addition, it will be difficult to properly operate in the ICU because of the prolonged length of stay of VOPs with chronic illnesses and the shortage of human resources for the other patients. Nguyen et al. [5] suggested that the outcome of admission to the ICU of a VOP for postoperative care after an unplanned surgery or severe medical problems would be dismal, and recommended admission to a regular ward or acute care unit rather than ICU admission. However, there is a lack of legal backgrounds for doctors and VOPs with chronic illness to make advanced plans for the proper end of life care in Korea. There have been significant concerns about a new law on hospice and palliative care and withdrawal of life-sustaining therapy, which will be implemented in 2018 because it includes ambiguous or difficult provisions to be applied in medical reality [6]. Its improvement and supplement will be required. Additionally, advanced-care planning regarding ICU admission with patients or their families seems to improve the quality of life and satisfaction level and reduce related mental stress and depression. In these cases, we need to make an early and active palliative care consultation system after ICU admission to reduce life-sustaining therapy and decrease ICU length of stay [7,8]. No potential conflict of interest relevant to this article was reported.

Background 1) Introduction Status epilepticus (SE) and refractory SE (RSE) are some of the most complex and expensive conditions encountered in the neurological intensive care unit (ICU). For example, in the United States, SE represents 0.07% of hospital admissions [1], but accounts for a larger proportion of hospital costs, including $4 billion in direct inpatient costs annually [2] with a mean length of stay of 14 days [2,3]. In those who fail to respond to conventional treatments, costs increase exponentially in proportion to treatment intensity. RSE provides diagnostic, management, and ethical challenges as treatment options become limited and prolonged hospital stays are accompanied by several potential complications and worse patient outcomes. This review provides a brief overview of SE, and a recent update on the management of RSE and super-RSE (SRSE) based on evidence from the literature, practice-based guidelines, and experiences at Keck Medical Center of the University of Southern California, an academic tertiary care center.

and worse patient outcomes. This review provides a brief overview of SE, and a recent update on the management of RSE and super-RSE (SRSE) based on evidence from the literature, practice-based guidelines, and experiences at Keck Medical Center of the University of Southern California, an academic tertiary care center. 2) Definitions Early definitions of SE described seizures persisting “for a sufficient length of time or is repeated frequently enough to produce a fixed or enduring epileptic condition [4].” A more practical definition was introduced in 1993 as a “seizure lasting more than 30 minutes or occurrence of two or more seizures without recovery of consciousness in between [5].” This somewhat arbitrary timeframe was based on animal models suggesting irreversible neuronal injury with prolonged seizure [6]. Conceptually, SE describes “a condition resulting either from the failure of the mechanisms responsible for seizure termination or from the initiation of mechanisms which lead to abnormally prolonged seizures,” as proposed by the International League Against Epilepsy Task Force on Classification of Status Epilepticus [7]. A pragmatic definition of SE as a seizure “≥5 minutes or two or more discreet seizures between which there is incomplete recovery of consciousness” was proposed by Lowenstein and Alldredge [8], and has largely been adopted by clinicians and clinical researchers. Based on the pathophysiology of SE described later, this definition is most useful and is endorsed by the Neurocritical Care Society (NCCS) Status Epilepticus Guideline Writing Committee [9].

onsciousness” was proposed by Lowenstein and Alldredge [8], and has largely been adopted by clinicians and clinical researchers. Based on the pathophysiology of SE described later, this definition is most useful and is endorsed by the Neurocritical Care Society (NCCS) Status Epilepticus Guideline Writing Committee [9]. SE fails to respond to standard medications in 31% to 43% of cases [10,11]. Although there is no consensus definition of RSE, it typically describes SE refractory to early benzodiazepines and one additional first-line anti-seizure medication. If seizures cannot be terminated with the use of an intravenous (IV) anesthetic in addition to benzodiazepines and standard anticonvulsants, the condition is termed SRSE. Relatively uncommon, up to 15% of SE cases become super-refractory [12], accounting for 4% of seizure-related hospital discharges [13]. As RSE and SRSE are commonly treated with therapeutic coma, signs of seizure become clinically absent. When electrographic seizures are not accompanied by corresponding motor movements, or by subtle motor signs, the term nonconvulsive SE (NCSE) is used. New-onset RSE (NORSE) describes RSE in previously healthy individuals where no cause of SE is immediately apparent [14].

coma, signs of seizure become clinically absent. When electrographic seizures are not accompanied by corresponding motor movements, or by subtle motor signs, the term nonconvulsive SE (NCSE) is used. New-onset RSE (NORSE) describes RSE in previously healthy individuals where no cause of SE is immediately apparent [14]. Causes and Pathophysiology 1) Causes SE and RSE most often occur in patients with known epilepsy. Up to 34% of SE cases are attributed to low anticonvulsant drug levels [10,15], due to suboptimal dosing, medication non-compliance, or recent medication change. In those without epilepsy, etiology varies significantly by age. In adults, other leading causes of SE include stroke, toxic-metabolic encephalopathy, and hypoxic-ischemic injury, while tumor and meningoencephalitis account for only 4% to 5% of RSE cases [10]. In comparison, central nervous system (CNS) infection is the leading cause of SE in children, with trauma and anoxic injury as other major etiologies [16,17]. Regardless of age, seizure etiology is not found in a large portion of patients. An analysis of 130 cases of NORSE by Gaspard et al. [14] determined that despite exhaustive search, 52% remained cryptogenic. In such cases, patients are commonly treated for presumed infectious, autoimmune or paraneoplastic encephalitis; however, an infectious agent is rarely found [18]. Similarly, neuropathologic study with biopsy or autopsy is often inconclusive [19-22].

determined that despite exhaustive search, 52% remained cryptogenic. In such cases, patients are commonly treated for presumed infectious, autoimmune or paraneoplastic encephalitis; however, an infectious agent is rarely found [18]. Similarly, neuropathologic study with biopsy or autopsy is often inconclusive [19-22]. 2) Pathophysiology (1) Biochemical phases of SE Seizures result from disruptions in the normal balance of excitatory and inhibitory processes. Hypersynchronous neuronal firing, the physiologic hallmark of seizure, is mediated predominantly by glutamate excitation and voltage-gated sodium and calcium channels. Seizure termination is largely dependent on the inhibitory effects of gamma-aminobutyric acid (GABA)-receptor activation and voltage-gated potassium channels [23]. After initiation, failure of intrinsic mechanisms or disruption of inhibition from extrinsic factors will lead to prolonged seizure activity. The maintenance phase of SE is characterized by synaptic internalization of GABA receptors and expression of excitatory N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors, resulting in self-sustaining seizures refractory to conventional antiepileptic drugs (AEDs) [24]. In animal models, the potency of benzodiazepines decrease 20-fold within 30 minutes of SE, while phenytoin loses effectiveness to a lesser degree [25]. Observations in humans suggest that a refractory state is reached much faster. In a video electroencephalogram (EEG) study of 120 generalized convulsive seizures, the mean seizure duration was 62 seconds with no events lasting >2 minutes [26]. In clinical practice, seizures are less likely to terminate spontaneously after this period, leading to Lowenstein and Alldredge [8] proposing SE be defined as continuous seizure for ≥5 minutes. In the minutes-to-hours that follow, SE is maintained by increased expression of proconvulsive neuropeptides and decrease in inhibitory ones [27-30].

izures are less likely to terminate spontaneously after this period, leading to Lowenstein and Alldredge [8] proposing SE be defined as continuous seizure for ≥5 minutes. In the minutes-to-hours that follow, SE is maintained by increased expression of proconvulsive neuropeptides and decrease in inhibitory ones [27-30]. Less is known about the biochemical changes associated with RSE and SRSE. In the range of hours-to-weeks, long-term changes in gene expression may occur secondary to recurrent seizure activity and neuronal injury, resulting in neural reorganization [24]. (2) Neuronal injury and death Neuronal loss in SE was described in early foundational work in primate models by Meldrum and Horton [6] and Meldrum et al. [31]. Cell death occurs even in the absence of hypoxia, acidosis, hypoglycemia and other confounding factors [32-34], and may be mediated by ‘programmed necrosis’ and apoptosis [35,36]. In humans, serum neuron-specific enolase, a biomarker for neural injury, has been shown to be elevated in SE [37,38], while autopsy studies demonstrate decreased hippocampal neuron density postmortem [39].

lycemia and other confounding factors [32-34], and may be mediated by ‘programmed necrosis’ and apoptosis [35,36]. In humans, serum neuron-specific enolase, a biomarker for neural injury, has been shown to be elevated in SE [37,38], while autopsy studies demonstrate decreased hippocampal neuron density postmortem [39]. To what extent similar pathologic changes occur in NCSE as compared to convulsive SE (CSE) is unclear. Early observational studies by Engel et al. [40] suggested long-standing neurocognitive changes following NCSE. Serum neuron-specific enolase has also been shown to be elevated [37,41,42]; however, it is difficult to determine whether NCSE directly causes neuronal injury, or if seizures reflect brain damage from other causes. NCSE also represents a heterogenous syndrome, with potential for neuronal injury dependent on the seizure type and etiology. Absence SE and NCSE in patients with underlying chronic epilepsy may be less prone to injury [43,44], while complex partial SE is most associated with pathologic changes [40-42].

om other causes. NCSE also represents a heterogenous syndrome, with potential for neuronal injury dependent on the seizure type and etiology. Absence SE and NCSE in patients with underlying chronic epilepsy may be less prone to injury [43,44], while complex partial SE is most associated with pathologic changes [40-42]. (3) Physiologic phases of SE Patients with CSE progress through physiologic stages [45]. In the early compensated phase, convulsions are accompanied by significant sympathetic activation. During this stage, hypertension, increased cardiac output, and increased cerebral blood flow is seen, and serum markers of hypermetabolism such as lactic acid and glucose will be elevated [46]. After prolonged seizure activity (>30 minutes), pathophysiologic decompensation occurs. This is characterized by cerebral dysautoregulation, cardiovascular dysfunction, and signs of systemic metabolic crisis: hypoxia, hypoglycemia, and acidosis. Failure to prevent profound physiologic disturbances may exacerbate secondary brain injury associated with RSE.

minutes), pathophysiologic decompensation occurs. This is characterized by cerebral dysautoregulation, cardiovascular dysfunction, and signs of systemic metabolic crisis: hypoxia, hypoglycemia, and acidosis. Failure to prevent profound physiologic disturbances may exacerbate secondary brain injury associated with RSE. Diagnosis 1) Clinical findings Differentiating seizure from mimics by clinical exam alone can be difficult. In hospitalized patients with encephalopathy out of proportion to known laboratory and imaging findings, non-convulsive seizures can be detected by continuous EEG (cEEG) in up to 18% [47]. Nonspecific motor movements, including posturing, rigidity, shivering, tremor, myoclonus, and spontaneous gaze deviation, can be misinterpreted as seizure. Facial twitching is associated with a higher incidence of electrographic seizures compared to other subtle signs [48]. 2) The use of EEG Because of the limitations of clinical exam in diagnosing seizures, EEG confirmation is necessary in many cases. Suspicion for SE should be higher in the critically ill, where seizure complicates the ICU stay of 8% to 10% of patients [49,50]. In the neuro-ICU, the incidence of seizure can exceed 30%, with NCSE occurring in up to 13% [51,52]. Eight percent of those with coma without obvious seizure activity have NCSE on cEEG [53].

Suspicion for SE should be higher in the critically ill, where seizure complicates the ICU stay of 8% to 10% of patients [49,50]. In the neuro-ICU, the incidence of seizure can exceed 30%, with NCSE occurring in up to 13% [51,52]. Eight percent of those with coma without obvious seizure activity have NCSE on cEEG [53]. In response, the use of cEEG has expanded. Indications for extended monitoring and the optimal study duration is unknown; however, early EEG findings can predict final diagnosis. The majority of seizures occur within the first 30 minutes of recording [54]. In this timeframe, generalized slowing is poorly predictive of seizure while lateralized periodic discharges has the strongest association [55]. In the absence of epileptiform discharges within the first 4 hours of EEG, seizures are unlikely to be found [56]. Conventional Management of RSE Effective management of SE and RSE occurs in three phases: seizure termination, prevention of SE recurrence, and minimization of complications. This is a dynamic process, where complications occur during all clinical stages and treatment plans should be adapted for seizure recurrence and new diagnostic findings. An initial treatment strategy simultaneously addresses airway, breathing and circulation while intervening quickly to abort seizure activity within 5 minutes of onset [9].

cess, where complications occur during all clinical stages and treatment plans should be adapted for seizure recurrence and new diagnostic findings. An initial treatment strategy simultaneously addresses airway, breathing and circulation while intervening quickly to abort seizure activity within 5 minutes of onset [9]. 1) Seizure termination in SE (1) Stage I treatment: benzodiazepine trial As described above, seizures are associated with profound pathophysiologic changes, that if left untreated, can lead to severe clinical consequences. When convulsive seizures occur, attention should be given to both basic life support measures while considering first line AED therapy. Evidence-based guidelines from both the NCCS and the American Epilepsy Society (AES) recommend benzodiazepines as initial therapy of choice [9,57]. Lorazepam is widely available, fast to administer, and terminates overt SE in 65% of cases [58]. Compared to diazepam, IV lorazepam is pharmacologically preferred because it is less lipid soluble and undergoes slower peripheral distribution [59]. Intramuscular midazolam is effective when IV access has not been established. Recurrent convulsive seizures in RSE and SRSE are treated similarly. In intubated patients, full-dose lorazepam (0.1 mg/kg) or equivalent benzodiazepine can be given, with smaller doses considered when attempting to avoid intubation or when significant hypotension is present.

1) Seizure termination in SE (1) Stage I treatment: benzodiazepine trial As described above, seizures are associated with profound pathophysiologic changes, that if left untreated, can lead to severe clinical consequences. When convulsive seizures occur, attention should be given to both basic life support measures while considering first line AED therapy. Evidence-based guidelines from both the NCCS and the American Epilepsy Society (AES) recommend benzodiazepines as initial therapy of choice [9,57]. Lorazepam is widely available, fast to administer, and terminates overt SE in 65% of cases [58]. Compared to diazepam, IV lorazepam is pharmacologically preferred because it is less lipid soluble and undergoes slower peripheral distribution [59]. Intramuscular midazolam is effective when IV access has not been established. Recurrent convulsive seizures in RSE and SRSE are treated similarly. In intubated patients, full-dose lorazepam (0.1 mg/kg) or equivalent benzodiazepine can be given, with smaller doses considered when attempting to avoid intubation or when significant hypotension is present. (2) Stage II treatment: first-line conventional AEDs Benzodiazepines lose effectiveness in established SE and are suboptimal for long-term anti-seizure management; therefore, next-line AEDs should be ordered and administered early, within 10 minutes of seizure onset [9]. First-line conventional AEDs are selected for their broad spectrum of activity and their ability to be given safely as an IV loading dose in attempts to abort SE and reach therapeutic levels rapidly (Table 1).

t; therefore, next-line AEDs should be ordered and administered early, within 10 minutes of seizure onset [9]. First-line conventional AEDs are selected for their broad spectrum of activity and their ability to be given safely as an IV loading dose in attempts to abort SE and reach therapeutic levels rapidly (Table 1). Historically, phenytoin has been the initial stage II AED given after benzodiazepines, terminating overt seizure in 44% in the Veterans Affairs Cooperative Study [58]. Many clinical concerns may limit its use. Phenytoin is not considered sedating; however, hypotension and cardiac arrhythmia can occur with bolus doses [60]. Other acute complications include ‘purple glove syndrome’ and metabolic acidosis, while long-term side effects including teratogenicity and coarsening of facial features make it nonideal for chronic therapy. Phenytoin is also limited by its narrow therapeutic range. Additionally, bioavailability is influenced by factors common to ICU patients: drug-drug interactions, renal insufficiency, and hypoalbuminemia.

g-term side effects including teratogenicity and coarsening of facial features make it nonideal for chronic therapy. Phenytoin is also limited by its narrow therapeutic range. Additionally, bioavailability is influenced by factors common to ICU patients: drug-drug interactions, renal insufficiency, and hypoalbuminemia. For many neurointensivists, levetiracetam has been explored as a stage I and II AED in SE. Preferred pharmacologic properties include its high bioavailability and few drug-drug interactions owing to low plasma protein binding and minimal hepatic metabolism [61]. Unfortunately, data for levetiracetam as an emergent treatment is limited. In a pilot study involving traditional AEDs, levetiracetam demonstrated equal efficacy in aborting overt seizure compared to lorazepam [62], and in a recent meta-analysis, showed similar activity in benzodiazepine-resistant SE (68.5% relative effectiveness) compared to phenytoin (50.2%), phenobarbital (73.6%), and valproic acid (75.7%) [63]. The drug is well tolerated, ideal in critically ill patients on multiple medications, and for long-term therapy in those who remain at risk for seizures after hospital discharge.

zodiazepine-resistant SE (68.5% relative effectiveness) compared to phenytoin (50.2%), phenobarbital (73.6%), and valproic acid (75.7%) [63]. The drug is well tolerated, ideal in critically ill patients on multiple medications, and for long-term therapy in those who remain at risk for seizures after hospital discharge. (3) Stage III treatment: termination of RSE seizures Optimal therapy after failure of benzodiazepines and first stage II AED is unknown. Although a second conventional AED is typically added in this setting, the likelihood of success is marginal and may delay seizure termination. Development of RSE should prompt planning for ICU admission, intubation, and initiation of a general anesthetic agent. Seizure activity is definitively aborted with use of a single anesthetic agent, or with combination of agents in SRSE, with no method or agent proven superior to another [12,64]. Anesthetic drugs used are outlined in Table 2. Whether propofol or midazolam are used first largely depends on provider preference. Propofol, a GABA agonist and NMDA-receptor antagonist, is often preferred because its highly lipophilic properties with large volume of distribution allow for rapid offset to facilitate neurologic examination. In patients with hypotension with or without propofol, midazolam is an alternate; however, it itself is associated with hypotension in 40% of patients and may exhibit either tachyphylaxis or drug accumulation resulting in prolonged sedation [65,66]. Adequate EEG suppression may be difficult to achieve with a single anesthetic agent.

th hypotension with or without propofol, midazolam is an alternate; however, it itself is associated with hypotension in 40% of patients and may exhibit either tachyphylaxis or drug accumulation resulting in prolonged sedation [65,66]. Adequate EEG suppression may be difficult to achieve with a single anesthetic agent. Because of its side effect profile, pentobarbital is typically reserved for cases of SRSE refractory to combined propofol and midazolam. Pentobarbital infusion can result in profound hypotension, cardiorespiratory depression, metabolic acidosis, and immunosuppression [67,68], which may outweigh any potential benefit of the drug. In a systematic review of outcomes in RSE and SRSE, barbiturate infusion achieved seizure control in 64%, but was associated with prolonged mechanical ventilation and increased mortality [12]. Whether poorer outcomes are due to the therapy and its complications, or a result of SRSE being associated with more severe illness is unclear.

review of outcomes in RSE and SRSE, barbiturate infusion achieved seizure control in 64%, but was associated with prolonged mechanical ventilation and increased mortality [12]. Whether poorer outcomes are due to the therapy and its complications, or a result of SRSE being associated with more severe illness is unclear. 2) Goals of EEG suppression Continuous EEG is fundamental to RSE management, and by AES guidelines, is recommended initially for all patients requiring anesthetics [57]. Because prolonged seizure is associated with electromechanical dissociation, where electrographic seizure no longer manifests overt clinical signs, successful termination is determined by EEG suppression. In SRSE, ‘burst suppression’ is typically maintained for an extended period; however, there is little evidence for efficacy of this practice, including the optimal level and duration of suppression. In a retrospective review by Krishnamurthy and Drislane [69] of 35 patients treated with pentobarbital infusion, “slow,” “burst-suppression” and “flat” EEG patterns were not strongly associated with the rate of seizure relapse. Similarly, a retrospective study by Rossetti et al. [70] suggested that outcome after RSE was independent of the extent of EEG suppression achieved. Meta-analysis of 193 RSE patients support these conclusions: despite EEG background suppression being associated with fewer breakthrough seizures compared to seizure suppression alone (4% and 53%, respectively), the study found no difference in withdrawal seizures after AED weaning, or in mortality [64].

sion achieved. Meta-analysis of 193 RSE patients support these conclusions: despite EEG background suppression being associated with fewer breakthrough seizures compared to seizure suppression alone (4% and 53%, respectively), the study found no difference in withdrawal seizures after AED weaning, or in mortality [64]. In the absence of high quality evidence, it is reasonable to titrate anesthetic agents to termination of ictal activity. If accomplished by reaching a diffusely slow EEG pattern, further suppression may not be needed. In the presence of persistent seizure or continuous epileptiform discharges, it is reasonable to consider suppression to one burst of EEG activity (<2 seconds duration) every 10–20 seconds, and to continue that depth for 12–48 hours before initiating an anesthetic wean. An Update on Management Strategies for SRSE After seizure termination, general anesthetics do not ensure freedom from relapse. Up to 15% of SE cases become SRSE [12], typically diagnosed with recurrence of ictal activity when anesthetic drugs are titrated down. With little high-quality data to guide therapy, and high complication rates, the management of SRSE can be challenging. The remainder of this review will provide a recent update on practice recommendations, based on expert opinion, limited data, and experiences at our institution. With the lack of compelling evidence, the NCCS offers only weak recommendations for SRSE management, while it is the beyond the scope of most recent AES guidelines [9,57].

of this review will provide a recent update on practice recommendations, based on expert opinion, limited data, and experiences at our institution. With the lack of compelling evidence, the NCCS offers only weak recommendations for SRSE management, while it is the beyond the scope of most recent AES guidelines [9,57]. 1) Optimizing scheduled antiepileptic drugs Appropriate management of multiple AEDs requires basic knowledge of drug mechanisms and pharmacokinetics. Drug properties to consider when choosing an agent include its formulation, speed of CNS penetration, protein binding, volume of distribution, presence of autoinduction, drug-drug interactions, half-life, and elimination kinetics [71]. (1) Adequate therapeutic dosing There are several reasons why AEDs fail to prevent seizure recurrence in SRSE. Although an epileptogenic focus may be refractory to certain medications, subtherapeutic drug levels commonly contribute to their ineffectiveness. AED levels should be checked serially throughout the early phase of SRSE, with titration of AEDs as appropriate.

reasons why AEDs fail to prevent seizure recurrence in SRSE. Although an epileptogenic focus may be refractory to certain medications, subtherapeutic drug levels commonly contribute to their ineffectiveness. AED levels should be checked serially throughout the early phase of SRSE, with titration of AEDs as appropriate. Loading doses are given at drug initiation to reach therapeutic range early. AEDs with linear pharmacokinetics and low protein-binding can be loaded with greater predictability than those without such properties. Phenytoin, in contrast to levetiracetam, demonstrates non-linear kinetics and is highly protein-bound, making therapeutic ranges difficult to achieve and maintain. Patient weight influences volume of distribution; therefore, weight-based dosing should be given when established. Similarly, general anesthetics should be bolused on initiation. Propofol is given as a 2–5 mg/kg bolus, midazolam as a 0.2 mg/kg bolus, and pentobarbital with 5–10 mg/kg.

difficult to achieve and maintain. Patient weight influences volume of distribution; therefore, weight-based dosing should be given when established. Similarly, general anesthetics should be bolused on initiation. Propofol is given as a 2–5 mg/kg bolus, midazolam as a 0.2 mg/kg bolus, and pentobarbital with 5–10 mg/kg. Drug bioavailability should be considered when choosing IV versus oral (PO) routes of administration. Although bioavailability of many first-line AEDs such as phenytoin, valproate, and levetiracetam is high in healthy subjects [71,72], alterations in gastric motility, gut absorption and drug metabolism in the critically ill likely interfere [73]. In a study of IV versus PO phenytoin loading, IV doses resulted in faster times to therapeutic drug concentrations (0.21 ± 0.28 hours) compared to PO (5.63 ± 0.28 hours) [74]. At our institution, IV formulations are typically continued until seizure termination and relapse has not occurred, and additionally during significant gastrointestinal illness. Failure to account for the half-life of an AED can result in sub-therapeutic drug troughs. Clinically, this can be observed as EEG suppression after drug dosing, but recurrent seizure or increased frequency of epileptiform discharges prior to next administration. In highly-tolerated medications with shorter half-lives, such as levetiracetam and valproate, dosing every 8 hours instead of twice daily should be considered. In cases of polytherapy, AEDs can be scheduled in staggered fashion to prevent long spans between drug doses.

leptiform discharges prior to next administration. In highly-tolerated medications with shorter half-lives, such as levetiracetam and valproate, dosing every 8 hours instead of twice daily should be considered. In cases of polytherapy, AEDs can be scheduled in staggered fashion to prevent long spans between drug doses. (2) Drug polytherapy Patients failing benzodiazepines and first-line scheduled AEDs are unlikely to achieve seizure termination without anesthetics; however, in SRSE, polytherapy is necessary to prevent seizure recurrence. Managing multiple AEDS requires consideration of several factors: risks and benefits of treatment, dosing strategies, and drug-drug interactions that alter drug efficacy and metabolism, and potentiate side effects. Second and third-line agents are listed in Table 1.

lytherapy is necessary to prevent seizure recurrence. Managing multiple AEDS requires consideration of several factors: risks and benefits of treatment, dosing strategies, and drug-drug interactions that alter drug efficacy and metabolism, and potentiate side effects. Second and third-line agents are listed in Table 1. In general, AEDs fall in to two broad categories: enzyme-inducers and enzyme-inhibitors. Table 3 categorizes commonly used AEDs. Concurrent use of an enzymeinducer and inhibitor can become problematic. This most commonly occurs when phenytoin, phenobarbital, or carbamazepine (inducers) are used in conjunction with valproic acid (an inhibitor), leading to supratherapeutic inducer levels and reductions in valproate concentrations by up to 50%–75% [75]. Free phenytoin, the active component of phenytoin, also increases with displacement from protein binding sites by valproate [76]. Increases in serum drug concentrations can result in toxicities at doses lower than expected. Carbamazepine, a second-line agent in SRSE, demonstrates auto-induction, where higher doses can result in increased drug metabolism [71].

henytoin, also increases with displacement from protein binding sites by valproate [76]. Increases in serum drug concentrations can result in toxicities at doses lower than expected. Carbamazepine, a second-line agent in SRSE, demonstrates auto-induction, where higher doses can result in increased drug metabolism [71]. (3) AED mechanism of action In polytherapy, the most effective regiments consider the pharmacokinetic properties and mechanisms of drug actions (MOAs). There is no proven benefit of one AED MOA compared to another in the treatment of SRSE. Similarly, no large study has investigated the efficacy of combining AEDs of different MOAs; however, this strategy is rationale, with early research suggesting polytherapy can be more effective and less toxic than monotherapy [77]. Use of consecutive AEDs with different MOAs is increasingly being employed in RSE protocols [78], and is common practice at our institution. First-line scheduled AEDs are typically sodium channel blockers (phenytoin), or have a broad spectrum of action (valproic acid, levetiracetam) [79]. The established MOAs of AEDs is outlined in Table 4. Some medications listed, including perampanel and ezogabine, have not yet been established in the treatment of SE, but are occasionally used at our hospital for SRSE given their novel MOA and overall tolerability. Although SE is often refractory to benzodiazepines, addition of a scheduled GABAergic benzodiazepine can be useful, especially in patients demonstrating responsiveness to midazolam infusion.

lished in the treatment of SE, but are occasionally used at our hospital for SRSE given their novel MOA and overall tolerability. Although SE is often refractory to benzodiazepines, addition of a scheduled GABAergic benzodiazepine can be useful, especially in patients demonstrating responsiveness to midazolam infusion. 2) Developments in antiepileptic drug therapy for RSE The development new AEDs for the treatment of SE has focused on optimizing drug pharmacodynamics and kinetics. Innovations in finding new molecular targets and advances in drug delivery systems has been promising. (1) New and investigational antiepileptic drugs Brivaracetam Brivaracetam has received 2016 approval in the United States and Europe for adjunctive therapy for partial seizures with potential application to the treatment of SE. It is a novel synaptic vesicle protein 2A (SV2A) ligand that has 10- to 20-fold higher affinity for its target compared to levetiracetam [80], with faster CNS entry and onset of activity [81]. In experimental models, brivaracetam has shown potent anticonvulsant properties, acting synergistically with ketamine and benzodiazepines in SE [77,82]. Further clinical experience with this new AED are needed, including trials investigating its efficacy in SE and RSE.

with faster CNS entry and onset of activity [81]. In experimental models, brivaracetam has shown potent anticonvulsant properties, acting synergistically with ketamine and benzodiazepines in SE [77,82]. Further clinical experience with this new AED are needed, including trials investigating its efficacy in SE and RSE. Allopregnanolone Allopregnanolone is a neurosteroid that modulates GABA-A receptors at both synaptic and extrasynaptic sites, increasing tonic neuronal inhibition [83]. Its efficacy in treating SRSE has been demonstrated in human case studies, allowing for successful weaning off anesthetics in both adults and children [84,85]. Given these results and the limited available therapies in SRSE, allopregnanolone is currently being studied in an international randomized, placebo-controlled phase III trial [86]. Other drugs under investigation Several other drugs have shown efficacy in animal models of SE, but are not yet widely available for human use. Valproic acid analogs, valnoctamide and sec-Butylpropylacetamide demonstrate broad spectrum anti-seizure activity and ability to terminate SE [87,88]. Carisbamate also shows broad spectrum of activity in seizure models by a yet undefined MOA that appears to be different than other established AEDs [89]. Preclinical data suggests carisbamate may protect against SE-induced neuronal damage [90].

e broad spectrum anti-seizure activity and ability to terminate SE [87,88]. Carisbamate also shows broad spectrum of activity in seizure models by a yet undefined MOA that appears to be different than other established AEDs [89]. Preclinical data suggests carisbamate may protect against SE-induced neuronal damage [90]. (2) New applications of established medications Ketamine Ketamine is an IV anesthetic for definitive seizure termination in SRSE that modulates GABA-A receptors and acts as a NMDA-receptor antagonist. Although there are no randomized controlled trials to support its use, there have been two case series and a number of case reports suggesting an overall success rate of up to 56% in RSE [59,91,92]. In these studies, a median of five AEDs were used prior to ketamine infusion [59]. The absence of cardiovascular depression with ketamine is appealing in patients receiving polytherapy; however, it may be associated with hypertension in a majority of cases [93]. In our experience, the addition of ketamine can allow for reductions in other anesthetic agents and avoidance of serious hypotensive episodes.

e absence of cardiovascular depression with ketamine is appealing in patients receiving polytherapy; however, it may be associated with hypertension in a majority of cases [93]. In our experience, the addition of ketamine can allow for reductions in other anesthetic agents and avoidance of serious hypotensive episodes. Inhaled anesthetics Isoflurane and desflurane are inhaled halogenated anesthetics with reported use in SE. Its ability to rapidly terminate seizure activity and achieve EEG suppression through uncomplicated drug titration is suggested from previous case series [94,95]; however, its widespread use has been limited. Side effects may be common and life-threatening, including severe hypotension and paralytic ileus, and machinery needed to administer the drug was not practical in most ICUs. The increasing availability of small self-contained vaporizers may result in new interest in this treatment modality. A recent case report suggests the efficacy of isoflurane in combination with mild hypothermia for SRSE [96]; however, the safety of inhaled anesthetics requires further study in larger trials.

ost ICUs. The increasing availability of small self-contained vaporizers may result in new interest in this treatment modality. A recent case report suggests the efficacy of isoflurane in combination with mild hypothermia for SRSE [96]; however, the safety of inhaled anesthetics requires further study in larger trials. 3) Additional treatments for SRSE (1) Non-surgical treatments Immunotherapy Immunotherapy in SRSE, even in the absence of any identifiable immunologic disease, may be reasonable in select cases. Of continued research interest, cerebral inflammation is an important pathologic process in SE that exacerbates neuronal damage and contributes to epileptogenesis [97,98]. Despite many cases of NORSE and SRSE remaining cryptogenic in etiology [14,18], immunotherapy can be empirically offered for treatment of occult autoimmune or paraneoplastic encephalitis, or other inflammatory disorder. Given the absence of high-quality evidence for this practice, the role of immunotherapy in SRSE is still unclear.

cases of NORSE and SRSE remaining cryptogenic in etiology [14,18], immunotherapy can be empirically offered for treatment of occult autoimmune or paraneoplastic encephalitis, or other inflammatory disorder. Given the absence of high-quality evidence for this practice, the role of immunotherapy in SRSE is still unclear. If provided, treatment typically begins with high-dose steroids (1 g IV methylprednisolone for 3 to 7 days). If there is no response to corticosteroids, plasma exchange or IV immunoglobulins over 3 to 5 sessions or doses may be employed. With partial response to immunotherapy and high suspicion for an underlying inflammatory disease, longstanding-immunotherapy can be continued with maintenance prednisone or other immunomodulators such as rituximab and cyclophosphamide [99]. Given the side effects associated with these medications, risks and benefits of therapy should be considered.

otherapy and high suspicion for an underlying inflammatory disease, longstanding-immunotherapy can be continued with maintenance prednisone or other immunomodulators such as rituximab and cyclophosphamide [99]. Given the side effects associated with these medications, risks and benefits of therapy should be considered. Ketogenic diet A ketogenic diet (4:1 ratio of fat-to-carbohydrate and protein) has both anti-seizure and anti-inflammatory effects, making it an appealing adjunct to AEDs [99]. Although largely studied in the pediatric population, given promising results, it has been similarly applied to adults. In a series of 10 patients with SRSE treated with ketogenic diet, nine patients achieved ketosis and SE termination in a median of 3 days [100]. A recent meta-analysis of 12 studies involving 270 patients with intractable epilepsy demonstrated a combined efficacy of 42%; however compliance was low [101]. Adequate ketosis is assessed through measuring urine and serum ketones. In our experience, involvement of dietary specialists is necessary to achieve and maintain ketogenesis in the ICU.

2 studies involving 270 patients with intractable epilepsy demonstrated a combined efficacy of 42%; however compliance was low [101]. Adequate ketosis is assessed through measuring urine and serum ketones. In our experience, involvement of dietary specialists is necessary to achieve and maintain ketogenesis in the ICU. Hypothermia Therapeutic hypothermia (TH) has long been an attractive treatment for brain injury, owing to a broad spectrum of activity that includes a reduction in inflammation, cerebral metabolic rate, oxidative stress, and cerebral edema. Early models of SE demonstrated anti-seizure and neuroprotective properties of TH [102,103]; however, until recently, there were only case reports to describe TH in human SRSE [99,104]. There has since been one multicentered randomized clinical trial comparing TH (goal, 32°C to 34°Cfor 24 hours) to standard medical treatment [105]. In the study of 270 patients, the rate of progression to EEG-confirmed SE on the first day was lower in the TH group; however, there was no significant difference between groups in 90-day outcomes. Some neurointensivists believe the effect of TH to be short-lasting, and should be used with caution in light of numerous adverse effects associated with systemic cooling and unclear efficacy [104,105].

first day was lower in the TH group; however, there was no significant difference between groups in 90-day outcomes. Some neurointensivists believe the effect of TH to be short-lasting, and should be used with caution in light of numerous adverse effects associated with systemic cooling and unclear efficacy [104,105]. Electroconvulsive therapy Electroconvulsive therapy is rarely used in the management of SRSE. This is a result of lacking data to support its use, limited availability in the ICU, and overall inexperience by most neurointensivists. A recent systemic review for electroconvulsive therapy for RSE identified 14 retrospective studies including a total of 19 patients. The review found seizure reduction in 57.9%, but highlighted the low quality of evidence in the current literature [106]. (2) Role of surgical evaluation and treatment Surgical evaluation When AEDs fail to control SRSE, surgical evaluation should be considered. In patients with a known etiologic structural lesion, early neurosurgical consultation is recommended; however, in the absence of such lesions, work-up and surgical management is challenging. The optimal timing for surgery is unknown, but some authors have suggested evaluation is appropriate after 2 weeks of failed medical therapy [107]. Future study should further identify criteria to select patients most likely to benefit from surgical intervention.

lesions, work-up and surgical management is challenging. The optimal timing for surgery is unknown, but some authors have suggested evaluation is appropriate after 2 weeks of failed medical therapy [107]. Future study should further identify criteria to select patients most likely to benefit from surgical intervention. Determining epileptogenic focality on EEG is typically the first diagnostic step. Lateralized and focal epileptiform discharges or seizures suggest a cortical target, while larger regionalized and widespread abnormalities suggest a deeper seizure focus, or diffuse or multi-focal injury. High resolution magnetic resonance imaging with and without contrast is recommended in all SRSE patients when feasible. Common findings include periictal edema and T2-weighted hyperintensities. Diffusion-weighted imaging lesions may reverse, while persistent changes signify irreversible neuronal damage and gliosis [108]. Ictal and inter-ictal positron emission tomography and single-photon emission computed tomography have been used in several case reports [109-112]. If a surgical lesion is suggested by non-invasive diagnostic studies, subdural EEG grid placement is necessary to further define the target and map eloquent brain regions.

Ictal and inter-ictal positron emission tomography and single-photon emission computed tomography have been used in several case reports [109-112]. If a surgical lesion is suggested by non-invasive diagnostic studies, subdural EEG grid placement is necessary to further define the target and map eloquent brain regions. Electrical stimulation therapies Vagal nerve stimulator placement has been described in the treatment of RSE in both children and adults. It is appealing for its widespread and tonic effect, while avoiding intracranial instrumentation. A recent systematic review identified 17 studies describing 28 patients, and found vagal nerve stimulator to be potentially effective in generalized RSE, but with poor response in focal RSE [113]. Deep brain stimulation and responsive neurostimulation may be the subject of future interest, but no recommendation for their use can be currently given. Focal resection and subpial transection In adults, focal resection and multiple subpial transections can be considered if a cortical target separated from eloquent brain regions is identified. It is the subject of a number of case reports [109-112], but the exact prevalence of these surgeries is unknown. Future investigations of electrical stimulation and surgery for SRSE will require combined interest from neurointensivists, epileptologists, and neurosurgeons. Whether earlier surgical evaluation is effective in terminating SRSE and improving patient outcomes should be determined.

Focal resection and subpial transection In adults, focal resection and multiple subpial transections can be considered if a cortical target separated from eloquent brain regions is identified. It is the subject of a number of case reports [109-112], but the exact prevalence of these surgeries is unknown. Future investigations of electrical stimulation and surgery for SRSE will require combined interest from neurointensivists, epileptologists, and neurosurgeons. Whether earlier surgical evaluation is effective in terminating SRSE and improving patient outcomes should be determined. Monitoring and Management of Complications Neurointensivists are responsible for considering the risks and benefits of any treatment. Expecting and managing complications is important in SRSE, where favorable outcomes are possible even after weeks-to-months of therapeutic coma, especially when no irreversible disease is found. A systemic review of RSE cases suggests longer SE duration is associated with worse outcomes; however, it is unknown if this is a result of the treatment itself, the underlying cause, or complications that arise with prolonged therapy [114].

-months of therapeutic coma, especially when no irreversible disease is found. A systemic review of RSE cases suggests longer SE duration is associated with worse outcomes; however, it is unknown if this is a result of the treatment itself, the underlying cause, or complications that arise with prolonged therapy [114]. Following termination of prolonged convulsive seizures, patients should be monitored for rhabdomyolysis and renal injury, affecting the choice of AED. Direct adverse effects of medications include cardiac arrhythmia, hypotension, transaminitis, renal injury, and cutaneous reactions. The risk of complications increases substantially with use of polytherapy and anesthetic agents. Because patients are often in coma, clinical manifestations of adverse events may be blunted. Patients on high-dose anesthetics, particularly propofol, should have serial chemistries and trigylcerides followed. Indirect adverse effects from prolonged sedation are cause of significant morbidity and mortality in SRSE. Most concerning to neurointensivists, immunosuppression from anesthetic therapy increases the risk of nosocomial infection by 3-fold [115]. Prolonged immobility additionally increases susceptibility to venous thromboembolism, paralytic ileus and decubitus ulcer formation [49]. Rarely, if attention is not paid to the duration of cEEG monitoring, clinically significant scalp ulcerations can occur. An EEG ‘vacation’ should be considered after extended monitoring if no drug titration is expected and EEG findings are unlikely to change management.

ralytic ileus and decubitus ulcer formation [49]. Rarely, if attention is not paid to the duration of cEEG monitoring, clinically significant scalp ulcerations can occur. An EEG ‘vacation’ should be considered after extended monitoring if no drug titration is expected and EEG findings are unlikely to change management. Conclusions SE and RSE are complex neurologic conditions. Although much has been learned from animal models, the pathophysiology of SRSE is still largely unknown. In a large proportion of cases, no underlying cause is found. There are increasingly more drug options to treat SE, but rational polytherapy should consider the pharmacodynamics and kinetics of established and new antiepileptic drugs. When seizures cannot be controlled with conventional medical therapy, non-conventional treatments, including early surgical evaluation can be considered; however, high-quality evidence for these strategies are lacking. Neurointensivists are challenged to reduce secondary brain injury by managing common complications. Future research should aim to identify specific drug regiments that are most effective, and to select which patients will most benefit from alternate therapies. No potential conflict of interest relevant to this article was reported. Table 1. Non-benzodiazepine first-line and second-line antiepileptic medications

Conclusions SE and RSE are complex neurologic conditions. Although much has been learned from animal models, the pathophysiology of SRSE is still largely unknown. In a large proportion of cases, no underlying cause is found. There are increasingly more drug options to treat SE, but rational polytherapy should consider the pharmacodynamics and kinetics of established and new antiepileptic drugs. When seizures cannot be controlled with conventional medical therapy, non-conventional treatments, including early surgical evaluation can be considered; however, high-quality evidence for these strategies are lacking. Neurointensivists are challenged to reduce secondary brain injury by managing common complications. Future research should aim to identify specific drug regiments that are most effective, and to select which patients will most benefit from alternate therapies. No potential conflict of interest relevant to this article was reported. Table 1. Non-benzodiazepine first-line and second-line antiepileptic medications Drug Initial dose Initial maintenance dose Clinical consideration First-line scheduled antiepileptic drug Phenytoin/fosphenytoin 15–20 mg/kg IV 100 mg every 8 h Narrow therapeutic range; calculated levels should be corrected for reduced GFR and hypoalbuminemia. Levetiracetam 30 mg/kg IV 500–1,000 mg every 12 h Few drug interactions; may cause agitation; unclear how rapid CNS penetration is Valproic acid 20–30 mg/kg IV 500 mg every 12 h May increase bleeding risk due to thrombocytopenia, reduced fibrinogen; high teratogenicity Phenobarbital 10–20 mg/kg IV bolus 1 mg/kg every 12 h High dose phenobarbital can aid in weaning off anesthetic agents. Patients can develop drug tolerance while maintaining therapeutic levels. Second-line scheduled antiepileptic drug Lacosamide 200–400 mg IV 200 mg every 12 h Associated with PR-prolongation on electrocardiogram; few drug interactions Topiramate 200–400 mg PO 300 mg every 6 h May be sedating; cannot be rapidly titrated Gabapentin 300–900 mg PO 300–900 mg every 8 h Few drug interactions; useful in patients with neuropathic pain First-line agents are commonly chosen for their ability to be given safely as an IV loading dose. Third line antiepileptic drugs include carbamazepine, oxcarbamazepine, zonisamide, vigabatrin, rufinamide, ezogabine, and perampanel.

PO 300–900 mg every 8 h Few drug interactions; useful in patients with neuropathic pain First-line agents are commonly chosen for their ability to be given safely as an IV loading dose. Third line antiepileptic drugs include carbamazepine, oxcarbamazepine, zonisamide, vigabatrin, rufinamide, ezogabine, and perampanel. IV: intravenous; GFR: glomerular filtration rate; CNS: central nervous system; PO: per oral. Table 2. Anesthetic infusions used for definite treatment of refractory status epilepticus Drug Loading dose (mg/kg) Maintenance dose (mg/kg/h) Clinical consideration Propofol 2–5 0.2–2.0 Rapid onset and offset facilitates neurologic examination; monitoring for propofol infusion syndrome with extended use Midazolam 0.1–0.3 5–30 Alternate to propofol that may cause less cardiovascular depression; associated with tachyphylaxis and drug accumulation Ketamine 1–3 0.5–10 Associated with hypertension; least amount of evidence to support its use Pentobarbital 5–10 0.5–5 Reserved for cases of failure of propofol and midazolam; associated with hypotension, hypothermia, and immunosuppression On initiation, anesthetic agents should be given as a bolus dose to reach therapeutic drug concentrations early. Table 3. Broad classification of antiepileptic drugs as significant enzyme inducers or inhibitors Enzyme inducer Enzyme inhibitor Phenytoin Valproic acid Primidone Zonisamidea Phenobarbital Carbamazepine Many antiepileptic drugs used in the management of status epilepticus induce or inhibit the activity of cytochrome P450 enzymes.

Table 3. Broad classification of antiepileptic drugs as significant enzyme inducers or inhibitors Enzyme inducer Enzyme inhibitor Phenytoin Valproic acid Primidone Zonisamidea Phenobarbital Carbamazepine Many antiepileptic drugs used in the management of status epilepticus induce or inhibit the activity of cytochrome P450 enzymes. a The addition of zonisamide will cause the increase of the carbamazepineepoxide only. Table 4. Mechanism of action of scheduled antiepileptic drugs used in refractory status epilepticus Drug GABA-agonist Glutamate antagonism Na+ channel Ca+ channel CA-inhibition Other Benzodiazepines ○ Phenytoin ○ ○ Levetiracetam ○ Presynaptic SV2A ligand Phenobarbital ○ ○ ○ Valproate ○ ○ ○ Lacosamide ○ Modulates CRMP2 protein Topiramate ○ ○ ○ ○ ○ Gabapentin ○ ○ Carbamazepine ○ ○ ○ Oxcarbamazepine ○ ○ ○ Zonisamide ○ ○ ○ Vigabatrin ○ Rufinamide ○ Ezogabine ○ Modulates voltage-gated potassium channels Perampanel Postsynaptic AMPA-receptor antagonism Medications commonly have a predominant mechanism of action, but may show anti-seizure activity by other secondary mechanisms. GABA: gamma-aminobutyric acid; CA: carbonic anhydrase; SV2A: synaptic vesicle protein 2A; CRMP2: collapsing-response mediator protein 2; AMPA: alpha-amino- 3-hydroxy-5-methyl-4-isoxazoleproprionic acid.

Introduction Human metapneumovirus (hMPV) was first identified in 2001, the Netherlands from a pediatric patient who had symptoms similar to those of respiratory syncytial virus (RSV) infection. It is a member of the paramyxovirus family and is genetically similar to RSV [1]. Typically hMPV infections occur between March and April, and account for 7% of respiratory tract infections [2]. A hMPV infection commonly occurs in children less than 2 years old and manifests as mild flu-like symptoms, similar to RSV [3]. Furthermore, hMPV is a major contributor to the burden of wintertime respiratory illness in older adults that is peak incidence at 65 years of age and immunocompromised individuals [2,4,5]. hMPV infections in children are usually mild and self-limiting, but in elderly and immunocompromised patients, the clinical course can progress to acute respiratory distress syndrome (ARDS) [6]. Studies of patients with hMPV who develop severe illness have focused on children; few have involved adults [7,8]. Nosocomial infection has been reported in several studies as a mode of transmission [5,9-11]. Nosocomial hMPV infection of adults occurs predominantly in human immunodeficiency virus-infected persons [12].

Studies of patients with hMPV who develop severe illness have focused on children; few have involved adults [7,8]. Nosocomial infection has been reported in several studies as a mode of transmission [5,9-11]. Nosocomial hMPV infection of adults occurs predominantly in human immunodeficiency virus-infected persons [12]. There are few studies on hMPV infection of adults in Korea. And the number of immunocompromised patients is increasing in hospitals. These patients are also vulnerable to previously neglected pathogens. Therefore, we designed a retrospective review of hMPV-infected adults. The clinical characteristics of the patients—including demographic data, comorbidities, presence of pneumonia or ARDS, acquisition site (community-acquired or nosocomial), and risk factors for ARDS—were reviewed.

e to previously neglected pathogens. Therefore, we designed a retrospective review of hMPV-infected adults. The clinical characteristics of the patients—including demographic data, comorbidities, presence of pneumonia or ARDS, acquisition site (community-acquired or nosocomial), and risk factors for ARDS—were reviewed. Materials and Methods This study was a retrospective review of medical records; institutional review board approval was obtained (No. GAIRB2016-179). We identified all patients at Gachon University Gil Medical Center, a tertiary referral hospital, with a laboratory-confirmed hMPV infection as diagnosed by positive multiplex real time-polymerase chain reaction (mRT-PCR) between January 2012 and April 2016. Nasopharyngeal or oropharyngeal samples were obtained in the emergency department, in an inpatient setting, or in an outpatient setting. All diagnostic testing was performed at the Gil Medical Center Laboratory Department. Multiplex real time-PCR was performed using Anyplex II RV16 Detection kit (Seegene, Seoul, Korea). This kit detected in parallel for the following 16 respiratory viruses: human bocavirus, human enterovirus, influenza virus A and B, parainfluenza virus 1, 2, 3, and 4, RSV A and B, coronavirus OC43, 229E, NL63, human rhinovirus A/B/C, and metapneumovirus. A total of 652 cases of hMPV infection were identified by mRT-PCR from nasopharyngeal or oropharyngeal swabs among patients. We collected hospitalized adult patients over 18 years with positive hMPV mRT-PCR assay result among 652 patients. The 110 of hospitalized adult patients with hMPV infection were included in the primary analysis.

ses of hMPV infection were identified by mRT-PCR from nasopharyngeal or oropharyngeal swabs among patients. We collected hospitalized adult patients over 18 years with positive hMPV mRT-PCR assay result among 652 patients. The 110 of hospitalized adult patients with hMPV infection were included in the primary analysis. Demographic, comorbidity, and hospitalization data of 110 adult patients were collected. The data included chronic respiratory disease, cardiac disease, immunocompromised state, end-stage renal disease (ESRD) requiring dialysis, cirrhosis, corticosteroid use, and immunosuppressive therapy. Viral co-infection was defined as identification by mRT-PCR from a nasopharyngeal or oropharyngeal swab. Patients who had either a positive blood culture or a positive sputum culture from a satisfactory specimen were identified as having bacterial co-infection. We reviewed the medical records of the patients to identify those admitted to the intensive care unit (ICU) with a nosocomial infection and who met the criteria for ARDS using the Berlin definition [13]. The incubation period is estimated to be 4 to 6 days [14]. Nosocomial infection was defined as an infection that occurred up to 5 days after discharge or more than 5 days in admission. Death information of patients was confirmed using hospital medical record and obliterated information of national health insurance service which means death.

is estimated to be 4 to 6 days [14]. Nosocomial infection was defined as an infection that occurred up to 5 days after discharge or more than 5 days in admission. Death information of patients was confirmed using hospital medical record and obliterated information of national health insurance service which means death. Clinically relevant parameters were identified on the day of execution of mRT-PCR and admission to the ICU or ward; these included temperature, respiratory rate, applied fraction of inspired oxygen, partial pressure of arterial oxygen, white blood cell count, blood urea nitrogen (BUN)/creatinine, total bilirubin, C-reactive protein (CRP) level, and procalcitonin level. The patients were divided into non-ARDS and ARDS groups, and the risk factors and laboratory values for ARDS were assessed. 1) Statistical analysis All data analyses were performed using SPSS version 20.0 (IBM Corp., Armonk, NY, USA). Categorical variables were compared by Fisher exact test, and continuous variables by Wilcoxon’s rank-sum test. The impact of potential risk factors on the development of ARDS was analyzed with univariate logistic regression analysis. Statistically significant variables at univariate analysis were included into a multivariate logistic regression analysis with backward elimination to identify independent risk factors of ARDS. The independent influences of risk factors for ARDS were expressed as the odds ratio (OR) with their 95% confidence intervals. Significance was taken as P < 0.05.

at univariate analysis were included into a multivariate logistic regression analysis with backward elimination to identify independent risk factors of ARDS. The independent influences of risk factors for ARDS were expressed as the odds ratio (OR) with their 95% confidence intervals. Significance was taken as P < 0.05. Results 1) Characteristics of adults with a hMPV infection The mean age of the adult patients was 61.4 ± 16.6 years, and there was no difference according to sex (Table 1). Overall, 19 patients (17.3%) had viral co-infections, and 22 (20.0%) had bacterial co-infections. Co-infection did not affect event of ARDS, nosocomial infection and mortality. Most patients (n = 105, 95.5%) had comorbidities; these included diabetes, malignancy, pulmonary disease, cardiac disease, ESRD, and liver cirrhosis and they were being treated with corticosteroid or immunosuppressive medication. Almost patients had pneumonia on chest X-ray (93.6%), and 22 patients (20%) had ARDS (Table 1). Half of the patients with hMPV-associated ARDS had severe disease according to the Berlin definition. The in-hospital mortality rate was 10.9%, and the 1-year all-cause mortality rate was 15.5%. The patients with ARDS showed higher in-hospital (36.4%) and 1-year all-cause (40.9%) mortality rates than those without ARDS (Table 1). Forty-three patients were found to have nosocomial infections (39%) (Table 1). Regarding laboratory findings, lymphocytopenia (<1,500/mm3 ), high levels of CRP, BUN and creatinine, and low levels of albumin were observed (Table 2).

(36.4%) and 1-year all-cause (40.9%) mortality rates than those without ARDS (Table 1). Forty-three patients were found to have nosocomial infections (39%) (Table 1). Regarding laboratory findings, lymphocytopenia (<1,500/mm3 ), high levels of CRP, BUN and creatinine, and low levels of albumin were observed (Table 2). 2) Risk factors for hMPV-associated ARDS The non-ARDS group included 88 patients while the ARDS group included 22 patients. The mean age of the non-ARDS group was 59.8 years, while that of the ARDS group was 68.0 years (OR, 1.034; P = 0.040). However, multivariate logistic regression showed age was not significant (Table 3). The pattern of viral and bacterial co-infection did not differ between the ARDS and non-ARDS groups (Table 1). The rates of comorbidities in the hMPV-associated ARDS patients were similar in hMPV-associated ARDS patients, with the exception of congestive heart failure (OR, 5.249; P = 0.044) (Table 3). The inhospital and 1-year all-cause mortality rates of the ARDS patients were 36.4% and 40.9%, respectively (Table 1). Additional analysis showed Kendal rank correalation coefficient was 0.752 (P = 0.001) between in hospital mortality and 1-year mortality. Therefore 1-year all-cause mortality was not independent variable.

inhospital and 1-year all-cause mortality rates of the ARDS patients were 36.4% and 40.9%, respectively (Table 1). Additional analysis showed Kendal rank correalation coefficient was 0.752 (P = 0.001) between in hospital mortality and 1-year mortality. Therefore 1-year all-cause mortality was not independent variable. 3) Characteristics of nosocomial hMPV infection Overall, 67 patients had community-acquired hMPV infections and 43 had nosocomial infections. The patients with nosocomial infection showed the presence of comorbidities (e.g., solid tumor, hematologic disorder, liver cirrhosis, and immunosuppressive therapy; P < 0.05). However, the rates of mortality and hMPV-associated ARDS were not different between the community-acquired and nosocomial infection groups (Table 4).

tients with nosocomial infection showed the presence of comorbidities (e.g., solid tumor, hematologic disorder, liver cirrhosis, and immunosuppressive therapy; P < 0.05). However, the rates of mortality and hMPV-associated ARDS were not different between the community-acquired and nosocomial infection groups (Table 4). Discussion Our study is the first to characterize ARDS and mortality in the adult patients with hMPV infections in Korea. This study evaluated the clinical features of 110 hospitalized adults with a positive hMPV mRT-PCR assay result at a tertiary referral hospital in South Korea from 2012 to 2016. This was a retrospective study of hMPV-infected adults and provided detailed clinical data, including comorbidities, radiologic findings, and laboratory findings, as well as the demographic data of the subjects. Most of the adults with hMPV infection had pneumonia on chest X-ray, and 20% had ARDS. The in-hospital mortality rate of ARDS patients was 36.4%, and 10.9% of all patients died during hospitalization. Nosocomial infection occurred in 39.1% of the subjects. Other study shows forty-eight percent of ICU patient is ARDS and mortality is 18% [8]. Although the scope of the studies, it is necessary to focus on high ARDS incidence and high mortality rates.

ts was 36.4%, and 10.9% of all patients died during hospitalization. Nosocomial infection occurred in 39.1% of the subjects. Other study shows forty-eight percent of ICU patient is ARDS and mortality is 18% [8]. Although the scope of the studies, it is necessary to focus on high ARDS incidence and high mortality rates. The increasing availability and broadened scope of viral respiratory polymerase chain reaction panels helps our understanding of viral pathogens. The mRT-PCR assay had a sensitivity and specificity higher than 88% and 98.6%. And the mRT-PCR could detect co-infecting respiratory viruses, even at low viral loads that cannot be detected using culture techniques [15]. During the study period, the peak incidence of hMPV infection was in April, after that of influenza which had peak in February. A study in a similar area conducted from 2000 to 2005 reported a peak incidence of hMPV infection in January to March [16], and another in March to April [17]. In this study, 17% of the subjects had viral co-infections, most frequently with influenza virus (Table 4). This is in agreement with a previous report [4]. These co-infections may be due to the overlap of the periods of peak incidence of these viruses. hMPV infections occur between March and April. And RSV infection is in December and Influenza is in January and February (Supplementary Figure 1). A high percentage (90% to 100%) of children in 5–10 years old have been infected. 78%, under 5 years of age in this study (Supplementary Figure 2).

cidence of these viruses. hMPV infections occur between March and April. And RSV infection is in December and Influenza is in January and February (Supplementary Figure 1). A high percentage (90% to 100%) of children in 5–10 years old have been infected. 78%, under 5 years of age in this study (Supplementary Figure 2). The mean age of the adult patients was 61.4 years, and all but five had comorbidities such as diabetes, malignancy, airway obstructive diseases, ESRD, and heart failure. Such comorbidities and old age can result in an immunocompromised state, possibly leading to the development of severe viral infections such as pneumonitis or ARDS. hMPV infections in healthy adults usually show mild clinical features from asymptomatic to upper respiratory symptoms, similar to RSV infection [18]. However, frail elderly patients with an hMPV infection could progress to pneumonitis, ARDS, or death [8,12,18]. In this study, severe hMPV infection was associated with a risk of pneumonia, ARDS, and mortality in immunocompromised elderly patients. The findings of this study suggest that 20% of hospitalized adults with an hMPV infection will develop ARDS. The risk factor for hMPV-associated ARDS was congestive heart failure. It is known that viral infection is one of the risk factors for aggravation of congestive heart failure. When patients with congestive heart failure are infected by hMPV, they have risks both aggravation of congestive heart failure and ARDS [8]. The patients with ARDS and congestive heart failure may also have had pulmonary edema; however, this could not be determined because of the limitations inherent to a retrospective study. The patients with ARDS among hMPV-infected adults showed lymphocytopenia, higher levels of BUN, creatinine, and CRP and lower levels of albumin compared to non-ARDS patients. In-hospital mortality of severe ARDS (OR, 9.48) was higher than ARDS (OR, 5.24). The in-hospital mortality rate of the patients with hMPV infection was 10.9%. The in-hospital mortality rate of the ARDS group (36.4%) was significantly higher than that of the non-ARDS group (5.7%). The risk factors for hMPV-associated ARDS should be considered as congestive heart failure. These results suggest that hMPV is an important and previously underappreciated cause of ARDS in adult patients. Severe clinical manifestations such as ARDS should be considered when hMPV infection is diagnosed in adult patients with congestive heart failure.

iated ARDS should be considered as congestive heart failure. These results suggest that hMPV is an important and previously underappreciated cause of ARDS in adult patients. Severe clinical manifestations such as ARDS should be considered when hMPV infection is diagnosed in adult patients with congestive heart failure. The hMPV infection has been reported in immunocompromised patients, including lung transplant and hematopoietic stem cell transplant recipients [3,19]. Moreover, outbreaks in healthcare facilities have been reported [2,9]. When considering the implications of these findings, it should be noted that the risk of nosocomial hMPV infection is on the rise among patients with blood cancer or patients with a solid tumor [20] who were recently admitted for chemotherapy. The additional analysis showed that immunosuppressive therapy is interaction variable with cancer. A patient with liver cirrhosis was not explained as risk factor of nosocomial infection because small cases and a further research is needed. Such patients typically reside in shared rooms containing four to six patients. Therefore, the possibility of hMPV infection should be considered in long-term immunocompromised patients residing in shared rooms. Droplet respiratory precautions are appropriate for preventing the transmission of hMPV, particularly in immunocompromised patients.

reside in shared rooms containing four to six patients. Therefore, the possibility of hMPV infection should be considered in long-term immunocompromised patients residing in shared rooms. Droplet respiratory precautions are appropriate for preventing the transmission of hMPV, particularly in immunocompromised patients. This study has multiple limitations. Whether hMPV was the pathogen in all patients enrolled in this study was unclear. However, the mRT-PCR assay used has high sensitivity and specificity. Also, the rate of co-infection with other pathogens was not high (Table 5). Future studies should assess the clinical features of hMPV infection in adults, determine the clinical potential of the virus as a respiratory pathogen to induce severe ARDS, and develop novel antiviral agents or vaccines. In conclusion, this study suggests that hMPV is an important respiratory pathogen that causes pneumonia/ARDS in elderly, immunocompromised individuals and that it is transmitted via the nosocomial route. Clinicians need to consider mMPV virus and sufficient attention is required. No potential conflict of interest relevant to this article was reported. Supplementary Materials The online-only Supplement data are available with this article online: https://doi.org/10.4266/kjccm.2017.00038. Table 1. Characteristics of hMPV-infected patients according to ARDS or non-ARDS group

In conclusion, this study suggests that hMPV is an important respiratory pathogen that causes pneumonia/ARDS in elderly, immunocompromised individuals and that it is transmitted via the nosocomial route. Clinicians need to consider mMPV virus and sufficient attention is required. No potential conflict of interest relevant to this article was reported. Supplementary Materials The online-only Supplement data are available with this article online: https://doi.org/10.4266/kjccm.2017.00038. Table 1. Characteristics of hMPV-infected patients according to ARDS or non-ARDS group All Non-ARDS ARDS P-value No. of patients 110 88 22 Age (yr) 61.4 ± 16.6 59.8 ± 16.6 68.0 ± 15.2 0.036 Male sex 46 (41.8) 36 (40.9) 10 (45.5) 0.439 Co-infection Virus 19 (17.3) 17 (19.3) 2 (9.1) 0.211 Bacteria 22 (20.0) 15 (17.0) 7 (31.8) 0.108 Comorbidity Solid tumor 19 (17.3) 16 (18.2) 3 (13.6) 0.681 Hematologic disorder 10 (9.1) 8 (9.1) 2 (9.1) 0.442 DM 29 (26.4) 21 (23.9) 8 (36.4) 0.178 Asthma 12 (10.9) 10 (11.4) 2 (9.1) 0.555 COPD 12 (10.9) 9 (10.2) 3 (13.6) 0.445 ILD 2 (1.8) 1 (1.1) 1 (4.5) 0.361 ESRD 7 (6.4) 4 (4.5) 3 (13.6) 0.141 LC 5 (4.5) 5 (5.7) 0 0.20 KT 3 (2.7) 3 (3.4) 0 0.508 CHF 7 (6.3) 3 (3.4) 4 (18.2) 0.029 IHD 4 (3.6) 2 (2.3) 2 (9.1) 0.178 Steroid useda 15 (13.6) 13 (14.8) 2 (9.1) 0.383 Immunotherapyb 17 (15.5) 14 (15.9) 3 (13.6) 0.545 Nosocomial infection 43 (39.1) 36 (40.9) 7 (31.8) 0.299 Pneumonia 103 (93.6) 83 (94.3) 20 (90.9) 0.560 In-hospital mortality 12 (10.9) 5 (5.7) 8 (36.4) 0.001 1-Year all-cause mortality 17 (15.5) 9 (10.2) 9 (40.9) 0.002 Values are presented as mean ± standard deviation or number (%).

b 17 (15.5) 14 (15.9) 3 (13.6) 0.545 Nosocomial infection 43 (39.1) 36 (40.9) 7 (31.8) 0.299 Pneumonia 103 (93.6) 83 (94.3) 20 (90.9) 0.560 In-hospital mortality 12 (10.9) 5 (5.7) 8 (36.4) 0.001 1-Year all-cause mortality 17 (15.5) 9 (10.2) 9 (40.9) 0.002 Values are presented as mean ± standard deviation or number (%). hMPV: human metapneumovirus; ARDS: acute respiratory distress syndrome; DM: diabetes mellitus; COPD: chronic obstructive pulmonary disease; ILD: interstitial lung disease; ESRD: end-stage renal disease; LC: liver cirrhosis; KT: kidney transplantation; CHF: congestive heart failure; IHD: ischemic heart disease. a Use of an inhaled or systemic corticosteroid; b Cytostatics, antirejection medications, and disease-modifying antirheumatic drugs. Table 2. Laboratory parameters of hMPV-infected patients according to non-ARDS and ARDS group All Non-ARDS ARDS P-value No. of patients 110 88 22 Hemoglobin (g/dl) 11.4 ± 2.2 11.6 ± 2.1 10.6 ± 2.5 0.053 Platelet (×103/mm3) 195.7 ± 92.9 195.5 ± 96.2 196.5 ± 80.5 0.967 WBC count (×103/mm3) 8.1 (9.6) 8.3 (10.2) 6.3 (9.2) 0.791 Lymphocytes (/mm3) 913.8 (961.1) 983.2 (851.2) 638.8 (1,971.6) 0.019 BUN (mg/dl) 16.4 (21.8) 14.6 (15.7) 35.5 (35.9) 0.002 Cr (mg/dl) 0.85 (0.7) 0.8 (0.6) 1.0 (2.4) 0.023 Total bilirubin (mg/dl) 0.65 (0.6) 0.7 (0.6) 0.5 (0.4) 0.247 Albumin (g/dl) 3.5 ± 0.6 3.6 ± 0.6 3.1 ± 0.6) 0.002 CRP (mg/dl) 9.48 ± 8.5 7.8 ± 7.3 16.1 ± 9.6 0.001 Procalcitonin (ng/ml) 0.83 (2.31) 0.57 (1.15) 2.35 (6.57) 0.165 Values are presented as mean ± standard deviation or median (interquartile range).

2.4) 0.023 Total bilirubin (mg/dl) 0.65 (0.6) 0.7 (0.6) 0.5 (0.4) 0.247 Albumin (g/dl) 3.5 ± 0.6 3.6 ± 0.6 3.1 ± 0.6) 0.002 CRP (mg/dl) 9.48 ± 8.5 7.8 ± 7.3 16.1 ± 9.6 0.001 Procalcitonin (ng/ml) 0.83 (2.31) 0.57 (1.15) 2.35 (6.57) 0.165 Values are presented as mean ± standard deviation or median (interquartile range). hMPV: human metapneumovirus; ARDS: acute respiratory distress syndrome; WBC: white blood cell; BUN: blood urea nitrogen; Cr: serum creatinine; CRP: C-reactive protein. Table 3. Risk factors for ARDS among patients with hMPV infection Variable Univariate analysis Multivariate analysis OR 95% CI P-value OR 95% CI P-value Age 1.034 1.00–1.07 0.040 1.031 0.99–1.07 0.070 CHF 6.296 1.30–30.60 0.023 5.249 1.04–26.39 0.044 Risk factor analysis (Nagelkerke R2 value = 0.120, Hosmer and Lemeshow goodness of fit with a χ2 value = 0.762). ARDS: acute respiratory distress syndrome; hMPV: human metapneumovirus; OR: odds ratio; CI: confidence interval; CHF: congestive heart failure. Table 4. Characteristics of hMPV-infected patients according to acquisition site: community-acquired or nosocomial infection

Variable Univariate analysis Multivariate analysis OR 95% CI P-value OR 95% CI P-value Age 1.034 1.00–1.07 0.040 1.031 0.99–1.07 0.070 CHF 6.296 1.30–30.60 0.023 5.249 1.04–26.39 0.044 Risk factor analysis (Nagelkerke R2 value = 0.120, Hosmer and Lemeshow goodness of fit with a χ2 value = 0.762). ARDS: acute respiratory distress syndrome; hMPV: human metapneumovirus; OR: odds ratio; CI: confidence interval; CHF: congestive heart failure. Table 4. Characteristics of hMPV-infected patients according to acquisition site: community-acquired or nosocomial infection Community-acquired Nosocomial P-value No. of patients 67 43 Age (yr) 63.5 ± 17.1 58.3 ± 15.5 0.113 Male sex 28 (41.8) 18 (41.9) 0.575 Co-infection Virus 9 (13.4) 10 (23.3) 0.142 Bacteria 13 (19.4) 9 (20.9) 0.515 Comorbidity Solid tumor 6 (9.0) 13 (30.2) 0.005 Hematologic disorder 1 (1.5) 9 (20.9) 0.001 DM 18 (26.9) 11 (25.6) 0.532 Asthma 10 (14.9) 2 (4.7) 0.081 COPD 9 (13.4) 3 (7.0) 0.231 ILD 2 (3.0) 0 0.369 ESRD 4 (6.0) 3 (7.0) 0.564 LC 0 5 (11.6) 0.008 KT 1 (1.5) 2 (4.7) 0.338 CHF 6 (9.0) 1 (2.3) 0.162 IHD 2 (3.0) 2 (4.7) 0.510 Steroid useda 11 (16.4) 2 (9.3) 0.221 Immunosuppressive therapyb 4 (6.0) 13 (30.2) 0.001 Initial pneumonia 61 (91.1) 42 (97.7) 0.166 ARDS 15 (22.4) 7 (16.3) 0.299 In-hospital mortality 8 (11.9) 5 (11.6) 0.606 1-Year all-cause mortality 8 (11.9) 10 (23.3) 0.098 Values are presented as mean ± standard deviation or number (%).

id useda 11 (16.4) 2 (9.3) 0.221 Immunosuppressive therapyb 4 (6.0) 13 (30.2) 0.001 Initial pneumonia 61 (91.1) 42 (97.7) 0.166 ARDS 15 (22.4) 7 (16.3) 0.299 In-hospital mortality 8 (11.9) 5 (11.6) 0.606 1-Year all-cause mortality 8 (11.9) 10 (23.3) 0.098 Values are presented as mean ± standard deviation or number (%). hMPV: human metapneumovirus; DM: diabetes mellitus; COPD: chronic obstructive pulmonary disease; ILD: interstitial lung disease; ESRD: end-stage renal disease; LC: liver cirrhosis; KT: kidney transplantation; CHF: congestive heart failure; IHD: ischemic heart disease. a Use of an inhaled or systemic corticosteroid; b Cytostatics, antirejection medications, and disease-modifying antirheumatic drugs. Table 5. Etiologies of co-infection in patients with hMPV infection No. Confirmed species (no.) Bacteria Pneumonia 5 S.pneumoniae (2) VAP 2 K.pneumoniae (1) Sepsis 3 E.coli (1) / Chlamydiae (1) CRBSI 2 MRSA (3) Colits 2 C. difficile (2) UTI 2 Liver abscess 1 K. pneumoniae (1) Wound infection 4 Virus Coronavirus 5 Influenza A 5 Influenza B 1 Adenovirus 2 Rhinovirus 2 Cytomegalovirus 2 RSV 1 Bocavirus 1 hMPV: human metapneumovirus; VAP: ventilator associated pneumonia; CRBSI: catheter related bloodstream infection; MRSA: methicillin resistant Staphylococcus aureus; UTI: urinary tract infection; RSV: respiratory syncytial virus.

Cardiopulmonary resuscitation (CPR) is an inevitable procedure used to save lives during sudden cardiac arrest. However, mechanical resuscitation measures, such as chest compression, can result in resuscitation-related trauma such as thoracic injuries. Herein, we describe a case of traumatic aortic dissection associated with cardiac compression in a patient with anaphylactic cardiac arrest who survived after CPR with extracorporeal membrane oxygenation (ECMO) support.

es, such as chest compression, can result in resuscitation-related trauma such as thoracic injuries. Herein, we describe a case of traumatic aortic dissection associated with cardiac compression in a patient with anaphylactic cardiac arrest who survived after CPR with extracorporeal membrane oxygenation (ECMO) support. Case Report The patient was a 54-year-old man who was scheduled to undergo open subtotal gastrectomy for gastric cancer. He had no underlying diseases, except dyslipidemia, and no history of allergic responses to drugs. The patient was transferred to the operating room without any abnormal signs or symptoms. Midazolam (2 mg), lidocaine (20 mg), propofol (60 mg), and cisatracurium (8 mg) were administered intravenously for induction of general anesthesia. After endotracheal intubation but before skin incision, 1 g cefotetan was administered intravenously as a prophylactic antibiotic. A few minutes later, the patient’s heart rate suddenly dropped to less than 30 beats/min, according to electrocardiography monitoring, and he went into cardiac arrest. CPR was initiated immediately, and transesophageal echocardiography performed during CPR showed akinesia and low ejection fraction in the left ventricle. Because normal cardiac contractility was not recovered during CPR, veno-arterial ECMO was performed. The catheters for arterial inflow (catheter size: 15 Fr) and venous outflow (catheter size, 21 Fr) for ECMO were inserted through the right femoral artery and the left femoral vein, respectively. CPR including chest compressions was performed for 35 minutes. After return of spontaneous circulation, the patient’s blood pressure was 110/60 mmHg, his pulse rate was 70 beats/min, and the level of lactic acid reached 12.1 mmol/L. The patient was transferred to the intensive care unit (ICU) without performing the planned operation. His initial vital signs on admission to the ICU were 114/64 mmHg, 110 beats/min, and a lactic acid level of 9.7 mmol/L. Initial chest radiography in the ICU showed diffuse pulmonary edema and subsegmental atelectasis (Figure 1). Transthoracic echocardiography (TTE) performed on the day of ICU admission confirmed normal left and right ventricle contractility and 59% ejection fraction. Repositioning of the ECMO catheter was considered to correct the malposition; however, ECMO was stopped after 9 hours due to confirmation of normal cardiac function on TTE. Mechanical ventilation was maintained for 18 days.

ICU admission confirmed normal left and right ventricle contractility and 59% ejection fraction. Repositioning of the ECMO catheter was considered to correct the malposition; however, ECMO was stopped after 9 hours due to confirmation of normal cardiac function on TTE. Mechanical ventilation was maintained for 18 days. Owing to persistent loss of renal function, renal replacement therapy was continued after recovery. Twenty days after his cardiac arrest, abdominal computed tomography (CT) was performed for re-evaluation and therapeutic planning of the remaining malignant gastric mass. The examination revealed left fourth and fifth rib fractures as well as an aortic dissection extending from the proximal descending thoracic aorta to the abdominal aorta; these findings were not observed on previous evaluations and were detected incidentally. The DeBakey type III aortic dissection extended from the distal arch of the thoracic aorta to the proximal level of the renal artery involving the celiac trunk (Figure 2). Twenty-five days after CPR, the patient complained of atypical chest pain in the epigastric area, and angiographic CT and TTE were performed to evaluate the extent of the aortic dissection and associated complications compared with the previous abdominal CT scan. Because there were no interval changes, it was considered an uncomplicated type B aortic dissection with no sign of malperfusion of the major vessels. Medical treatment including antihypertensive drugs was continued. One month later, the patient was diagnosed with an anaphylactic reaction to cefotetan based on a positive response to a skin prick test.

anges, it was considered an uncomplicated type B aortic dissection with no sign of malperfusion of the major vessels. Medical treatment including antihypertensive drugs was continued. One month later, the patient was diagnosed with an anaphylactic reaction to cefotetan based on a positive response to a skin prick test. Discussion CPR is a critical procedure for saving patients experiencing cardiac arrest by maintaining auxiliary cardiac circulation and ventilation until return of spontaneous cardiac circulation. In 1960, Kouwenhoven et al. [1] developed a method of external cardiac massage without thoracotomy for patients in cardiac arrest. Approximately 200,000 cases of in-hospital cardiac arrest occur each year in the United States[2]. The rate of survival to discharge is an estimated 15-20% [3,4]. The survival rate after in-hospital cardiac arrest has improved steadily owing to quality improvement efforts [4]. In addition, several studies have suggested that ECMO during CPR has improved the survival rates of in-hospital cardiac arrest [5-7]. Along with increased survival after CPR, there has been a corresponding increase in concern about CPR-related complications in survivors of cardiac arrest.

ality improvement efforts [4]. In addition, several studies have suggested that ECMO during CPR has improved the survival rates of in-hospital cardiac arrest [5-7]. Along with increased survival after CPR, there has been a corresponding increase in concern about CPR-related complications in survivors of cardiac arrest. Resuscitation measures include invasive manipulations such as external cardiac compression, endotracheal intubation, and electric impulses for defibrillation. These manipulations support recovery of spontaneous circulation and termination of CPR, but patients may also experience traumatic injuries due to mechanical chest compression. In a study of approximately 1,000 autopsy cases, resuscitation-related injuries were reported in an estimated 21% to 65% of cases [8]. The most common complications after CPR are thoracic skeletal injuries, particularly fractures of the ribs (13% to 97%) and sternum (1% to 43%) [9]. Fractured ribs can cause mediastinal hematoma, hemothorax, pneumothorax, and pulmonary contusion [10]. Direct cardiac laceration and great vessel injuries are relatively rare but can lead to death [8].

after CPR are thoracic skeletal injuries, particularly fractures of the ribs (13% to 97%) and sternum (1% to 43%) [9]. Fractured ribs can cause mediastinal hematoma, hemothorax, pneumothorax, and pulmonary contusion [10]. Direct cardiac laceration and great vessel injuries are relatively rare but can lead to death [8]. Aortic dissection after CPR has rarely been reported in survivors [11-13]. Most blunt aortic injuries occur in the proximal descending aorta, where the relatively mobile aortic arch joins the fixed descending aorta via the ligamentum arteriosum. The shearing forces of sudden deceleration that occur in motor vehicle collisions can cause traumatic aortic dissection. Other mechanisms for this injury include sudden elevated intraluminal aortic pressure due to compression between the sternum and thoracic vertebrae [14]. The symptoms, clinical signs, and physical examination findings are nonspecific and difficult to differentiate from aortic dissection after CPR. A majority of CPR survivors cannot complain of symptoms because of decreased mentality due to hypoxic brain damage or a sedated state for post-CPR management, including mechanical ventilation. In these cases, the first clinical sign may be the onset of hemodynamic instability.

rentiate from aortic dissection after CPR. A majority of CPR survivors cannot complain of symptoms because of decreased mentality due to hypoxic brain damage or a sedated state for post-CPR management, including mechanical ventilation. In these cases, the first clinical sign may be the onset of hemodynamic instability. Chest radiography is typically used for screening, but it has a low sensitivity for diagnosing traumatic aortic injuries. Aortic injury after CPR may be suspected when chest radiography findings suggest mediastinal widening, loss of the aortopulmonary window, rightward tracheal/nasogastric tube deviation, left main stem bronchus depression, and left apical cap [15]. Imaging studies useful for diagnosis include CT and transesophageal echocardiography (TEE). Chest CT has high diagnostic sensitivity, exceeding 97% to 100% for aortic dissection [16]. Chest CT can also be used to evaluate other associated damage, such as thoracic skeletal injuries and lung parenchymal injuries. TEE can be performed at bedside in patients with hemodynamic instability to evaluate the myocardium for wall motion abnormalities. However, TEE has limitations such as dependency on operator skill and poor visualization of the distal ascending aorta and proximal arch [17].

eletal injuries and lung parenchymal injuries. TEE can be performed at bedside in patients with hemodynamic instability to evaluate the myocardium for wall motion abnormalities. However, TEE has limitations such as dependency on operator skill and poor visualization of the distal ascending aorta and proximal arch [17]. Aortic dissection after CPR is rare but can be fatal. In the absence of previous evaluation, it can also be difficult to confirm a causal relationship between CPR-induced injury and the underlying disease that caused the sudden cardiac arrest. The 2010 American Heart Association (AHA) guidelines recommend harder (a depth of at least 5 cm) and faster compressions (a rate of at least 100 compressions per minute) than the 2005 guidelines [18]. Increasing the requirements of compression depth and rate may increase the risk of CPR-related complications, including life-threatening events. Still, to prevent complications, chest compressions should be performed correctly and in the accurate position rather than reducing compressive efforts [19]. With increasing rates of survival with CPR, potential traumatic injuries caused by chest compression should be considered. Accurate diagnosis and proper treatment of unavoidable CPR-related complications are important to decrease traumatic damages. No potential conflict of interest relevant to this article was reported.

Aortic dissection after CPR is rare but can be fatal. In the absence of previous evaluation, it can also be difficult to confirm a causal relationship between CPR-induced injury and the underlying disease that caused the sudden cardiac arrest. The 2010 American Heart Association (AHA) guidelines recommend harder (a depth of at least 5 cm) and faster compressions (a rate of at least 100 compressions per minute) than the 2005 guidelines [18]. Increasing the requirements of compression depth and rate may increase the risk of CPR-related complications, including life-threatening events. Still, to prevent complications, chest compressions should be performed correctly and in the accurate position rather than reducing compressive efforts [19]. With increasing rates of survival with CPR, potential traumatic injuries caused by chest compression should be considered. Accurate diagnosis and proper treatment of unavoidable CPR-related complications are important to decrease traumatic damages. No potential conflict of interest relevant to this article was reported. Figure 1. Anteroposterior chest radiography performed after cardiopulmonary resuscitation. (A) Initial chest radiography in the ICU shows diffuse pulmonary edema and subsegmental atelectasis in the left lower lung field. (B) Follow-up chest radiography performed after 6 hours reveals that pulmonary edema was improved due to the improvement in cardiac function. ICU: intensive care unit.

lmonary resuscitation. (A) Initial chest radiography in the ICU shows diffuse pulmonary edema and subsegmental atelectasis in the left lower lung field. (B) Follow-up chest radiography performed after 6 hours reveals that pulmonary edema was improved due to the improvement in cardiac function. ICU: intensive care unit. Figure 2. Aortic dissection after cardiopulmonary resuscitation. (A) Transverse contrast-enhanced computed tomography (CT) image shows dissection and saccular aneurysm of the proximal descending aorta and associated fracture of the left rib. (B) Transverse view at the abdominal level shows aortic dissection involvement of the celiac trunk; however, there is no malperfusion of the peripheral organs and periaortic hematoma. (C) Sagittal view of the contrast-enhanced CT image shows that the aortic dissection extends from the distal arch of the thoracic aorta to the proximal level of the renal artery.

Introduction High morbidity and mortality from severe sepsis and septic shock is still an unresolved challenging [1,2]. Sepsis-induced tissue hypoperfusion and organ dysfunction result from vasodilation, decreased systemic vascular resistance, and increased microvascular leakage [3]. Hypotension triggers a complex neurohumoral response during severe sepsis to maintain blood pressure homeostasis, resulting in increased sympathetic responses and hypothalamic–pituitary–adrenal axis stimulation [4]. The lack of this response in critically ill patients is related to a poor prognosis [5,6]. To preserving blood pressure homeostasis, the renin-angiotensin-aldosterone system (RAAS) is required. By loss of blood volume or a drop in blood pressure, RAAS induces hormonal activation sequentially with renin-angiotensin I-angiotensin II-aldosterone. This sophisticated system controls microvascular, blood pressure, and organ function [7], as well as inflammatory responses. The dissociation of plasma renin activity (PRA) and aldosterone production, namely hyperreninemic hypoaldosteronism, has been observed in approximately 20%–30% of critically ill patients, increasing the development of acute renal failure, other organ dysfunction, and death [8]. Plasma aldosterone concentrations (PACs) are useful biomarker to detect septic shock patients with a high risk of renal dysfunction [8].

ic hypoaldosteronism, has been observed in approximately 20%–30% of critically ill patients, increasing the development of acute renal failure, other organ dysfunction, and death [8]. Plasma aldosterone concentrations (PACs) are useful biomarker to detect septic shock patients with a high risk of renal dysfunction [8]. However, little is known regarding the usefulness of PRA and PAC measurements in prediction of mortality or relations with mortality in patients with septic shock. We hypothesized that prolonged septic shock increases PRA and PACs, resulting in increased inflammatory activity, organ failure, and mortality. In this study, we examined whether PRA and PAC measurement compared with conventional severity indicators are associated with 28-day mortality in patients with septic shock.

k. We hypothesized that prolonged septic shock increases PRA and PACs, resulting in increased inflammatory activity, organ failure, and mortality. In this study, we examined whether PRA and PAC measurement compared with conventional severity indicators are associated with 28-day mortality in patients with septic shock. Materials and Methods 1) Study design and setting This study was a prospective blood sampling and retrospective data analysis of a cohort of patients admitted to the medical intensive care unit (ICU) from August 2008 to January 2009. We enrolled 140 patients aged older than 19 years who had been admitted to the ICU with septic shock from August 2008 to January 2009. All of the patients met the following criteria for septic shock, as established by the International Sepsis Definitions Conference [9]: (1) clinical diagnosis of septic shock made within the past 48 hours with a documented bacterial infection and (2) at least two of the following: fever (body temperature, >38°C) or hypothermia (<36°C), heart rate >90 beats min-1, respiratory rate >20 breaths min-1, white blood cell count >12,000 cells mm-3 or <4,000 cells mm-3, or >10% immature (band) forms. Patients also had to meet the additional following criteria for septic shock: the presence of at least two signs of organ dysfunction (metabolic acidosis, arterial hypoxemia [PaO2/FiO2 ratio <250], oliguria [<30 ml kg-1 for 3 hours], intravascular disseminated coagulopathy, or an abrupt change in mental status); and persistent hypotension with systolic blood pressure of <90 mmHg for at least 1 hour, despite adequate fluid replacement (assessed by a central venous pressure >8 mmHg) and continuous administration of inotropic or vasopressor support [9].

travascular disseminated coagulopathy, or an abrupt change in mental status); and persistent hypotension with systolic blood pressure of <90 mmHg for at least 1 hour, despite adequate fluid replacement (assessed by a central venous pressure >8 mmHg) and continuous administration of inotropic or vasopressor support [9]. We excluded two patients without PAC or PRA measurements. We also excluded 33 patients who had previously taken angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (n = 12), who previously took aldosterone antagonists (n = 5), who had a history of steroid usage (n = 12), and who had previously used hemodialysis (n = 4). A total of 105 patients were included in the study. Demographic, clinical, and biochemical data at the time of ICU admission were recorded. Baseline characteristics are shown in Table 1 and Figure 1.

rone antagonists (n = 5), who had a history of steroid usage (n = 12), and who had previously used hemodialysis (n = 4). A total of 105 patients were included in the study. Demographic, clinical, and biochemical data at the time of ICU admission were recorded. Baseline characteristics are shown in Table 1 and Figure 1. 2) Data collection To assess the severity of illness and prognosis of the underlying disease, the Acute Physiology and Chronic Health Evaluation (APACHE) II score and Sequential Organ Failure Assessment (SOFA) score were evaluated at the initial time of ICU admission. The SOFA score was also recorded on days 1, 3, and 7. Data of the following parameters were also collected: age, sex, total fluid input and output, serum creatinine levels, arterial blood gas analysis, and serum electrolytes. Additionally, patients were assessed with respect to their need for mechanical ventilation, ventilator-free days, and/or renal replacement therapy. All of the patients were evaluated over their entire ICU stay, and ICU-free days and outcomes were recorded. The definitions of ventilator free-days and ICU free-days are as follows: Ventilator free-days = number of days from day 1 to day 28 on which a patient breathed without assistance (if the period of unassisted breathing lasted at least 48 consecutive hours) ICU free-days = number of days from day 1 to day 28 on which a survived patient is not in ICU

2) Data collection To assess the severity of illness and prognosis of the underlying disease, the Acute Physiology and Chronic Health Evaluation (APACHE) II score and Sequential Organ Failure Assessment (SOFA) score were evaluated at the initial time of ICU admission. The SOFA score was also recorded on days 1, 3, and 7. Data of the following parameters were also collected: age, sex, total fluid input and output, serum creatinine levels, arterial blood gas analysis, and serum electrolytes. Additionally, patients were assessed with respect to their need for mechanical ventilation, ventilator-free days, and/or renal replacement therapy. All of the patients were evaluated over their entire ICU stay, and ICU-free days and outcomes were recorded. The definitions of ventilator free-days and ICU free-days are as follows: Ventilator free-days = number of days from day 1 to day 28 on which a patient breathed without assistance (if the period of unassisted breathing lasted at least 48 consecutive hours) ICU free-days = number of days from day 1 to day 28 on which a survived patient is not in ICU 3) Plasma biomarker sampling Blood samples were collected with the patient in the supine position in the early morning. On the day of sampling, blood samples were immediately sent to a laboratory for measuring biomarkers, such as PACs (by radioimmunoassay; normal range, 29.9 to 158.8 ng dl-1; Cobra II gamma counter, Hewlett-Packard, Meriden City, KS, USA), PRA (by radioimmunoassay; normal range, 0.15 to 2.23 ng ml-1 h-1; Cobra II gamma counter, Hewlett-Packard), plasma cortisol levels (by comparative binding immunoenzymatic assay; normal range, 6.7 to 22.6 pg ml-1; UNICEL DXI 800, Beckman Coulter Inc., Brea, CA, USA), and C-reactive protein (CRP) levels (by immunoturbidimetry; normal range, 0 to 8.0 mg L-1; Hitachi 7600, Sekisui, Japan). Serial PACs, PRA, the PAC/PRA ratio, cortisol levels, and CRP levels were measured on days 1, 3, and 7 to analyze changes over time. Based on the previous study, the definition of hyperreninemic hypoaldosteronism was defined when PAC/PRA ratio was less than 2 [8].

range, 0 to 8.0 mg L-1; Hitachi 7600, Sekisui, Japan). Serial PACs, PRA, the PAC/PRA ratio, cortisol levels, and CRP levels were measured on days 1, 3, and 7 to analyze changes over time. Based on the previous study, the definition of hyperreninemic hypoaldosteronism was defined when PAC/PRA ratio was less than 2 [8]. 4) Statistical analyses Continuous variables are presented as the mean ± standard deviation or median (interquartile range) and categorical variables as number (percentage) of the total sample. Baseline characteristics of the groups were compared using the χ2 test or Fisher exact test for categorical variables. And nonnormally distributed variables were compared using the Mann-Whitney U-test, and normally distributed variables using Student t-test when analyzing continuous variables. Pearson correlation was performed to investigate the relationships between SOFA scores and PRA or PACs. We evaluated 28-day all-cause mortality as the study endpoint. We conducted receiver operating characteristic (ROC) analysis to compare the predictive accuracy of PRA or PACs and APACHE II scores or SOFA scores. The area under the curve (AUC) was calculated. The Youden index was used to determine the optimal cutoff value of PRA or PACs for predicting 28-day mortality. Survival curves were prepared using the Kaplan-Meier method, and comparisons were made using the log-rank test. Prognostic variables for mortality were analyzed using the univariate Cox proportional hazards model, and variables with P-values <0.1 were used in the multivariate Cox proportional hazard model with backward selection. Results of univariate and multivariate Cox regression analyses are presented as hazard ratios and 95% confidence intervals (CIs). Differences were considered statistically significant at P < 0.05 (two-tailed). Statistical analyses were performed using the SPSS software package version 18.0 (SPSS Inc., Chicago, IL, USA).

esults of univariate and multivariate Cox regression analyses are presented as hazard ratios and 95% confidence intervals (CIs). Differences were considered statistically significant at P < 0.05 (two-tailed). Statistical analyses were performed using the SPSS software package version 18.0 (SPSS Inc., Chicago, IL, USA). 5) Ethics statement The protocol was approved by our institutional review board (No. 4-2008-0099). Written informed consent was obtained from the patients or their next of kin.

esults of univariate and multivariate Cox regression analyses are presented as hazard ratios and 95% confidence intervals (CIs). Differences were considered statistically significant at P < 0.05 (two-tailed). Statistical analyses were performed using the SPSS software package version 18.0 (SPSS Inc., Chicago, IL, USA). 5) Ethics statement The protocol was approved by our institutional review board (No. 4-2008-0099). Written informed consent was obtained from the patients or their next of kin. Results 1) Baseline characteristics of the study population A total of 105 patients with septic shock were enrolled. There were 65 (61.9%) men and 40 (38.1%) women. The mean age was 63.6 years (range, 55 to 74 years). The 28-day mortality was 43.8% (n = 46) and these patients were classified as non-survivors. The remaining patients were classified as survivors. Table 1 shows the comparison of demographics and clinical parameters on day 1 between non-survivors and survivors. Age, sex, primary infection site, the rate of positive blood culture, the comorbidities are similar in both groups. However, the survivor group showed lower APACHE II scores (P = 0.001) and SOFA scores (P = 0.001), and a much more prevalent history of coronary arterial occlusive disease or cancer than did the non-survivor group. In terms of critical care needs, patients in the non-survivor group had less ventilator-free days (P < 0.001) and ICU-free days (P < 0.001), and required a higher number of steroid use and continuous renal replacement therapy applications within 7 days after admission than did those in the non-survivor group. The non-survivor group showed a higher heart rate (P = 0.036) and net fluid balance on day 1 (P = 0.022), but a lower PaO2/FiO2 ratio (P = 0.002), than did the survivor group. The survivor group showed significantly lower PRA (P = 0.003) and PACs (P = 0.008) than did the non-survivor group. However, there were no significant differences in other blood biomarkers, such as CRP levels, cortisol levels, and the PAC/PRA ratio, between the non-survivor and survivor groups (Tables 1 and 2).

group. The survivor group showed significantly lower PRA (P = 0.003) and PACs (P = 0.008) than did the non-survivor group. However, there were no significant differences in other blood biomarkers, such as CRP levels, cortisol levels, and the PAC/PRA ratio, between the non-survivor and survivor groups (Tables 1 and 2). 2) Time course of PRA and PACs during the first 7 ICU days in the non-survivor and survivor groups The time course of PRA in he survivors and non-survivors was as follows. PRA and PACs in the survivors were significantly lower than in the non-survivors on days 1, 3, and 7 (Figure 2, Supplementary Table 1). PACs were significantly lower in the survivor group than in the non-survivor group on days 1 (P = 0.008), 3 (P = 0.001), and 7 (P = 0.012). CRP levels, cortisol levels, and the PAC/PRA ratio were not different between the non-survivor and survivor groups on day 1, 3, and 7 (Supplementary Table 1). When serial PAC, PRA, cortisol, and CRP level were measured, PAC and PRA level were consistently lower in the surviving group, but CRP and cortisol level were not (Supplementary Table 1). 3) Correlations between PRA or PACs and the APACHE II score and SOFA score PRA was positively correlated with the APACHE II score (r = 0.230, P = 0.018) and the SOFA score (r = 0.373, P < 0.001) (Figure 3A and B). PACs were only positively correlated with the SOFA score (r = 0.316, P = 0.001) (Figure 3C and D). However, both biomarkers showed weak positive correlation.

II score and SOFA score PRA was positively correlated with the APACHE II score (r = 0.230, P = 0.018) and the SOFA score (r = 0.373, P < 0.001) (Figure 3A and B). PACs were only positively correlated with the SOFA score (r = 0.316, P = 0.001) (Figure 3C and D). However, both biomarkers showed weak positive correlation. 4) Risk analysis for 28-day mortality The areas under the ROC curve for PRA and PAC values on day 1 were 0.69 (95% CI, 0.58 to 0.79; P = 0.001) and 0.67 (95% CI, 0.56 to 0.77; P = 0.003), respectively. These values were similar to the APACHE II scores (AUC, 0.68; 95% CI, 0.58 to 0.78; P = 0.001) and SOFA scores (AUC, 0.67; 95% CI, 0.57 to 0.77; P = 0.003) (Figure 4). Therefore, based on the maximal value of the Youden index (J = 0.25), the most discriminatory markers were PRA with a cutoff of 3.5 ng ml-1 h-1 (sensitivity of 76.1% and specificity of 64.4%) and PAC with a cutoff of 112 ng dl-1 (sensitivity of 58.7% and specificity of 71.2%) for predicting 28-day mortality. Kaplan-Meier survival analysis showed that the 28-day mortality of patients with PRA values ≥3.5 ng ml-1 h-1 (n = 49) was higher than that of patients with PRA values <3.5 ng ml-1 h-1 (n = 56) (log-rank test, P < 0.001, Figure 5). Using Cox regression analysis, PRA and PACs, as well as several other clinical characteristics to predict 28-day mortality, were evaluated (Table 3). After adjusting for confounding factors, the SOFA score (hazard ratio, 1.11; 95% CI, 1.01 to 1.22), PRA values ≥3.5 ng ml-1 h-1 (hazard ratio, 3.25; 95% CI, 1.60 to 6.60), previous history of cancer (hazard ratio, 3.44; 95% CI, 1.72 to 6.90) and coronary arterial occlusive disease (hazard ratio, 2.99; 95% CI, 1.26 to 7.08) were independently associated with 28-day mortality.

hazard ratio, 1.11; 95% CI, 1.01 to 1.22), PRA values ≥3.5 ng ml-1 h-1 (hazard ratio, 3.25; 95% CI, 1.60 to 6.60), previous history of cancer (hazard ratio, 3.44; 95% CI, 1.72 to 6.90) and coronary arterial occlusive disease (hazard ratio, 2.99; 95% CI, 1.26 to 7.08) were independently associated with 28-day mortality. Discussion In the current study, we assessed clinical parameters, including serial serum PRA and PACs of septic shock patients on days 1, 3, and 7 after admission to ICU in a single-center cohort. The main findings of this study were as follows: (1) 28-day survivors showed lower PRA and PAC values than did 28-day non-survivors; (2) PRA and PACs were positively correlated with severity scores; and (3) in particular, patients with septic shock with PRA values ≥3.5 ng ml-1 h-1 on day 1 were related to 28-day mortality.

main findings of this study were as follows: (1) 28-day survivors showed lower PRA and PAC values than did 28-day non-survivors; (2) PRA and PACs were positively correlated with severity scores; and (3) in particular, patients with septic shock with PRA values ≥3.5 ng ml-1 h-1 on day 1 were related to 28-day mortality. APACHE II score or SOFA score are used as a good indicator of the severity stratification or prognosis predication of patients with sepsis in the ICU. In our study, PRA and PACs were positively correlated with the SOFA score on days 1, 3, and 7. Their prognostic accuracy was similar to that of SOFA scores for mortality. PRA values ≥3.5 ng ml-1h-1 on day 1 was identified as an independent factor associated with 28-day mortality, even after adjusting for other parameters and severity scores (APACHE II and SOFA score). Although a single biomarker cannot substitute a composite indicator of severity such as APACHE II score or SOFA score, our findings may help to understand the pathophysiology in septic shock. And these results could be explained as follows. Over the last decade, new components of the RAAS have been found to be involved in many essential pathophysiological processes, such as development, inflammation, and remodeling [10,11]. Recently, the relevance between the potentially proinflammatory effects of angiotensin II and the pathogenesis of sepsis and acute lung injury has been of interest [12-14]. Blocking of a pro-renin receptor attenuates an inflammatory response in an animal model of sepsis using rat [15]. Many serum markers in the inflammatory cascade have been assessed in sepsis as potential indicators of infection [16]. In particular, the RAAS plays important roles in controlling blood pressure, microvascular regulation, and organ function, as well as in inflammatory responses. In our study, there was a positive linear correlation between PRA or PACs and the organ failure score (SOFA score) in septic shock patients. We initially expected that elevated PRA or PACs may also contribute to a stronger inflammatory response and may be a similarly useful prognostic indicator as serial CRP measurements in prediction of infection [17] and in monitoring treatment responses [18-20]. However, there was no linear correlation between PRA or PACs and CRP levels. Further studies are required to determine the role of PRA and PACs in the inflammatory response using other inflammatory markers.

dicator as serial CRP measurements in prediction of infection [17] and in monitoring treatment responses [18-20]. However, there was no linear correlation between PRA or PACs and CRP levels. Further studies are required to determine the role of PRA and PACs in the inflammatory response using other inflammatory markers. This was the first study to elucidate the relationship between PRA and mortality in patients with septic shock. There have been no studies on the association between PRA and mortality in patients with septic shock. However, in animal experiments, there is evidence between RAAS activity and sepsis [11]. In some studies, hyperreninemic hypoaldosteronism has been identified as a potential prognostic indicator of septic shock-induced acute renal failure [8] and liver cirrhosis [21]. du Cheyron et al. [8] showed that hyperreninemic hypoaldosteronism might be related with acute kidney injury in patients with septic shock. In our study, 55.2% of patients with septic shock had hyperreninemic hypoaldosteronism. du Cheyron et al. [8] found hyperreninemic hypoaldosteronism in 48% of septic shock patients, which is similar to our study. However, we did not find any correlations between hyperreninemic hypoaldosteronism and any clinical outcome, such as renal failure, ventilator-free days, ICU-free days, and 28-day mortality (data not shown). Similar to our study, du Cheyron et al. [8] also found that patients with septic shock showed markedly elevated PRA levels, associated with volume depletion and catecholamine-induced stimulation of β-adrenergic receptors. In our study, the non-survivor group showed higher PRA and PACs than did the survivor group, despite a greater net fluid balance on day 1. This finding suggests that patients with higher PRA and PACs need much more fluid administration for intravascular volume depletion than expected. In our data, PRA was consistently higher in non-survival group (Figure 2A). Therefore, mortality in septic shock patients may be relevant with prolonged plasma renin activation not merely hyperreninemic hypoaldosteronism defined as a fixed cutoff (PAC/PRA <2).

istration for intravascular volume depletion than expected. In our data, PRA was consistently higher in non-survival group (Figure 2A). Therefore, mortality in septic shock patients may be relevant with prolonged plasma renin activation not merely hyperreninemic hypoaldosteronism defined as a fixed cutoff (PAC/PRA <2). The effects of the RAAS are no longer considered to be limited to the circulatory system. This is because system components are also produced and activated in situ in other tissues, including the brain, blood vessels, adrenals, adipose tissue, kidney, and heart, acting through paracrine and autocrine mechanisms [22]. In particular, angiotensin II can promote cell growth and inflammation by enhancing expression of endothelium-derived adhesion molecules [23] and production of pro-inflammatory cytokines and chemokines [24]. Angiotensin II also has procoagulant activity [25], and can stimulate reactive oxygen species production [26]. This results in increased microvascular hydraulic permeability [27] and capillary leakage in severe sepsis. Doerschug et al. [7] showed that PRA was correlated with plasma concentrations of angiotensin II, microvascular dysregulation, and organ failure in patients with septic shock. Accordingly, long-term RAAS over-activation may be harmful in patients with severe sepsis and septic shock, regardless of hyperreninemic hypoaldosteronism. In our serial follow-up data (day 1, day 3, and day 7), consistently high PRA levels were observed in non-survivor groups. And we found the relation between high PRA level and mortality in septic shock in Cox regression model. Therefore we support a negative effect on survival in septic shock patients by long-term RAAS over-activation. Hirano et al. [15] showed that a (pro) renin receptor blocker significantly improved survival in rats with clinically relevant sepsis induced through cecal ligation and puncture. However, further human data on RAAS activity in sepsis are required. In a practical situation, it may be important to identify the appropriate time and subject to receive a RAAS antagonist because of the beneficial role of angiotensin II in maintaining vascular tone during septic shock. Our study showed that renin activity ≥3.5 ng ml-1 h-1 was significantly correlated with 28-day mortality in patients with septic shock. This finding can be used as the basis for selection of (pro) renin blocker candidates for future clinical studies. Interestingly, Vergaro et al.

ning vascular tone during septic shock. Our study showed that renin activity ≥3.5 ng ml-1 h-1 was significantly correlated with 28-day mortality in patients with septic shock. This finding can be used as the basis for selection of (pro) renin blocker candidates for future clinical studies. Interestingly, Vergaro et al. [28] showed that a cutoff value for PRA ≥2.3 ng ml-1 h-1 was the best predictor of cardiac mortality in patients with heart failure, which is similar to our finding. In future studies, cutoff values for PRA should be investigated in various diseases.

ning vascular tone during septic shock. Our study showed that renin activity ≥3.5 ng ml-1 h-1 was significantly correlated with 28-day mortality in patients with septic shock. This finding can be used as the basis for selection of (pro) renin blocker candidates for future clinical studies. Interestingly, Vergaro et al. [28] showed that a cutoff value for PRA ≥2.3 ng ml-1 h-1 was the best predictor of cardiac mortality in patients with heart failure, which is similar to our finding. In future studies, cutoff values for PRA should be investigated in various diseases. We are aware of some limitations of our research. First, this small, single-center, retrospective cohort study should not be generalized. Large-scale, multicenter, prospective validation studies are required to confirm these findings. Although this was a retrospective cohort study, we attempted to remove confounding factors from the data, such as prior use of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, aldosterone antagonists, and steroids, as well as the effect of terminal renal failure. Second, we did not perform adrenocorticotropic hormone stimulation tests to assess the adequacy of adrenal function. This is a major limitation, even though an adrenocorticotropic hormone stimulation test is usually not recommended in patients with septic shock [29]. In addition, no data about thyroid function test can affect the results of our study. Third, we did not examine plasma angiotensin II concentrations because angiotensin II are not correlated with SOFA score in the later course of sepsis [7]. Additionally, a previous study showed that angiotensin II activity is only correlated with organ failure only during initial treatment. However, we wished to evaluate PRA and PACs over the course of septic shock. Fourth, in this study, the ROC AUC of APACHE II and SOFA score were relatively low because it may be due to the difference in the timing of beginning of septic shock and the time of admission to the ICU. And elevated PRA levels cannot be generalized to the entire sepsis patients because our study was aimed only at patients with septic shock. These points required careful interpretation. Finally, we only assessed CRP as an inflammatory marker with PRA and PACs. We did not measure blood lactate levels, procalcitonin, or prothrombin time for inflammation or coagulation system. Other biomarkers may show different results.

ed only at patients with septic shock. These points required careful interpretation. Finally, we only assessed CRP as an inflammatory marker with PRA and PACs. We did not measure blood lactate levels, procalcitonin, or prothrombin time for inflammation or coagulation system. Other biomarkers may show different results. In conclusion, based on our findings, elevated PRA might be a useful prognostic biomarker for risk stratification of patients with septic shock, and an indicator of 28-day mortality. Therefore, the results of our study support that PRA can be a potential therapeutic target. In the future, multicenter, large-scale studies on RAAS inhibition, such as PRA antagonists may be needed for septic shock treatment. No potential conflict of interest relevant to this article was reported. Supplementary Materials The online-only Supplement data are available with this article online: https://doi.org/10.4266/kjccm.2017.00094. Figure 1. Flow chart of inclusion and exclusion of patients in the study. A total of 140 patients were enrolled between August and December 2008, and 105 patients were included in the analysis. Patients without available renin and aldosterone measurements (n = 2), those using an angiotensinconverting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) (n = 12), those using an aldosterone antagonist (n = 5), those using steroids (≥15 mg/d) (n = 12), and those who had previous usual hemodialysis (n = 4) within 7 days were excluded. ICU: intensive care unit.

ments (n = 2), those using an angiotensinconverting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) (n = 12), those using an aldosterone antagonist (n = 5), those using steroids (≥15 mg/d) (n = 12), and those who had previous usual hemodialysis (n = 4) within 7 days were excluded. ICU: intensive care unit. Figure 2. Time courses of PRA and PACs in non-survivors versus survivors. Levels of PRA (A) and PACs (B) were measured on days 1, 3, and 7 after admission. The circles and bars indicate the mean values and SEM, respectively. Significant differences between nonsurvivors and survivors are indicated by an asterisk (*P < 0.05, survivors versus non-survivors on each day). PRA: plasma renin activity; PAC: plasma aldosterone concentration; SEM: standard error of the mean. Figure 3. Correlations between PRA or PAC, SOFA score, and APACHE II score on day 1. The PRA levels correlated with better with APACHE II score (A) or SOFA score (B) than the PAC levels (C, D) although there were weak correlations. The open circles indicate survivors, and the solid circles indicate non-survivors. APACHE: Acute Physiologic and Chronic Health Evaluation; PRA: plasma renin activity; SOFA: Sequential Organ Failure Assessment; PAC: plasma aldosterone concentration.

(B) than the PAC levels (C, D) although there were weak correlations. The open circles indicate survivors, and the solid circles indicate non-survivors. APACHE: Acute Physiologic and Chronic Health Evaluation; PRA: plasma renin activity; SOFA: Sequential Organ Failure Assessment; PAC: plasma aldosterone concentration. Figure 4. Receiver operating characteristic curves for PRA and PAC on day 1 for predicting 28-day mortality. The prognostic accuracy of PRA and PAC was not inferior to that of the APACHE II and SOFA scores. APACHE: Acute Physiologic and Chronic Health Evaluation; AUC: area under the curve; CI: confidence interval; SOFA: Sequential Organ Failure Assessment; PRA: plasma renin activity; PAC: plasma aldosterone concentration. Figure 5. Kaplan-Meier survival curves for 28 days according to PRA. The cutoff value was the optimal cutoff limit that predicted 28-day mortality (PRA <3.5 ng/ml/h or ≥3.5 ng/ml/h). PRA: plasma renin activity; ICU: intensive care unit. Table 1. Baseline characteristics according to 28-day mortality

Figure 4. Receiver operating characteristic curves for PRA and PAC on day 1 for predicting 28-day mortality. The prognostic accuracy of PRA and PAC was not inferior to that of the APACHE II and SOFA scores. APACHE: Acute Physiologic and Chronic Health Evaluation; AUC: area under the curve; CI: confidence interval; SOFA: Sequential Organ Failure Assessment; PRA: plasma renin activity; PAC: plasma aldosterone concentration. Figure 5. Kaplan-Meier survival curves for 28 days according to PRA. The cutoff value was the optimal cutoff limit that predicted 28-day mortality (PRA <3.5 ng/ml/h or ≥3.5 ng/ml/h). PRA: plasma renin activity; ICU: intensive care unit. Table 1. Baseline characteristics according to 28-day mortality Characteristic Non-survivors (n = 46 [43.8]) Survivors (n = 59 [56.2]) P-value Age (yr) 65.0 (54.0–73.3) 65.0 (56.0–74.0) 0.824 Male sex 28 (60.9) 37 (62.7) 1.000 APACHE II score 26.0 (23.8–31.0) 22.0 (18.0–28.0) 0.001 SOFA score 9.0 (7.0–12.3) 7.0 (4.0–10.0) 0.001 Comorbidity CAOD 7 (15.2) 2 (3.4) 0.040 Hypertension 13 (28.3) 28 (47.5) 0.069 Diabetes 10 (21.7) 21 (35.6) 0.137 Heart failure 6 (13.0) 6 (10.2) 0.760 Chronic lung disease 7 (15.2) 11 (18.6) 0.795 Chronic renal disease 6 (13.0) 6 (10.2) 0.760 Chronic liver disease 12 (26.1) 9 (15.3) 0.220 Cancer 35 (76.1) 24 (40.7) <0.001 Primary site of infection 0.814 Pulmonary 34 (73.9) 41 (69.5) Pancreato-biliary 1 (2.2) 5 (8.5) Liver 4 (8.7) 3 (5.1) GI tract 2 (4.3) 3 (5.1) Soft tissue and bone 1 (2.2) 1 (1.7) Brain and CSF 0 1 (1.7) Kidney 1 (2.2) 1 (1.7) Urogenital 1 (2.2) 3 (5.1) Miscellaneous 2 (4.3) 1 (1.7) Positive blood culture 17 (37.0) 18 (30.5) 0.535 Acute renal failure 0.489 At risk 11 (23.9) 15 (25.4) Injury 9 (19.6) 14 (23.7) Failure 3 (6.5) 8 (13.6) Chronic renal disease acute exacerbation, 6 (13.0) 6 (10.2) 0.760 Steroid use within 7 days 29 (63.0) 24 (40.7) 0.031 No. of CRRT applied within 7 days 18 (39.1) 10 (16.9) <0.001 Values are presented as number (%) or median (interquartile range).

sk 11 (23.9) 15 (25.4) Injury 9 (19.6) 14 (23.7) Failure 3 (6.5) 8 (13.6) Chronic renal disease acute exacerbation, 6 (13.0) 6 (10.2) 0.760 Steroid use within 7 days 29 (63.0) 24 (40.7) 0.031 No. of CRRT applied within 7 days 18 (39.1) 10 (16.9) <0.001 Values are presented as number (%) or median (interquartile range). APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; CAOD: coronary arterial occlusive disorder; CSF: cerebrospinal fluid; CRRT: continuous renal replacement therapy. Table 2. Clinical and biological characteristics according to 28-day mortality

sk 11 (23.9) 15 (25.4) Injury 9 (19.6) 14 (23.7) Failure 3 (6.5) 8 (13.6) Chronic renal disease acute exacerbation, 6 (13.0) 6 (10.2) 0.760 Steroid use within 7 days 29 (63.0) 24 (40.7) 0.031 No. of CRRT applied within 7 days 18 (39.1) 10 (16.9) <0.001 Values are presented as number (%) or median (interquartile range). APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; CAOD: coronary arterial occlusive disorder; CSF: cerebrospinal fluid; CRRT: continuous renal replacement therapy. Table 2. Clinical and biological characteristics according to 28-day mortality Variable Non-survivor (n = 46 [43.8]) Survivor (n = 59 [56.2]) P-value Mechanical ventilation used 43 (93.5) 49 (83.1) 0.140 Heart rate (beats min-1) 120.0 (90.5–136.5) 105.0 (81.0–126.0) 0.036 Mean arterial pressure (mmHg) 69.2 (62.8–92.6) 79.0 (65.7–99.3) 0.129 pH 7.4 (7.3–7.5) 7.4 (7.3–7.5) 0.395 PaCO2 (mmHg) 35.7 (29.5–41.0) 32.2 (27.4–36.6) 0.320 P/F ratio 197.7 (150.7–257.9) 264.6 (180.5–365.5) 0.002 PEEP (cmH2O) 10.0 (8.0–12.0) 10.0 (6.0–12.0) 0.679 Platelet (103/µl) 128.0 (52.5–227.8) 159 (76.0–314.0) 0.278 Serum potassium (mmol L-1) 3.8 (3.4–4.6) 3.7 (3.3–4.1) 0.301 Serum sodium (mmol L-1) 137.0 (134.8–144.0) 138.0 (134.0–142.0) 0.279 Blood urea nitrogen (mg dl-1) 24.7 (16.2–40.1) 25.6 (15.6–41.0) 0.853 Serum creatinine (µmol L-1) 1.3 (0.7–1.9) 1.2 (0.9–2.2) 0.760 Ejection fraction (%) 66.0 (54.0–70.0) 66.0 (60.0–70.0) 0.551 Net fluid balance at 1 days (L) 2.0 (0.5–3.4) 1.0 (0.2–2.3) 0.022 Blood Biomarkers CRP (mg L-1) 14.3 (5.8–28.7) 13.8 (8.3–23.3) 0.578 Cortisol (pg ml-1) 23.8 (15.2–46.0) 25.0 (15.8–45.3) 0.210 PRA (ng ml-1h-1) 8.7 (3.2–23.4) 2.4 (0.7–7.9) 0.003 PAC (ng dl-1) 125.9 (37.0–255.9) 53.8 (23.0–135.7) 0.008 PAC/PRA ratio 1.2 (0.6–3.0) 1.9 (0.7–5.9) 0.581 PAC/PRA <2 23 (50.0) 35 (59.3) 0.429 Values are presented as number (%) or median (interquartile range).

Cortisol (pg ml-1) 23.8 (15.2–46.0) 25.0 (15.8–45.3) 0.210 PRA (ng ml-1h-1) 8.7 (3.2–23.4) 2.4 (0.7–7.9) 0.003 PAC (ng dl-1) 125.9 (37.0–255.9) 53.8 (23.0–135.7) 0.008 PAC/PRA ratio 1.2 (0.6–3.0) 1.9 (0.7–5.9) 0.581 PAC/PRA <2 23 (50.0) 35 (59.3) 0.429 Values are presented as number (%) or median (interquartile range). P/F: PaO2/FIO2; PEEP: positive end-expiratory pressure; CRP: C-reactive protein; PRA: plasma renin activity; PAC: plasma aldosterone concentration. Table 3. Univariate and multivariate Cox model for 28-day mortality in patients with septic shock Covariable Univariate analysis Multivariate analysis HR 95% CI P-value HR 95% CI P-value APACHE II score 1.073 1.028–1.121 0.001 SOFA score 1.169 1.082–1.263 <0.001 1.112 1.011–1.224 0.029 Previous cancer history 3.140 1.591–6.197 0.001 3.442 1.717–6.899 <0.001 CAOD history 2.645 1.179–5.935 0.018 2.989 1.263–7.075 0.013 Steroid use within 7 days 1.945 1.068–3.542 0.030 CRRT within 7 days 2.290 1.265–4.147 0.006 1.769 0.924–3.385 0.085 Heart rate 1.011 1.000–1.022 0.048 Net fluid balance at day 1 1.000 1.000–1.000 0.011 PaO2/FIO2 ratio 0.998 0.995–1.001 0.130 PRA ≥3.5 ng/ml/h 3.943 1.996–7.790 <0.001 3.245 1.596–6.597 0.001 PAC ≥112 ng/dl 2.642 1.466–4.760 0.001 HR: hazard ratio; CI: confidence interval; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; CAOD: coronary artery obstructive disease; CRRT: continuous renal replacement therapy; PRA: plasma renin activity; PAC: plasma aldosterone concentration.

Dear Editor: A kind of herbs, aconite is known for cardiac toxicity [1,2]. Hemodynamic support using extracorporeal life support (ECLS) may be good method if failed conventional resuscitation. We report two experiences using ECLS in aconite intoxication. A 47-year-old man, who had taken 20 herbal tablets containing aconite, visited the emergency room because of chest discomfort. An initial electrocardiography (ECG) showed persistent multifocal ventricular tachycardia (Figure 1). He repeatedly became pulseless and unconscious. All conventional resuscitation methods including antiarrhythmic medicines, chest compression, and electric cardioversion failed to maintain a stable condition. After 10 minutes of resuscitation, extracorporeal membrane oxygenation (ECMO) was inserted immediately. A 15-F arterial and 22-F venous catheter were percutaneously inserted into the right femoral vessels. The initial flow rate was set at 2 L/min. Although ventricular tachyarrhythmia occurred frequently on the first hospital day, soon after, the vital signs were stabilized. ECG showed a normal sinus rhythm after 33 hours of ECLS. The ECLS was removed on hospital day 2. He was discharged on hospital day 10.

moral vessels. The initial flow rate was set at 2 L/min. Although ventricular tachyarrhythmia occurred frequently on the first hospital day, soon after, the vital signs were stabilized. ECG showed a normal sinus rhythm after 33 hours of ECLS. The ECLS was removed on hospital day 2. He was discharged on hospital day 10. Other case, a 31-year-old man, ingested an unknown number of tablets containing aconite and had difficulty in moving and chest discomfort, was referred to Samsung Medical Center. The initial ECG showed an irregular rhythm with a narrow QRS (Figure 2). Despite conventional resuscitation, the ventricular tachycardia was sustained. After a few minutes, a 15-F arterial and 22-F venous catheter were inserted. The patient’s vital signs were stabilized and ECG rhythm regained normal sinus rhythm after 9 hours of ECLS support. The ECLS was removed on hospital day 2. He was discharged on hospital day 7. Aconite induces refractory ventricular arrhythmia. The symptoms of poisoning appear 10 minutes to 3 hours after aconite is ingested [3,4]. When cardiogenic shock are refractory to medical treatment, it is most important to maintain blood pressure and tissue oxygenation by the use of a percutaneous cardiopulmonary bypass [5]. During the clearance of aconite from the body, ECMO may substitute heart function [6]. Our report shows that ECLS was an effective modality for repetitive life-threatening arrhythmia due to aconite poisoning. We believe that ECPR is a viable alternative to traditional cardiopulmonary resuscitation for patients with acute aconite intoxication.

aconite from the body, ECMO may substitute heart function [6]. Our report shows that ECLS was an effective modality for repetitive life-threatening arrhythmia due to aconite poisoning. We believe that ECPR is a viable alternative to traditional cardiopulmonary resuscitation for patients with acute aconite intoxication. No potential conflict of interest relevant to this article was reported. Figure 1. Initial and post-cardioversion electrocardiography: persistent multifocal ventricular tachycardia. Figure 2. Electrocardiography upon arrival at the emergency room: persistent multifocal ventricular tachycardia with right bundle branch block pattern.

Introduction Aging of the population is a significant issue worldwide, especially in developed countries. The Organization for Economic Cooperation and Development (OECD) report in 2006 showed that the number of people aged 65 years or older will double by the year 2050 [1]. Compared to neighboring East Asian countries, the aging rate of the population in Korea is rapid [2]. According to a report by the Korea Statistical Office, the proportion of people aged 65 years or above surpassed 11.7% of the whole population in the year 2015, and is expected to account for 40% of the national population in 2050. This increase in the percentage of elderly people is a consequence of low birth rate in the past 2 decades, increased life expectancy, and high birth rates during the 1950s and 1960s after the Korean War [3]. Along with an aging population, hospitalization for the elderly population will increase, and intensive care units (ICUs) will face increased admissions and demand by older patients. Elderly patients comprise a substantial proportion of ICU patients and are responsible for higher healthcare costs and longer ICU durations compared to relatively younger patients [4,5]. In Korea, hospital costs and duration of hospital stay also increase with patient age [3].

d admissions and demand by older patients. Elderly patients comprise a substantial proportion of ICU patients and are responsible for higher healthcare costs and longer ICU durations compared to relatively younger patients [4,5]. In Korea, hospital costs and duration of hospital stay also increase with patient age [3]. The percent of elderly ICU patients varies among nations, with values ranging from 3.0% [6] to 13.4% [7]. A study in France involving 75,000 ICU admissions reported that very elderly patients accounted for about 9.6% of the total admissions [8]. Many studies, mostly in developed nations, reported increased mean age of patients admitted to ICUs and increased percentage of ICU admissions attributable to elderly patients. Bagshaw et al. [9] analyzed ICU patients in Australia and New Zealand and showed that the proportion of ICU patients aged 80 years or more has increased, and that in-hospital mortality increased with patient age. Reinikainen et al. [10] reported that the proportion of elderly patients increased in a Finnish ICU, and that the hospital mortality rate increased with increasing age. Other large prospective studies have also reported that older age was associated with higher mortality [11-14].

l mortality increased with patient age. Reinikainen et al. [10] reported that the proportion of elderly patients increased in a Finnish ICU, and that the hospital mortality rate increased with increasing age. Other large prospective studies have also reported that older age was associated with higher mortality [11-14]. The demographic changes in ICU patients should also be studied, with a focus on resource utilization. Adhikari et al. [15] suggested that the burden of critical care increases as the population ages, and other studies have suggested that the aging population is strongly associated with more medical resources. A study by Teno et al. [16] showed that ICU use for elderly Medicare beneficiaries has increased during the last decade. To the best of our knowledge, no study evaluating the change in proportions of age groups of patients admitted to ICUs has been reported in Korea. The aim of this study is to evaluate the demographic changes of recent ICU patients in Korea, and to determine if clinical outcomes such as in-hospital mortality of ICU patients, length of ICU stay, and hospital stay have changed.

The demographic changes in ICU patients should also be studied, with a focus on resource utilization. Adhikari et al. [15] suggested that the burden of critical care increases as the population ages, and other studies have suggested that the aging population is strongly associated with more medical resources. A study by Teno et al. [16] showed that ICU use for elderly Medicare beneficiaries has increased during the last decade. To the best of our knowledge, no study evaluating the change in proportions of age groups of patients admitted to ICUs has been reported in Korea. The aim of this study is to evaluate the demographic changes of recent ICU patients in Korea, and to determine if clinical outcomes such as in-hospital mortality of ICU patients, length of ICU stay, and hospital stay have changed. Materials and Methods 1) Study design The present study is a retrospective analysis of electronic data of ICU patients from a single center. The patients admitted to either the medical or surgical ICU of St. Paul’s Hospital, The Catholic University of Korea, from January 2005 to December 2014 were analyzed. The medical and surgical ICUs have 13 and 11 beds, respectively, and are equipped with necessary equipment for the treatment and monitoring of patients on mechanical ventilation and/or vasopressor therapy. Patient age, sex, duration of hospital stay, duration of ICU stay, inhospital mortality, and readmission to ICU within the same hospital admission period were retrospectively collected. This study was reviewed and approved by the institutional review board of St. Paul’s Hospital, College of Medicine, The Catholic University of Korea (No. PC-16OISI0002).

tal stay, duration of ICU stay, inhospital mortality, and readmission to ICU within the same hospital admission period were retrospectively collected. This study was reviewed and approved by the institutional review board of St. Paul’s Hospital, College of Medicine, The Catholic University of Korea (No. PC-16OISI0002). 2) Outcome measures The proportion of total admissions of patients aged ≥65 and ≥80 years were described annually and cumulatively. The number of admissions of age groups of <50, 50–64, 65–79, and ≥80 years were compared annually. The mean age of the overall ICU patients was also evaluated each year. For clinical outcome measures, in-hospital mortality, length of stay in the hospital, and length of stay in the ICU were evaluated annually. The ICU readmission rate, defined as the number of patients readmitted to the ICU during the same hospital admission period, was assessed annually as well. 3) Inclusion criteria We extracted data from our hospital electronic medical record on ICU admissions attributable to patients aged at least 18 years. Patients were admitted for medical reasons, elective surgery, or emergency surgery between January 1, 2005 and December 31, 2014. We excluded the admissions to the ICU for observation after coronary angiography due to a sudden increase in coronary angiography cases after 2012 and possible bias on clinical outcomes of overall ICU patients (Figure 1). In case of multiple admissions by the same patient, admissions other than first admission were excluded.

luded the admissions to the ICU for observation after coronary angiography due to a sudden increase in coronary angiography cases after 2012 and possible bias on clinical outcomes of overall ICU patients (Figure 1). In case of multiple admissions by the same patient, admissions other than first admission were excluded. 4) Definition of elderly patients Upon analysis of past literature concerning the ICU and elderly patients [17,18], we defined elderly patients as those aged 65 years or older. Patients older than 80 years were defined as the very elderly according to previous studies [9,19]. 5) Statistical analysis Descriptive statistics included mean and standard deviation values except when stated otherwise. To see if there was a linear association between study year and age group, both of which are categorical variables, we performed the modified chi-square test and linear by linear association. When the P-value was less than the significance level, 0.05, linear association was counted as valid. Second, we analyzed if each demographic showed any increasing or decreasing trend with time, using a nonparametric method, Mann-Kendall trend analysis. To verify any differences in demographic or clinical data of the four age groups, we performed analysis of variance. Lastly, in order to evaluate the associations between decreasing inhospital mortality and vasopressors, intubation, and operations, we performed simple linear regression.

od, Mann-Kendall trend analysis. To verify any differences in demographic or clinical data of the four age groups, we performed analysis of variance. Lastly, in order to evaluate the associations between decreasing inhospital mortality and vasopressors, intubation, and operations, we performed simple linear regression. 6) Total cost per person Total costs per person during hospital admission were evaluated based on insurance data. All costs are presented in Korean won (KRW). 7) Diagnosis at admission to ICU The ICU patients with diagnoses associated with ICU mortality were classified into 10 subcategories. Diagnoses were sorted according to main 10th revision of the International Statistical Classification of Diseases (ICD-10) codes of the patients. Regardless of organs involved, patients were classified into the neoplasm category if they were coded with ICD-10 code related to malignant cancer. Patients with A418 or A419 ICD-10 codes were classified into the sepsis category. 8) Operation The operations undergone be patients admitted to the ICU included surgeries under general anesthesia, regional anesthesia, and local anesthesia. The patients who received operations were classified into five categories (abdominal, thoracic, neurosurgery, orthopedic, and others).

7) Diagnosis at admission to ICU The ICU patients with diagnoses associated with ICU mortality were classified into 10 subcategories. Diagnoses were sorted according to main 10th revision of the International Statistical Classification of Diseases (ICD-10) codes of the patients. Regardless of organs involved, patients were classified into the neoplasm category if they were coded with ICD-10 code related to malignant cancer. Patients with A418 or A419 ICD-10 codes were classified into the sepsis category. 8) Operation The operations undergone be patients admitted to the ICU included surgeries under general anesthesia, regional anesthesia, and local anesthesia. The patients who received operations were classified into five categories (abdominal, thoracic, neurosurgery, orthopedic, and others). 9) Treatment days attributable to elderly patients We obtained the total number of ICU treatment days and the number attributable to the elderly by summing the days of individual admissions by elderly patients. Percentages of total treatment days attributable to elderly patients were calculated as the sum of treatment days for all elderly patients divided by the sum of treatment days for patients of all age groups.

ays and the number attributable to the elderly by summing the days of individual admissions by elderly patients. Percentages of total treatment days attributable to elderly patients were calculated as the sum of treatment days for all elderly patients divided by the sum of treatment days for patients of all age groups. Results During the 10-year period, 10,366 ICU patients were included in the analysis set constructed from the electronic database of St. Paul’s Hospital. During the 10 years, the absolute number and the proportion of patients aged ≥65 years admitted to the ICU significantly increased annually. The proportion of patients aged 65 years or older increased from 47.9% in 2005 to 63.7% in 2014 (Table 1). The cumulative proportion of patients aged ≥80 years admitted during the study period was 15.4% (n = 1,601) (Table 1, Figure 2). 1) In-hospital mortality In-hospital mortality of overall patients was 12.1% in 2005, but decreased to 9.6% in 2014 (P = 0.004). For elderly patients, in-hospital mortality also significantly decreased from 16.7% to 10.6% (P = 0.0042) (Table 2). Over the 10 years, in-hospital mortality increased with age group; 6.9% in patients <50 years and 19.1% in patients aged ≥80 years [4,9]. The rate of readmission to the ICU in the same admission period showed a significant difference between age groups (P = 0.0298) (Table 3). 2) Diagnosis at admission The percentage of patients with respiratory diagnoses increased with age group. Of all diagnostic categories, the highest proportion of patients in each age groups had a neurologic diagnosis (Table 3).

1) In-hospital mortality In-hospital mortality of overall patients was 12.1% in 2005, but decreased to 9.6% in 2014 (P = 0.004). For elderly patients, in-hospital mortality also significantly decreased from 16.7% to 10.6% (P = 0.0042) (Table 2). Over the 10 years, in-hospital mortality increased with age group; 6.9% in patients <50 years and 19.1% in patients aged ≥80 years [4,9]. The rate of readmission to the ICU in the same admission period showed a significant difference between age groups (P = 0.0298) (Table 3). 2) Diagnosis at admission The percentage of patients with respiratory diagnoses increased with age group. Of all diagnostic categories, the highest proportion of patients in each age groups had a neurologic diagnosis (Table 3). 3) ICU treatment days attributable to elderly patients The percentage of ICU treatment days attributable to patients aged 65 years or more relative to total treatment days of all age groups increased from 51.1% in 2005 to 64.0% in 2014, and the percent of hospital treatment days increased from 49.1% in 2005 to 59.5% in 2012, followed by a gradual decrease in the next 2 years. For very elderly patients aged 80 years or over, the percent of ICU treatment days increased from 12.5% in 2005 to 24.66% in 2014, and the hospital treatment days increased from 13.5% to 22.0% (Table 2).

al treatment days increased from 49.1% in 2005 to 59.5% in 2012, followed by a gradual decrease in the next 2 years. For very elderly patients aged 80 years or over, the percent of ICU treatment days increased from 12.5% in 2005 to 24.66% in 2014, and the hospital treatment days increased from 13.5% to 22.0% (Table 2). 4) Dobutamine, norepinephrine, dopamine, and intubation For patients aged ≥80, 7.7% received dobutamine; 9.7% received norepinephrine; 27.7% received dopamine; and 27.7% were intubated. Compared to other younger age groups, the proportion of patients who received aggressive treatment was higher in the very elderly patients. Use of norepinephrine (P = 0.016) and the proportion of intubated patients (P = 0.01) showed significant differences between the age groups (Table 3). 5) Operations The proportion of patients who underwent operations was 27% in the patients aged <50 years and 38.2% for patients aged from 65 to 74 years. Nevertheless, the proportion of very elderly patients who received operation decreased to 30.5%. For patients aged 65 years or more, orthopedic surgeries were the second most frequently performed surgeries after abdominal surgeries (Table 3).

the patients aged <50 years and 38.2% for patients aged from 65 to 74 years. Nevertheless, the proportion of very elderly patients who received operation decreased to 30.5%. For patients aged 65 years or more, orthopedic surgeries were the second most frequently performed surgeries after abdominal surgeries (Table 3). 6) Total cost per person Table 4 shows that the total cost per person in overall patients, elderly patients, and very elderly patients increased annually. Medical expenses incurred during admission for elderly patients significantly increased from 3,078,981 KRW in 2005 to 5,552,785 KRW in 2014 (P < 0.001). For the very elderly patients, total cost per person was 1,499,765 KRW in 2005 and significantly increased to 2,103,833 KRW in 2014 (P = 0.0023). The association between increased medical expenses in elderly patients and increase in the number of operations among elderly patients was evaluated by linear regression and showed a significant correlation (P = 0.0113, R2 = 0.5726). Table 3 shows the total cost per person from each age group, which increased with age group (P = 0.001). 7) Factors that influence in-hospital mortality of elderly patients Vasopressor, intubation, and operation were evaluated for statistical association with change of in-hospital mortality in elderly patients. Use of at least one kind of vasopressors was not significantly associated with in-hospital mortality of patients aged 65 years or more (P = 0.799, R2 = 0.1085), nor was intubation or surgery (P = 0.4736, R2 = 0.1085 and P = 0.1365, R2 = 0.255, respectively.)

ion with change of in-hospital mortality in elderly patients. Use of at least one kind of vasopressors was not significantly associated with in-hospital mortality of patients aged 65 years or more (P = 0.799, R2 = 0.1085), nor was intubation or surgery (P = 0.4736, R2 = 0.1085 and P = 0.1365, R2 = 0.255, respectively.) Discussion In this study, we examined changes in the percentage of ICU admissions by elderly patients in the period 2005 to 2014 in two ICU units of a tertiary university hospital. Over the studied decade, the proportion of elderly patients and the mean age of overall ICU patients increased. To our knowledge, the present study is the first to evaluate changes in the demographic structure and concurrent changes in clinical outcomes of ICU patients in Korea. Furthermore, comparison of clinical outcomes and medical costs by different age groups in ICU has not been reported in Korea. The Korean population increased from 48,138,007 in 2005 to 50,423,955 in 2014, and the proportion of elderly citizens aged 65 years or older increased from 9.1% to 12.7%. The mean age of the Korean population was 35.5 years in 2005 and increased to 39.8 years in 2014 [2]. The aging of the Korean population was evident when compared to demographic data of other countries [2].

o 50,423,955 in 2014, and the proportion of elderly citizens aged 65 years or older increased from 9.1% to 12.7%. The mean age of the Korean population was 35.5 years in 2005 and increased to 39.8 years in 2014 [2]. The aging of the Korean population was evident when compared to demographic data of other countries [2]. Korea experienced rapid population expansion in the 1960s after the Korean War, and those currently in their 50–60s are usually called “baby boomers.” A similar population increase was observed in Europe after World War II [20]. However, the population policy enacted by the Korean government in the 1970s and the change in family value led to a decreased birth rate, thus resulting Table in an increasing percentage of elderly population. The present study is a single center study and cannot represent the entire Korean population; however, the number of patients is large enough to show both the demographic change over the last decade and clinical characteristics of elderly ICU patients. The ICU admissions by patients aged ≥80 accounted about 15.4% of overall ICU admissions during the study period. Compared to studies in other countries such as Australia and New Zealand [9], Finland [10], and Saudi Arabia [21], the percentage of ICU admissions attributable to very elderly patients was higher. However, the percentage was lower than the result of an Italian study by Pavoni et al. [22].

ions during the study period. Compared to studies in other countries such as Australia and New Zealand [9], Finland [10], and Saudi Arabia [21], the percentage of ICU admissions attributable to very elderly patients was higher. However, the percentage was lower than the result of an Italian study by Pavoni et al. [22]. With an increasing proportion of elderly patients, inhospital mortality was expected to increase [9,10]. Unlike our presumption, however, in-hospital mortality decreased during the study period. This result should be approached with consideration of the unique clinical setting. More patients received operations during 2010–2014 compared to earlier years, suggesting increased ICU admission for postoperative observations. Thus, in-hospital mortality, which was expected to increase with the larger proportion of elderly patients, has not increased. However, the decrease in in-hospital mortality of elderly patients was not significantly associated with an increase in operations among the elderly (P = 0.1365, R2 = 0.255). Nevertheless, in-hospital mortality showed a significant difference between age groups, increasing with age, consistent with previous results [9,11,23]. The very elderly group received the highest proportion of respiratory diagnoses at the time of admission to the ICU. Most of the respiratory diagnoses were acute respiratory distress syndrome or pneumonia, which showed higher proportions among the elderly patients in a previous study [24].

With an increasing proportion of elderly patients, inhospital mortality was expected to increase [9,10]. Unlike our presumption, however, in-hospital mortality decreased during the study period. This result should be approached with consideration of the unique clinical setting. More patients received operations during 2010–2014 compared to earlier years, suggesting increased ICU admission for postoperative observations. Thus, in-hospital mortality, which was expected to increase with the larger proportion of elderly patients, has not increased. However, the decrease in in-hospital mortality of elderly patients was not significantly associated with an increase in operations among the elderly (P = 0.1365, R2 = 0.255). Nevertheless, in-hospital mortality showed a significant difference between age groups, increasing with age, consistent with previous results [9,11,23]. The very elderly group received the highest proportion of respiratory diagnoses at the time of admission to the ICU. Most of the respiratory diagnoses were acute respiratory distress syndrome or pneumonia, which showed higher proportions among the elderly patients in a previous study [24]. Intensity of ICU treatment [25] and use of medical resources [15,26] are also significant issues. Of all treatment days for the four age groups combined, both the percentages of ICU and hospital treatments days attributable to the very elderly patients increased annually. A significant proportion of the elderly ICU patients received intensive ICU treatment such as inotropics, vasopressors, or intubation, and the proportion was relatively higher than in the younger age groups. Even though our data lacked clinical details including kind and severity of the diseases of the patients, the results showed that vasopressors and airway management were not spared for the elderly patients. The proportion of patients who underwent operation tended to increase with age strata; however, for the very elderly group, the surgery rate was lower compared to the younger elderly group.

the diseases of the patients, the results showed that vasopressors and airway management were not spared for the elderly patients. The proportion of patients who underwent operation tended to increase with age strata; however, for the very elderly group, the surgery rate was lower compared to the younger elderly group. In terms of total cost per person, elderly patients seem to incur lower medical expenses compared to younger age groups (Tables 3 and 4); however, the total cost per person in both elderly and very elderly patients gradually increased over the 10 years (Table 4). These results suggest that medical costs of elderly patients will increase in the future. An increase in the number of operations in elderly patients might have influenced the increased medical cost as the association between these two factors was significant (P = 0.0113, R2 = 0.5726). With relatively longer ICU stays and increasing total cost per person, our results suggest that the need for medical resources for elderly patients is increasing [15,27,28]. As the proportion of elderly patients increased over the last 10 years, it is likely that medical resources spent on overall ICU patients increased as well [16].

longer ICU stays and increasing total cost per person, our results suggest that the need for medical resources for elderly patients is increasing [15,27,28]. As the proportion of elderly patients increased over the last 10 years, it is likely that medical resources spent on overall ICU patients increased as well [16]. Our study has some limitations. First, it is a single-center retrospective study and does not represent the entire Korean ICU population; it is very likely that the demographic data were influenced by factors unique to our hospital. Coronary angiography cases increased more than two-fold after 2012 and resulted in increased ICU admissions for the sole purpose of observation after the procedure. For that reason, we excluded simple coronary angiography cases. Second, we excluded multiple admissions by same person, using only the first admission in analysis. In this process, admission for more severe illnesses could have been omitted. Third, disease severity scores were not analyzed. Our hospital adopted the electronic Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system prior to ICU admissions in the year 2012, so disease severity data were unreliable for retrospective analysis.

or more severe illnesses could have been omitted. Third, disease severity scores were not analyzed. Our hospital adopted the electronic Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system prior to ICU admissions in the year 2012, so disease severity data were unreliable for retrospective analysis. The present study is the largest single-center, retrospective analysis to show the increased mean age of ICU patients and the percentage attributable to elderly patients in Korea. Despite an increasing proportion of elderly patients admitted to ICUs, overall in-hospital mortality has not increased. We suggest that the rapidly aging national population had a considerable effect on the age structure in our ICU, and the increased proportion of elderly patients is expected to lead to more economic burdens in the ICU. In the future, multicenter studies are needed for a better representation of the Korean elderly ICU population and to evaluate the subtype of elderly patients who will benefit the most from intensive medical care. No potential conflict of interest relevant to this article was reported. Figure 1. Patients included in the study analysis. ICU: intensive care unit; CAG: coronary angiography. Figure 2. Graph showing annual changes in proportions of two age groups (65–79 and ≥80 yr). P-value indicates statistical difference between two age groups (P-value <0.05). Table 1. Annual percentages of the four different age groups of intensive care unit patients

Year <50 yr 50–64 yr 65–79 yr ≥80 yr 2005 248 (22.7) 320 (29.3) 383 (35.1) 140 (12.8) 2006 230 (19.9) 303 (26.2) 462 (40.0) 160 (13.9) 2007 223 (20.9) 297 (27.9) 407 (38.2) 138 (13.0) 2008 191 (18.6) 250 (24.4) 435 (42.4) 150 (14.6) 2009 173 (17.9) 241 (24.9) 397 (41.0) 158 (16.3) 2010 194 (17.4) 285 (25.5) 479 (42.8) 160 (14.3) 2011 165 (15.3) 293 (27.1) 451 (41.8) 171 (15.8) 2012 162 (15.1) 278 (25.9) 435 (40.6) 197 (18.4) 2013 111 (12.1) 277 (30.2) 382 (41.7) 146 (15.9) 2014 95 (10.9) 222 (25.4) 376 (43.0) 181 (20.7) P-value <0.0001a Value are presented as number (%). a Result of modified chi-square test for linear by linear association. Table 2. Patients characteristics by year Variable 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 P-valuea Patient 1,091 1,155 1,065 1,026 969 1,118 1,080 1,072 916 874 0.0491 Mean age (yr) 61.8 ± 15.6 63.3 ± 15.5 62.5 ± 15.4 64.1 ± 15.1 64.6 ± 15.0 64.3 ± 15.0 64.9 ± 15.3 65.8 ± 14.8 65.8 ± 14.4 67.7 ± 14.3 0.0014 Female 602 (55.2) 610 (52.8) 550 (51.6) 551 (53.7) 474 (48.9) 577 (51.6) 572 (53.0) 583 (54.4) 492 (53.7) 453 (51.8) 0.1524 ICU days attributable to very elderly 582 (12.51) 761 (17.79) 624 (14.21) 627 (17.79) 587 (15.53) 786 (19.46) 828 (18.64) 816 (19.25) 603 (18.85) 807 (24.66) 0.1524 Hospital days attributable to very elderly 2,707 (13.46) 3,019 (15) 2,661 (13.34) 2,147 (11.91) 2,552 (15.44) 2,714 (14.14) 2,970 (16.15) 3,788 (19.85) 2,812 (18.83) 3,189 (21.96) 0.1524 In-hospital mortality 132 (12.1) 152 (13.2) 133 (12.5) 141 (13.7) 111 (11.5) 130 (11.6) 123 (11.4) 121 (11.3) 87 (9.5) 84 (9.6) 0.0073 In-hospital mortality of elderly patients 64 (16.7) 77 (16.7) 53 (13.0) 61 (14.0) 43 (10.8) 55 (11.5) 51 (11.3) 46 (10.6) 43 (11.3) 40 (10.6) 0.0042 ICU type Medical 578 (53.0) 595 (51.5) 618 (58.0) 605 (59.0) 585 (60.4) 660 (59.0) 451 (41.8) 440 (41.0) 333 (36.4) 354 (40.5) 0.0736 Surgical 513 (47.0) 560 (48.5) 447 (42.0) 421 (41.0) 384 (39.6) 458 (41.0) 629 (58.2) 632 (59.0) 583 (63.6) 520 (59.5) 0.3711 ICU readmission rate 106 (9.7) 109 (9.4) 84 (7.9) 91 (8.9) 85 (8.8) 101 (9.0) 80 (7.4) 74 (6.9) 57 (6.2) 65 (7.4) 0.0073 Proportion of surgical patients 150 (13.7) 267 (23.1) 270 (25.4) 359 (35.0) 356 (36.7) 482 (43.1) 488 (45.2) 441 (41.1) 373 (40.7) 326 (37.3) 0.0035 Value are presented as number, mean ± standard deviation, or number (%).

06 (9.7) 109 (9.4) 84 (7.9) 91 (8.9) 85 (8.8) 101 (9.0) 80 (7.4) 74 (6.9) 57 (6.2) 65 (7.4) 0.0073 Proportion of surgical patients 150 (13.7) 267 (23.1) 270 (25.4) 359 (35.0) 356 (36.7) 482 (43.1) 488 (45.2) 441 (41.1) 373 (40.7) 326 (37.3) 0.0035 Value are presented as number, mean ± standard deviation, or number (%). ICU: intensive care unit. a Result of Mann-Kendall trend analysis. Table 3. Patients characteristics and clinical outcomes by age group

06 (9.7) 109 (9.4) 84 (7.9) 91 (8.9) 85 (8.8) 101 (9.0) 80 (7.4) 74 (6.9) 57 (6.2) 65 (7.4) 0.0073 Proportion of surgical patients 150 (13.7) 267 (23.1) 270 (25.4) 359 (35.0) 356 (36.7) 482 (43.1) 488 (45.2) 441 (41.1) 373 (40.7) 326 (37.3) 0.0035 Value are presented as number, mean ± standard deviation, or number (%). ICU: intensive care unit. a Result of Mann-Kendall trend analysis. Table 3. Patients characteristics and clinical outcomes by age group Variable <50 yr (n = 1,792) 50–64 yr (n = 2,766) 65–79 yr (n = 4,207) ≥80 yr (n = 1,601) P-value Male 1,097 (61.2) 1,716 (62.0) 2,107 (50.1) 544 (34.0) <0.0001a ICU type Medical 953 (53.2) 1349 (48.8) 2037 (48.4) 880 (55.0) 0.9372a Surgical 839 (46.8) 1417 (51.2) 2170 (51.6) 721 (45.0) 0.823a Age (yr) 39.7 ± 7.9 57.4 ± 4.4 71.8 ± 4.1 84.6 ± 3.9 - Median duration of ICU stay (d) 1 ± 5.7 2 ± 6.5 2 ± 10.4 2 ± 7.2 <0.0001b Median duration of hospital stay (d) 9 ± 25.5 12 ± 18.8 15 ± 18.2 14 ± 18.0 <0.0001b In-hospital mortality 124 (6.9) 251 (9.1) 533 (12.7) 306 (19.1) <0.0001a Intubation 181 (10.1) 348 (12.6) 662 (15.7) 316 (19.7) <0.0001a Norepinephrine 75 (4.2) 171 (6.2) 300 (7.1) 156 (9.7) <0.0001a Dopamine 330 (18.4) 613 (22.2) 1106 (26.3) 444 (27.7) <0.0001a Dobutamine 106 (5.9) 188 (6.8) 305 (7.2) 123 (7.7) 0.0188a ICU readmission rate 138 (7.7) 227 (8.2) 329 (7.8) 158 (9.9) 0.0298a Total cost per person (KRW) 18,213,069 ± 9,498,514 9,098,337 ± 2,253,776 4,324,299 ± 1,815,562 1,632,010 ± 994,939 <0.001a Diagnosis at ICU admission 0.001a Cardiovascular 156 (8.7) 266 (9.6) 368 (8.7) 152 (9.5) Respiratory 96 (5.4) 161 (5.8) 367 (8.7) 205 (12.8) Gastrointestinal 167 (9.3) 203 (7.3) 246 (5.8) 101 (6.3) Hepatobiliary 171 (9.5) 230 (8.3) 207 (4.9) 83 (5.2) Renal 37 (2.1) 74 (2.7) 117 (2.8) 49 (3.1) Neurologic 253 (14.1) 511 (18.5) 636 (15.1) 207 (12.9) Neoplasm 142 (7.9) 391 (14.1) 551 (13.1) 111 (6.9) Sepsis 7 (0.4) 14 (0.5) 33 (0.8) 14 (0.9) Metabolic 49 (2.7) 92 (3.3) 132 (3.1) 36 (2.2) Operation 483 (27.0) 935 (33.8) 1,605 (38.2) 489 (30.5) <0.001a Abdominal 159 (32.9) 314 (33.5) 471 (29.3) 111 (22.7) Thoracic 54 (11.2) 85 (9.1) 62 (3.9) 14 (2.9) Neurosurgery 120 (24.8) 259 (27.7) 218 (13.6) 50 (10.2) Orthopedic 53 (11.0) 177 (18.9) 666 (41.5) 269 (55.1) Others 17 (3.5) 20 (2.1) 39 (2.4) 7 (1.4) Value are presented as number (%) or mean ± standard deviation.

001a Abdominal 159 (32.9) 314 (33.5) 471 (29.3) 111 (22.7) Thoracic 54 (11.2) 85 (9.1) 62 (3.9) 14 (2.9) Neurosurgery 120 (24.8) 259 (27.7) 218 (13.6) 50 (10.2) Orthopedic 53 (11.0) 177 (18.9) 666 (41.5) 269 (55.1) Others 17 (3.5) 20 (2.1) 39 (2.4) 7 (1.4) Value are presented as number (%) or mean ± standard deviation. ICU: intensive care unit; KRW, Korean won. a Result of analysis of variance; b Result of Kruskal-Wallis test. Table 4. Annual change in medical costs expended Patient 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 P-valuea Overall 6,901,938 ± 5750344 6,818,923 ± 6233102 7,452,438 ± 6411112 7,209,678 ± 6388350 7,283,513 ± 6172670 7,839,203 ± 7088482 8,198,807 ± 8023929 7,921,556 ± 7037811 8,345,987 ± 7296308 8,079,804 ± 8811788 0.0042 Elderly 3,078,981 ± 1432996 3,324,927 ± 1444655 3,669,665 ± 1578033 4,030,537 ± 1768428 4,083,576 ± 1484201 4,479,332 ± 1674485 4,708,820 ± 1695110 5,040,944 ± 1693953 5,390,068 ± 1757865 5,552,785 ± 1809353 <0.001 Aged ≥80 yr 1,499,765 ± 1432996 1,202,370 ± 1444655 1,347,979 ± 1578033 1,422,915 ± 1768428 1,517,404 ± 1484201 1,628,645 ± 1674485 1,705,852 ± 1695110 1,816,514 ± 1693953 1,914,464 ± 1757865 2,103,833 ± 1809353 0.0023 Value are presented as mean ± standard deviation. All the costs are indicated in Korean won. a Result of simple linear regression.

Introduction Lung hyperinflation lowers the efficiency of the respiratory muscles by expanding the thorax and moves the diaphragm downwards, thus aggravating respiratory muscle fatigue [1]. In obstructive lung disease, the positive endexpiratory pressure (PEEP) is increased by airway obstruction during expiration, which leads to lung hyperinflation [2,3]. However, the mechanism of unilateral lung hyperinflation is more complex than bilateral lung hyperinflation. Unilateral lung hyperinflation is mainly caused by dynamic hyperinflation which is influenced by the lung compliance, tidal volume, airway resistance, and respiratory rate [1,4]. Asymmetric lung compliance aggravates unilateral lung hyperinflation. It may lead to mediastinal shift, giant bullae rupture, and acute vital instability. A typical example is native lung hyperinflation, which is one of the major complications associated with unilateral lung transplantation [5]. In addition, there have been reports of unilateral lung hyperinflation during mechanical ventilation of patients with chronic obstructive pulmonary disease (COPD), who had asymmetrical lung compliance due to fibrosis or pneumonia of a single lung [6,7]. Here, changes in ventilatory distribution caused by asymmetric lung compliance accelerated unilateral lung hyperinflation particularly in patients who had airflow obstruction [6]. To the best of our knowledge, little is known about the respiratory dynamics during mechanical ventilation of lungs with asymmetric compliance.

e, changes in ventilatory distribution caused by asymmetric lung compliance accelerated unilateral lung hyperinflation particularly in patients who had airflow obstruction [6]. To the best of our knowledge, little is known about the respiratory dynamics during mechanical ventilation of lungs with asymmetric compliance. In the present study, we aimed to investigate the distribution of ventilatory volume during inspiration and the influence of airway internal diameter (ID) change, by measuring respiratory variables in an asymmetric lung compliance model.

e, changes in ventilatory distribution caused by asymmetric lung compliance accelerated unilateral lung hyperinflation particularly in patients who had airflow obstruction [6]. To the best of our knowledge, little is known about the respiratory dynamics during mechanical ventilation of lungs with asymmetric compliance. In the present study, we aimed to investigate the distribution of ventilatory volume during inspiration and the influence of airway internal diameter (ID) change, by measuring respiratory variables in an asymmetric lung compliance model. Materials and Methods 1) Lung model The lung model was set to simulate a patient mechanically ventilated under general anesthesia. An air filter, endotracheal tube with an ID of 8 mm, connector with a variable diameter, Y-type breathing circuit, and two test lungs (Test lung 190; Maquet, Rastatt, Germany) were serially connected (Figure 1). The connector was placed to represent variable grades of distal airway obstruction. To implement asymmetrical lung compliance, one of the test lungs (lung1, L1) was manipulated to represent a normal lung (C60, static compliance 60 ml/cmH2O), a lung with low compliance (C15, static compliance 15 ml/cmH2O) or high compliance (C120, static compliance 120 ml/cmH2O). The compliance of the other test lung (lung2, L2) was set to that of a normal lung. In total, three lung models were constructed. To measure the respiratory parameters of total lung and L1, two spirometers (D-LiteTM; Datex-Ohmeda, Madison, WI, USA) were set between the breathing circuit and air filter (proximal measurement) and between the Y-type breathing circuit and L1 (distal measurement), respectively. The method described by Park et al. [8] was used to introduce compliance changes to L1. At first, the cover plates of both test lungs were manipulated to establish a static compliance at 15 ml/cmH2O, 60 ml/cmH2O, and 120 ml/cmH2O. As a result, both plastic cover plates of the test lungs were either opened or fixed to establish a static compliance of 120 or 60 cmH2O, respectively, and 10 rubber bands were wound around the center of the fixed plastic cover plates to establish a static compliance of 15 cmH2O. Then, these conditions were applied to L1 of each group. Before measuring respiratory mechanics, proximally placed spirometer was used to observe static compliance. During the experiment, preset compliance changes were applied to L1. First, tidal volume of total lung and volumes distributed to each lung were measured proximally and distally, respectively. Second, other respiratory variables such as PEEP, plateau pressure time (P1-P2 time), and flow were measured proximally and distally.

the experiment, preset compliance changes were applied to L1. First, tidal volume of total lung and volumes distributed to each lung were measured proximally and distally, respectively. Second, other respiratory variables such as PEEP, plateau pressure time (P1-P2 time), and flow were measured proximally and distally. 2) Intervention The lung model was ventilated using a mechanical ventilator (Dräger Fabius GS, Lübeck, Germany). Mechanical ventilation was performed using the volume-controlled mode with an inspiration to expiration ratio 1:2, an oxygen flow of 4 L/min, a tidal volume of 600 ml, a respiratory rate of 10 “breaths” per minute, and a maximum inspiratory pressure <80 cmH2O. To observe the inspiratory plateau pressure, the inspiratory pause was set as 50% of the inspiration time (1 second). To measure the effect of airway ID change on the distribution of tidal volume during mechanical ventilation, the ID of the connector was decreased by 1-mm intervals from 8 to 3 mm. Every time the ID was changed, the new conditions were maintained for at least 1 min (10 cycles) before and 3 minnutes (30 cycles) during data collection. A total of 36 conditions (three lung models, proximal and distal measurement, six IDs) were measured per cycle, and the mean values of 30 measurements at each condition (30 cycles) were calculated. Three additional cycles were measured with a spirometer distally placed at L2 to calculate the compliance ratio between the two lungs. The compliance ratio was calculated by dividing the mean values of L1 compliance by L2 compliance.

and the mean values of 30 measurements at each condition (30 cycles) were calculated. Three additional cycles were measured with a spirometer distally placed at L2 to calculate the compliance ratio between the two lungs. The compliance ratio was calculated by dividing the mean values of L1 compliance by L2 compliance. 3) Data collection A data collection computer program (S5 collect®; Datex-Ohmeda Co., Helsinki, Finland) was used to obtain physical variables at a frequency of 25 Hz. Respiratory dynamics were analyzed based on this collected data. Flow, volume, and pressure were recorded as values measured by the spirometer, and static compliance was derived using the following equation: (equation1) Cstatic=VTPplateau-PEEP Cstatic, static compliance, VT; tidal volume; Pplateau, inspiratory plateau pressure; and PEEP, positive end-expiratory pressure. The time from the start (P1) to the end (P2) of the inspiratory plateau phase (P1-P2) was calculated to evaluate the flow during the inspiratory pause. 4) Statistical analysis Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). Differences between and within groups were analyzed using two-way and one-way analysis of variance (ANOVA), respectively. Post-hoc analyses were performed if the ANOVA results showed statistically significant differences. Depending on the results of Levene’s test, a Bonferroni correction for equal variance or a Tamhane’s T2 for unequal variance was used as the post-hoc test. The threshold for statistical significance in all tests was set at P < 0.05.

lyses were performed if the ANOVA results showed statistically significant differences. Depending on the results of Levene’s test, a Bonferroni correction for equal variance or a Tamhane’s T2 for unequal variance was used as the post-hoc test. The threshold for statistical significance in all tests was set at P < 0.05. Results The ratio of the volume distributed to the two lungs (VL1/VL2) was proportionate to the ratio of lung compliance (CL1/CL2) in all groups (Table 1). VL1/VL2 and CL1/CL2 were similar among the different airway IDs in C15 and C60 groups. However, in C120 group, VL1/VL2 (P < 0.001) and CL1/CL2 (P < 0.001) were significantly reduced at airway ID of 3 and 4 mm when compared with airway ID 8 mm. The total tidal volume (volume proximally measured) and volume distributed to L1 is summarized in Table 2. The distally measured (P1-P2) times were significantly shorter than those measured proximally in C15 and C60 group, except for an airway ID of 3 mm (Figure 2). The proximally measured (P1-P2) times were comparable between groups (Figure 3A) and the distally measured (P1-P2) time durations were negatively correlated to lung compliance (Figure 3B).

) times were significantly shorter than those measured proximally in C15 and C60 group, except for an airway ID of 3 mm (Figure 2). The proximally measured (P1-P2) times were comparable between groups (Figure 3A) and the distally measured (P1-P2) time durations were negatively correlated to lung compliance (Figure 3B). In the C120 group, a PEEP > 1 cmH2O was observed for all IDs. PEEP was significantly increased for an ID ≤ 5 mm (cross sectional area ≤ 19.6 mm2) compared with an ID of 8 mm in the C120 (P < 0.001) and C60 (P < 0.001). In the C15 group, a statistically significant increase of PEEP occurred in ID 3 mm (cross sectional area = 7.1 mm2), compared to ID 8 mm (P < 0.001). When the distal measurements of PEEP values were compared between lungs, C15 and C60 groups were comparable, while C120 group was significantly higher than that in the other models (Figure 4). Flow-time curve of distal measurement in the three groups are shown in Figure 5. The inspiratory flow of L1 in C15 and C60 groups was maintained ≤ 5 and ≤ 10 ml/sec, respectively. This value was between 10 and 20 ml/sec in the C120 group, showing the larger volume distributed to L1 compared to other groups. Flow limitation was observed in smaller airway IDs, which was most significant in C120 group.

he inspiratory flow of L1 in C15 and C60 groups was maintained ≤ 5 and ≤ 10 ml/sec, respectively. This value was between 10 and 20 ml/sec in the C120 group, showing the larger volume distributed to L1 compared to other groups. Flow limitation was observed in smaller airway IDs, which was most significant in C120 group. Discussion Without airway obstruction, the tidal volume was distributed to each lung according to its compliance (Table 1). However, in the C120 group, although the total tidal volume was achieved, the volume distributed to L1 was significantly decreased at ID 3 and 4 mm, compared to ID 8 mm (Table 2). In visual inspection, a significant reduction of flow was observed in airway ID 3 and 4 mm in C120 group (Figure 5). While other groups did not show a significant reduction in the tidal volume distributed to L1, the distal measurement in C120 did show indications of statistically significant decreases at airway ID values of 3 and 4 mm (Table 2). The proportion of tidal volume distributed to L1 in the C120 group was reduced at airway ID 3 and 4 mm. This signifies amelioration of the asymmetry of ventilatory distribution, which can be explained by a few factors.

show indications of statistically significant decreases at airway ID values of 3 and 4 mm (Table 2). The proportion of tidal volume distributed to L1 in the C120 group was reduced at airway ID 3 and 4 mm. This signifies amelioration of the asymmetry of ventilatory distribution, which can be explained by a few factors. First, differences in the expansion rates of both lungs can be inferred by the differences of the time constants according to their compliance. The time constant is a product of the compliance and resistance and is therefore proportional to both compliance and resistance. Assuming that the resistances of both lungs are similar, the time constants of both lungs are proportional to their compliance. Having a small time constant means increased speed and pressure since the equilibrium volume is reached more quickly, and flow to that compartment will stop earlier. Meanwhile, having a larger time constant means that the time to reach the equilibrium volume is longer, and flow to that compartment will persist longer until equilibrium volume is reached [9-11]. Differences in time constants according to the differences in compliance can be confirmed by the (P1-P2) time results (Figure 3B). When distally measured, (P1-P2) times were significantly longer than proximally measured values in C15 and C60 groups; however, distally measure (P1-P2) times of C120 group were comparable with proximally measured values (Figure 2). From this finding, it can be inferred that the flow at L1 stops earlier than proximal measurement in the C15 and C60 groups. But in the C120 group, the flow at L1 persists until the flow at the proximal measurement stops, indicating that L1 in this group may not have enough time to reach the equilibrium volume distributed according to the compliance.

erred that the flow at L1 stops earlier than proximal measurement in the C15 and C60 groups. But in the C120 group, the flow at L1 persists until the flow at the proximal measurement stops, indicating that L1 in this group may not have enough time to reach the equilibrium volume distributed according to the compliance. Second, a decrease in the airway ID causes a limitation of flow and increases the time required for the lungs to reach their equilibrium volume. When the flow rate decreases, the time required to fill a certain lung volume is increased, which makes it more difficult for the lung with the larger time constant (i.e., the lung with higher compliance if the airway resistance is the same) to fill the same volume. The flow-time curve of the three distally measured groups (Figure 5) shows that at an airway ID was equal to or less than 4 mm, all groups failed to reach the tidal volume within the inspiration period.

r time constant (i.e., the lung with higher compliance if the airway resistance is the same) to fill the same volume. The flow-time curve of the three distally measured groups (Figure 5) shows that at an airway ID was equal to or less than 4 mm, all groups failed to reach the tidal volume within the inspiration period. It is widely known that reducing the tidal volume and respiration rate ameliorates lung hyperinflation, and these variables can be controlled by mechanical ventilators [12-14]. This is true even in the asymmetrical lung compliance cases in COPD patients reported by Kollef and Turner [6]. However, little is known about how airway diameter changes, such as those in airway obstruction, affect unilateral lung hyperinflation. Anglès et al. [15] reported that after unilateral lung transplantation, the incidence of hyperinflation in the remaining native lung reached 64% during hospitalization. Lung hyperinflation led to longer ICU stays and higher mortality rates (67% vs. 20%) compared with patients without lung hyperinflation.

yperinflation. Anglès et al. [15] reported that after unilateral lung transplantation, the incidence of hyperinflation in the remaining native lung reached 64% during hospitalization. Lung hyperinflation led to longer ICU stays and higher mortality rates (67% vs. 20%) compared with patients without lung hyperinflation. The results of the present study can be applied to patients under these conditions. However, improving the ventilatory distribution by reducing the airway ID ≤ 4 mm is not a desirable treatment for the mechanical ventilation of COPD patients with asymmetrical lung compliance. Because PEEP was significantly increased for an airway ID ≤ 5 mm, reducing the airway ID ≤ 4 mm may further exacerbate auto-PEEP. Clinical applications to limit volume redistribution in lungs with asymmetric compliance should focus on appropriate flow limitations that can maximize the time constant difference between the two lungs or on reducing the inspiratory pause time. The appropriate inspiratory flow rate for a patient may vary depending on the ratio of inspiration to expiration, respiratory rate, appropriate tidal volume, and lung compliance. As such, additional clinical studies are required to elucidate the appropriate values of these variables.

educing the inspiratory pause time. The appropriate inspiratory flow rate for a patient may vary depending on the ratio of inspiration to expiration, respiratory rate, appropriate tidal volume, and lung compliance. As such, additional clinical studies are required to elucidate the appropriate values of these variables. The limitation of our experimental model is that the model does not reflect the airway obstruction that occurs during expiration in obstructive lung disease. As such, our results should be limited to demonstrating the mechanism of unilateral lung hyperinflation during inspiration. Moreover, although the effects of respiratory muscles were not taken into account by assuming mechanical ventilation under general anesthesia, the interactions that may occur between both lungs due to the limited space of the thoracic cavity were not considered. Therefore, the measured values in the present study should be regarded as relative values showing the relationship between respiratory mechanical variables rather than absolute values. On the other hand, the fact that the respiratory variables were measured separately in unilateral lungs is a significant advantage of the present study as compared with animal studies or clinical trials. In conclusion, we were able to show that the distribution of volume during inspiration was positively correlated to lung compliance. The uneven distribution of volume might be reduced by changing the airway diameter equal to or less than 4 mm (cross sectional area, 12.6 mm2); however, this may exacerbate auto-PEEP.

The limitation of our experimental model is that the model does not reflect the airway obstruction that occurs during expiration in obstructive lung disease. As such, our results should be limited to demonstrating the mechanism of unilateral lung hyperinflation during inspiration. Moreover, although the effects of respiratory muscles were not taken into account by assuming mechanical ventilation under general anesthesia, the interactions that may occur between both lungs due to the limited space of the thoracic cavity were not considered. Therefore, the measured values in the present study should be regarded as relative values showing the relationship between respiratory mechanical variables rather than absolute values. On the other hand, the fact that the respiratory variables were measured separately in unilateral lungs is a significant advantage of the present study as compared with animal studies or clinical trials. In conclusion, we were able to show that the distribution of volume during inspiration was positively correlated to lung compliance. The uneven distribution of volume might be reduced by changing the airway diameter equal to or less than 4 mm (cross sectional area, 12.6 mm2); however, this may exacerbate auto-PEEP. No potential conflict of interest relevant to this article was reported. This study was presented at the Critical Care Symposium of the World Federation of Societies of Intensive and Critical Care Medicine (WFSICCM) on August 31, 2015 in Seoul, Korea.

In conclusion, we were able to show that the distribution of volume during inspiration was positively correlated to lung compliance. The uneven distribution of volume might be reduced by changing the airway diameter equal to or less than 4 mm (cross sectional area, 12.6 mm2); however, this may exacerbate auto-PEEP. No potential conflict of interest relevant to this article was reported. This study was presented at the Critical Care Symposium of the World Federation of Societies of Intensive and Critical Care Medicine (WFSICCM) on August 31, 2015 in Seoul, Korea. Figure 1. Schematic diagram of the two-lung model. (A) Proximal measurement setting. (B) Distal measurement setting. a: spirometer; b: filter; c: endotracheal tube with an internal diameter of 8 mm; d: connector with a variable internal diameter ranging from 3 mm to 8 mm; e: breathing circuit. Figure 2. Change of plateau pressure time in (A) C15, (B) C60, and (C) C120 groups. P1: start point of plateau pressure; P2: end point of plateau pressure; C15: static compliance of lung1 was manipulated as 15 ml/cmH2O; C60: static compliance of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. * P < 0.05 vs. proximal measurement.

(C) C120 groups. P1: start point of plateau pressure; P2: end point of plateau pressure; C15: static compliance of lung1 was manipulated as 15 ml/cmH2O; C60: static compliance of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. * P < 0.05 vs. proximal measurement. Figure 3. Change of plateau pressure time at (A) proximal and (B) distal measurements during change of internal diameter. P1: start point of plateau pressure; P2: end point of plateau pressure; C15: static compliance of lung1 was manipulated as 15 ml/cmH2O; C60: static compliance of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. * P < 0.05 vs. airway internal diameter of 8 mm. Figure 4. Distal measurements showing response of positive endexpiratory pressure according to internal diameter change. C15: static compliance of lung1 was manipulated as 15 ml/cmH2O; C60: static compliance of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. * P < 0.05 vs. airway internal diameter of 8 mm. Cross section area: internal diameter of 3, 4, 5, 6, 7 and 8 mm were measured to be 7.1, 12.6, 19.6, 28.3, 38.5, 50.3 mm2 respectively. † P < 0.05 vs. C60 group.

ce of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. * P < 0.05 vs. airway internal diameter of 8 mm. Cross section area: internal diameter of 3, 4, 5, 6, 7 and 8 mm were measured to be 7.1, 12.6, 19.6, 28.3, 38.5, 50.3 mm2 respectively. † P < 0.05 vs. C60 group. Figure 5. Flow-time curve of distal measurement in (A) C15, (B) C60, and (C) C120 groups. The area under the curve represents volume distributed to lung1. ID: internal diameter; C15: static compliance of lung1 was manipulated as 15 ml/cmH2O; C60: static compliance of lung1 was manipulated as 60 ml/cmH2O; C120: static compliance of lung1 was manipulated as 120 ml/cmH2O. Table 1. Comparison of VL1/VL2 and CL1/CL2 according to the airway internal diameter ID (mm) C15a group C60b group C120c group VL1/VL2 CL1/CL2d VL1/VL2 CL1/CL2d VL1/VL2 CL1/CL2d 3 0.10 ± 0.05 0.10 1.05 ± 0.16 1.05 1.46 ± 0.18e 1.67 4 0.11 ± 0.03 0.12 1.01 ± 0.09 1.03 3.06 ± 0.41e 2.74 5 0.12 ± 0.02 0.12 1.00 ± 0.07 1.02 3.72 ± 0.37 3.08 6 0.12 ± 0.02 0.13 0.97 ± 0.09 1.00 3.78 ± 0.47 3.32 7 0.12 ± 0.02 0.13 0.96 ± 0.06 1.00 3.77 ± 0.45 3.26 8 0.12 ± 0.02 0.13 0.97 ± 0.08 1.00 3.78 ± 0.60 3.39 Values are presented as mean ± standard deviation or mean, as appropriate. ID: internal diameter; VL1: volume distributed to lung1; VL2: volume distributed to lung2; CL1: static compliance of lung1; CL2: static compliance of lung2. a Static compliance of lung1 was manipulated as 15 ml/cmH2O. b Static compliance of lung1 was manipulated as 60 ml/cmH2O. c Static compliance of lung1 was manipulated as 120 ml/cmH2O.

ID: internal diameter; VL1: volume distributed to lung1; VL2: volume distributed to lung2; CL1: static compliance of lung1; CL2: static compliance of lung2. a Static compliance of lung1 was manipulated as 15 ml/cmH2O. b Static compliance of lung1 was manipulated as 60 ml/cmH2O. c Static compliance of lung1 was manipulated as 120 ml/cmH2O. d Three additional cycles were measured with a spirometer distally placed at lung2 to calculate the compliance ratio between the two lungs. e P < 0.05 vs. airway internal diameter of 8 mm. Table 2. Difference of lung1 volume between groups at increasing internal diameters of airway ID (mm) Tidal volume (ml) C15a group C60b group C120c group Total lung Lung1 Total lung Lung1 Total lung Lung1 3 524.6 ± 25.8 59.4 ± 17.9 530.5 ± 19.0 272.1 ± 16.4 551.0 ± 17.1 307.9 ± 12.1d 4 535.2 ± 27.8 55.2 ± 10.2 541.6 ± 21.8 272.4 ± 9.3 519.9 ± 15.8 401.6 ± 16.0d 5 541.5 ± 28.8 56.3 ± 9.7 549.3 ± 24.4 274.8 ± 9.2 534.8 ± 17.1 428.2 ± 15.8 6 544.2 ± 27.9 60.4 ± 4.7 552.9 ± 24.2 272.6 ± 8.9 543.3 ± 18.1 432.7 ± 18.9 7 544.7 ± 27.8 60.2 ± 6.4 557.3 ± 23.4 274.2 ± 10.7 547.2 ± 19.0 433.4 ± 18.5 8 549.8 ± 28.6 59.3 ± 8.1 557.2 ± 24.2 274.9 ± 10.0 549.3 ± 18.2 433.7 ± 20.4 Values are presented as mean ± standard deviation. ID: internal diameter. a Static compliance of lung1 was manipulated as 15 ml/cmH2O. b Static compliance of lung1 was manipulated as 60 ml/cmH2O. c Static compliance of lung1 was manipulated as 120 ml/cmH2O. d P < 0.05 vs. airway internal diameter of 8 mm.

Introduction Some patients experience serious adverse events while in the hospital, including sudden cardiac arrest, respiratory failure, acute changes in consciousness, unplanned intensive care unit (ICU) admission, and sudden death [1]; these may lead to irreversible organ damage, increased mortality rate, prolonged hospital stay, and increased medical costs [2,3]. Many studies have reported on the epidemiology and causes of unexpected adverse events in the hospital [2,4-6]. One of the main causes of such adverse events is insufficient, delayed, or incorrect medical detection systems [4,7]. To overcome these problems in medical detection, rapid response systems (RRSs) have been introduced, including the formation of rapid response teams, also known as medical emergency teams or critical care outreach [8,9]. Several studies have shown results, such as decreases in hospital mortality and increases in better care for acutely ill patients, after implementation of an RRS [10-12]. However, RRSs also have shortcomings. The RRS team commonly requires additional staff consisting of critical care staff or fellows, nurses, and respiratory therapists, who can resolve patients’ critical problems [9,13,14]. Maintaining the RRS generates additional expenses including educational expenses and communication system development costs. Furthermore, the RRS may lower the sense of responsibility for patients and interest from the primary doctor [9,15-18].

piratory therapists, who can resolve patients’ critical problems [9,13,14]. Maintaining the RRS generates additional expenses including educational expenses and communication system development costs. Furthermore, the RRS may lower the sense of responsibility for patients and interest from the primary doctor [9,15-18]. Our institute has not yet implemented an RRS owing to various limitations, including cost and lack of staff. Instead, we developed an automatic alarm system for unexpected unstable vital signs in admitted patients in the general ward. This automatic alarm system makes the best use of existing resources, such as primary doctors, including residents, nurses, electronic medical record systems (EMRs), and electronic communication systems. We have termed this the “medical emergency system” (MES). The aim of this pilot study was to determine the effectiveness of the MES before expanding this system to all of our departments. The primary objective was to prevent pre-ICU cardiac arrest and decrease mortality rates.

Rs), and electronic communication systems. We have termed this the “medical emergency system” (MES). The aim of this pilot study was to determine the effectiveness of the MES before expanding this system to all of our departments. The primary objective was to prevent pre-ICU cardiac arrest and decrease mortality rates. Materials and Methods 1) Study setting This was planned as a pilot study prior to expanding the MES to all departments in our institute. This was a retrospective, observational study comparing the performance of patients admitted to only one ward on the pulmonary department at a 3,000-bed (30-bed medical intensive care unit) university tertiary referral hospital in Seoul, Korea, with retrospective data from the same hospital before the application of the MES during the winter season. This study was conducted over three 3-month periods (before implementation of the MES, December 2013-February 2014; after implementation of the MES, December 2014-February 2015 and December 2015-February 2016). The primary outcomes were the rate of cardiopulmonary resuscitation (CPR), mortality rate, and ICU admission rate. Length of hospital stay and length of ICU stay were also analyzed.

on of the MES, December 2013-February 2014; after implementation of the MES, December 2014-February 2015 and December 2015-February 2016). The primary outcomes were the rate of cardiopulmonary resuscitation (CPR), mortality rate, and ICU admission rate. Length of hospital stay and length of ICU stay were also analyzed. 2) MES design The MES is an automatic alarm system that detects warning signs of disease progression or adverse events in patients and generates an appropriate awareness for primary care physicians. The MES uses the EMR and existing communication system. Figure 1 shows the design of the MES. In brief, if vital signs entered in the EMR by a patient’s nurse satisfy the criteria for MES, the MES will automatically alert the primary doctor, resident, and on-call doctor of the abnormal vital signs. MES inclusion criteria include abnormal respiration rate, oxygen saturation, heart rate, and systolic blood pressure (Table 1). The MES is not turned on by patients who agree to “do not resuscitate,” are younger than 18 years, or are admitted in the emergency room or ICU. Any doctor who receives the MES message manages the patient according to the MES manuals (Figure 2). After management, the doctor records the method of management, status of the patient, and the results of management. The doctor can then turn off the MES. The MES includes education for primary care physicians that relates to basic procedures and plans for situations such as acute respiratory distress, shock, and arrhythmia. Education was conducted periodically before and after implementing the MES.

nt, and the results of management. The doctor can then turn off the MES. The MES includes education for primary care physicians that relates to basic procedures and plans for situations such as acute respiratory distress, shock, and arrhythmia. Education was conducted periodically before and after implementing the MES. 3) Data collection The data from patients admitted to one ward of the pulmonary department were collected from the hospital electronic medical records. Clinical data on hospitalization path, length of stay, admission to the ICU, development of CPR, and mortality were evaluated. The Charlson comorbidity index (CCI) was calculated for evaluation of comorbidity [19]. Additionally, we inspected progress of patients admitted to the ICU after implementation of the MES. 4) Data analysis Statistical analysis was performed using SPSS version 23.0 (IBM Corp., Armonk, NY, USA). All continuous variables are expressed as mean with standard deviation or median with interquartile range. The chi-square test or Fisher exact test was used to assess differences among the groups. Continuous variables were analyzed using Kruskal-Wallis test or analysis of variance. 5) Ethical approval This study protocol was approved by the institutional review board of the university tertiary referral hospital in Seoul (No. 4-2016-0928).

4) Data analysis Statistical analysis was performed using SPSS version 23.0 (IBM Corp., Armonk, NY, USA). All continuous variables are expressed as mean with standard deviation or median with interquartile range. The chi-square test or Fisher exact test was used to assess differences among the groups. Continuous variables were analyzed using Kruskal-Wallis test or analysis of variance. 5) Ethical approval This study protocol was approved by the institutional review board of the university tertiary referral hospital in Seoul (No. 4-2016-0928). Results A total of 571 patients were admitted to one ward of the pulmonary department during the observation periods. The mean age of the patients was 64.2 years (range, 18-93 years) and 338 of the patients were male (59.2%). One hundred fifty-six patients were admitted in December 2013-February 2014 before implementation of the MES. The 203 patients hospitalized in December 2014-February 2015 and the 212 patients admitted in December 2015-February 2016 were under the MES. The number of hospitalized patients increased in 2014 and 2015. The sex, mean age, mean body mass index, and hospitalization path (such as emergency department and outpatient) were not significantly different between groups (Table 2). However, the CCI, indicating comorbidity to predict short- and long-term mortality, was significantly different between the groups (P = 0.038), and was significantly higher in 2015 than in 2013 (P = 0.032) (Table 2).

such as emergency department and outpatient) were not significantly different between groups (Table 2). However, the CCI, indicating comorbidity to predict short- and long-term mortality, was significantly different between the groups (P = 0.038), and was significantly higher in 2015 than in 2013 (P = 0.032) (Table 2). Table 3 shows the results of the analysis before and after implementing the MES. During this pilot study, the MES automatically turned on 568 alarms for 415 admitted patients. Among 568 alarms, 170 (29.9%) were caused by problems in respiration rate; 149 (26.2%) were caused by low oxygen saturation; 77 (13.6%) were caused by problems in heart rate; and 172 (30.3%) were caused by low systolic blood pressure. The response rate to MES alarms increased in the second year of MES implementation compared to the first year (82.7% vs. 43.79%). There was no significant difference in the rate of development of CPR. The mortality rate also did not differ between groups (P = 1.000). The admission rate of the ICU increased, but this was not statistically significant. The majority of patients were admitted to the ICU for ventilator care (2013-2014, 80%; 2014-2015, 80%; 2015-2016, 61%). Among them, four patients admitted from the general ward to the ICU in December 2015-February 2016 underwent CPR with 1 day of ICU admission.

ICU increased, but this was not statistically significant. The majority of patients were admitted to the ICU for ventilator care (2013-2014, 80%; 2014-2015, 80%; 2015-2016, 61%). Among them, four patients admitted from the general ward to the ICU in December 2015-February 2016 underwent CPR with 1 day of ICU admission. The median length of hospital stay and median length of ICU stay among all of the patients were not significantly different between the groups. However, there were differences in hospitalized days on excluding patients admitted to the ICU and length of ICU stay on ICU-admitted patients. Figure 3 shows the trend of the MES to turn on according to time.

y and median length of ICU stay among all of the patients were not significantly different between the groups. However, there were differences in hospitalized days on excluding patients admitted to the ICU and length of ICU stay on ICU-admitted patients. Figure 3 shows the trend of the MES to turn on according to time. Discussion This study examined the effects of the introduction of the MES, an automatic alarm system for unexpected unstable vital signs of patients on the general ward, through the composite incidence of CPR, ICU admissions, and mortality. Our study showed similar CPR rates before and after implementation of the MES. There were no significant differences in mortality of admitted patients or rate of ICU admission. Most studies on RRSs analyzed a large number of subjects and showed low CPR and mortality rates [1,2,9-15]. As the number of subjects in our study was relatively small compared to that in other studies, the results of CPR rate and mortality rates in our study were hard to accept as they are. However, our study showed that the number of admitted patients increased, as did comorbidity of the admitted patients, after implementation of the MES, but the rates of development of CPR and mortality remained the same. Additionally, as the response rate to MES alarms increased, the number of ICU admission also increased. We believe that the MES provides early detection of problems, aiding in the decision to admit patients to the ICU before the development of unexpected adverse events. In agreement with this hypothesis, four patients admitted to the ICU after MES implementation underwent CPR within 1 day of admission.

on also increased. We believe that the MES provides early detection of problems, aiding in the decision to admit patients to the ICU before the development of unexpected adverse events. In agreement with this hypothesis, four patients admitted to the ICU after MES implementation underwent CPR within 1 day of admission. The strong point of the MES is that it needs only an initial system construction cost and a low maintenance cost and does not require an additional workforce. The weakness of the MES is that necessity of more education can make startup difficult. Without educated physicians, the MES is only a simple alarm clock. To overcome this, our institution paid attention to periodic education for healthcare workers, and the response rate to the MES gradually increased after repeated education. Furthermore, repetitive education imbued healthcare workers with the ability to manage emergencies and it suggests that staff properly trained under the MES could efficiently cope with simultaneous emergencies on multiple wards.

are workers, and the response rate to the MES gradually increased after repeated education. Furthermore, repetitive education imbued healthcare workers with the ability to manage emergencies and it suggests that staff properly trained under the MES could efficiently cope with simultaneous emergencies on multiple wards. As this was a pilot study, it had several limitations. First, this study was retrospective in design; thus, we could not systematically analyze multiple factors. As this study does not have many variables, we did not know whether multiple variables were correlated. Second, the patients in this study were limited to admission of one ward on the pulmonary department. That ward had relatively low severity compared to other pulmonary departments because it included rooms for patients undergoing bronchoscopy with biopsy. Furthermore, healthcare workers, including the doctors and nurse of the pulmonology ward, are better trained than workers in other department wards, as they see many patients with serious illnesses who need more monitoring. This may have affected the results of this study. Third, the number of subjects in our study was relatively small compared to other studies. Fourth, our observation and implementation time may be too short. Some studies have reported that significant effects of RRSs did not emerge until 2 years after implementation [20,21]. Our study did not find any long-term effects of the MES, because we did not perform continuous studies.

mall compared to other studies. Fourth, our observation and implementation time may be too short. Some studies have reported that significant effects of RRSs did not emerge until 2 years after implementation [20,21]. Our study did not find any long-term effects of the MES, because we did not perform continuous studies. Although we did not find significant improvement in our primary outcomes with MES implementation, the CPR rate and mortality rate did not increase, despite increased comorbidity of patients. However, this study is a pilot study, so we expect that such limitations may be overcome after expanding the MES to all departments in our hospital. Thus, we believe that the MES stands a better chance of significant effects with expansion and increased duration. We will continue to expand the MES to all departments in our hospital. No potential conflict of interest relevant to this article was reported. Figure 1. Design of the medical emergency system (MES). When a nurse enters vital sign data into the electronic medical record system, the computer automatically analyzes this information. If vital signs meet the MES criteria, a message is automatically sent to the primary doctor, resident, and on-call doctor. A doctor who receives the message must treat the patient and chart the treatment to deactivate the MES. If the MES is not deactivated, the system will continue to send the message to the doctors. EMR: electronic medical record; SMS: short message service.

ically sent to the primary doctor, resident, and on-call doctor. A doctor who receives the message must treat the patient and chart the treatment to deactivate the MES. If the MES is not deactivated, the system will continue to send the message to the doctors. EMR: electronic medical record; SMS: short message service. Figure 2. Protocol for the medical emergency system. Management flow for (A) shock, (B) hypoxemia and tachypnea, (C) tachycardia. SBP: systolic blood pressure; SpO2: oxygen saturation; RR: respiratory rate; HR: heart rate; EMR; electronic medical record; GCS: glasgow coma scale; MES: medical emergency system; qSOFA: quick sepsis related organ failure assessment; CBC: complete blood count; diff: differential count; T.bil: total bilirubin; Cr: creatinine; ABGA: arterial blood gas analysis; EKG: electrocardiogram; CXR: chest X-ray; MAP: mean arterial pressure; ICU: intensive care unit; Tx: treatment; HAT; hypotension, altered mental status, tachypnea; BP: blood pressure; Resp: respiration; COPD: chronic obstructive pulmonary disease; PTE: pulmonary thromboembolism; CT: computed tomography; TTE: transthoracic echocardiogram; AVNRT: atrioventricular nodal reentrant tachycardia; AVRT: atrioventricular reentrant tachycardia; VT: ventricular tachycardia; VF: ventricular fibrillation; DDx: differential diagnosis; A-fib: atrial fibrillation; PSVT: paroxysmal supraventricular tachycardia. aEye response, 1–4; verbal response,1–5; motor response, 1–6. Figure 3. Number of medical emergency system (MES) to turn on and time of MES calls.

Figure 2. Protocol for the medical emergency system. Management flow for (A) shock, (B) hypoxemia and tachypnea, (C) tachycardia. SBP: systolic blood pressure; SpO2: oxygen saturation; RR: respiratory rate; HR: heart rate; EMR; electronic medical record; GCS: glasgow coma scale; MES: medical emergency system; qSOFA: quick sepsis related organ failure assessment; CBC: complete blood count; diff: differential count; T.bil: total bilirubin; Cr: creatinine; ABGA: arterial blood gas analysis; EKG: electrocardiogram; CXR: chest X-ray; MAP: mean arterial pressure; ICU: intensive care unit; Tx: treatment; HAT; hypotension, altered mental status, tachypnea; BP: blood pressure; Resp: respiration; COPD: chronic obstructive pulmonary disease; PTE: pulmonary thromboembolism; CT: computed tomography; TTE: transthoracic echocardiogram; AVNRT: atrioventricular nodal reentrant tachycardia; AVRT: atrioventricular reentrant tachycardia; VT: ventricular tachycardia; VF: ventricular fibrillation; DDx: differential diagnosis; A-fib: atrial fibrillation; PSVT: paroxysmal supraventricular tachycardia. aEye response, 1–4; verbal response,1–5; motor response, 1–6. Figure 3. Number of medical emergency system (MES) to turn on and time of MES calls. Table 1. Medical emergency system inclusion criteria

Figure 2. Protocol for the medical emergency system. Management flow for (A) shock, (B) hypoxemia and tachypnea, (C) tachycardia. SBP: systolic blood pressure; SpO2: oxygen saturation; RR: respiratory rate; HR: heart rate; EMR; electronic medical record; GCS: glasgow coma scale; MES: medical emergency system; qSOFA: quick sepsis related organ failure assessment; CBC: complete blood count; diff: differential count; T.bil: total bilirubin; Cr: creatinine; ABGA: arterial blood gas analysis; EKG: electrocardiogram; CXR: chest X-ray; MAP: mean arterial pressure; ICU: intensive care unit; Tx: treatment; HAT; hypotension, altered mental status, tachypnea; BP: blood pressure; Resp: respiration; COPD: chronic obstructive pulmonary disease; PTE: pulmonary thromboembolism; CT: computed tomography; TTE: transthoracic echocardiogram; AVNRT: atrioventricular nodal reentrant tachycardia; AVRT: atrioventricular reentrant tachycardia; VT: ventricular tachycardia; VF: ventricular fibrillation; DDx: differential diagnosis; A-fib: atrial fibrillation; PSVT: paroxysmal supraventricular tachycardia. aEye response, 1–4; verbal response,1–5; motor response, 1–6. Figure 3. Number of medical emergency system (MES) to turn on and time of MES calls. Table 1. Medical emergency system inclusion criteria Indicator Criteria 1 Acute respiratory distress RR ≤ 8/min or ≥ 30/min 2 Acute hypoxia SpO2 < 90% (regardless of oxygen therapy) 3 Tachycardia or bradycardia with symptoms HR ≤ 40/min or ≥ 140/min 4 Blood pressure Systolic BP <90 mmHg MES activation: patient fulfills any criterion. Exception: patients who sign a ‘do not resuscitate,’ are younger than 18 years old, or have been admitted to the emergency room or ICU. RR: respiratory rate; SpO2: oxygen saturation; HR: heart rate; BP: blood pressure; MES: medical emergency system; ICU: intensive care unit.

activation: patient fulfills any criterion. Exception: patients who sign a ‘do not resuscitate,’ are younger than 18 years old, or have been admitted to the emergency room or ICU. RR: respiratory rate; SpO2: oxygen saturation; HR: heart rate; BP: blood pressure; MES: medical emergency system; ICU: intensive care unit. Table 2. Baseline characteristics of total patients Variable 2013-2014 (n = 156) 2014-2015 (n = 203) 2015-2016 (n = 212) P-value Male sex 91(58.3) 132 (65.0) 115 (54.2) 0.080 Age (yr) 66.1 ± 14.3 63.5 ± 13.8 63.5 ± 15.1 0.160 BMI (kg/m2) 21.9 ± 3.3 22.5 ± 4.1 22.7 ± 3.5 0.131 Hospitalization path 0.116 Emergency room 55 (35.3) 71 (35.0) 93 (43.9) Outpatient 101 (64.7) 132 (65.0) 119 (56.1) CCI 2.0 ± 2.4 2.3 ± 2.7 2.7 ± 2.7 0.038 Values are presented as number (%) or mean ± standard deviation. BMI: body mass index; CCI: Charlson comorbidity index. Table 3. Outcomes after implementation of medical emergency system

Variable 2013-2014 (n = 156) 2014-2015 (n = 203) 2015-2016 (n = 212) P-value Male sex 91(58.3) 132 (65.0) 115 (54.2) 0.080 Age (yr) 66.1 ± 14.3 63.5 ± 13.8 63.5 ± 15.1 0.160 BMI (kg/m2) 21.9 ± 3.3 22.5 ± 4.1 22.7 ± 3.5 0.131 Hospitalization path 0.116 Emergency room 55 (35.3) 71 (35.0) 93 (43.9) Outpatient 101 (64.7) 132 (65.0) 119 (56.1) CCI 2.0 ± 2.4 2.3 ± 2.7 2.7 ± 2.7 0.038 Values are presented as number (%) or mean ± standard deviation. BMI: body mass index; CCI: Charlson comorbidity index. Table 3. Outcomes after implementation of medical emergency system From December to February 2013-2014 (n = 156) 2014-2015 (n = 203) 2015-2016 (n = 212) P-value Total MES turn on - 279 289 - RR - 73 (26.2) 97 (33.6) SpO2 - 88 (31.5) 61 (21.1) HR - 32 (11.5) 45 (15.6) SBP - 86 (30.8) 86 (29.8) Response rate to MES (%) - 43.8 82.7 - Primary outcome CPR 1 (0.6) 1 (0.4) 0 (0) 0.398 Mortality 1 (0.6) 1 (0.5) 1 (0.5) 1.000 ICU admission 10 (6.4) 15 (7.4) 18 (8.5) 0.753 Secondary outcome Days of hospital stay 4.0 (3.0-12.8) 3.0 (2.0-11.0) 3.0 (2.0-11.0) 0.051 Days of ICU stay 9.5 (3.8-16.3) 9.0 (4.0-13.0) 5.5 (3.8-22.0) 0.984 Days of hospital stay excluding ICU patients 3.0 (3.0-10.0) 3.0 (2.0-10.0) 3.0 (2.0-9.0) 0.038 Days of hospital stay of ICU patients 28 (14.0-63.5) 27 (19.0-41.0) 13 (12.0-22.3) 0.021 Values are presented as number (%) or median (interquartile range) unless otherwise indicated. MES: medical emergency system; RR: respiration rate; SpO2: oxygen saturation; HR: heart rate; SBP: systolic blood pressure; CPR: cardiopulmonary resuscitation; ICU: intensive care unit.

Introduction Rapid response teams (RRTs) have been implemented to prevent serious adverse events, such as cardiac arrest, unexpected admission to the intensive care unit (ICU), and death [1]. According to several studies, serious adverse events are preceded for some hours by warning signs including abnormal vital signs, physiological instability, and an altered mental status [2-5]. Therefore, the goal of the RRT is early identification of a deteriorating patient showing warning signs, and an appropriate response before the occurrence of unintended adverse consequences [6]. To ensure patient safety and improve hospital quality of care, RRTs have been implemented in a number of countries. In the United States, the 100,000 lives campaign of the Institute for Healthcare Improvement recommended that hospitals implement RRTs as one of six strategies to reduce preventable in-hospital deaths [7]. Nowadays, most US hospitals have implemented an RRT in some form or other [8,9]. In Australia, an RRT system exists in two-thirds of all hospitals [10]. In Korea, RRTs have been mainly adopted in tertiary medical centers [11-13]. The clinical benefits of an RRT are still unclear [7,9]. Some studies showed favorable outcomes in terms of a decreased incidence of mortality from cardiac arrest and reduced ICU admission rate through the use of an RRT [14-17]. In contrast, other studies found no significant difference or had ambiguous results [18,19]. This discrepancy may be due to the heterogeneity of the study populations, and differences in RRT type, activation criteria, and quality of activation.

ac arrest and reduced ICU admission rate through the use of an RRT [14-17]. In contrast, other studies found no significant difference or had ambiguous results [18,19]. This discrepancy may be due to the heterogeneity of the study populations, and differences in RRT type, activation criteria, and quality of activation. In this study, we retrospectively reviewed the RRT activation records of our institution to determine the clinical characteristics and predictors of survival of Korean patients who required an RRT activation.

ac arrest and reduced ICU admission rate through the use of an RRT [14-17]. In contrast, other studies found no significant difference or had ambiguous results [18,19]. This discrepancy may be due to the heterogeneity of the study populations, and differences in RRT type, activation criteria, and quality of activation. In this study, we retrospectively reviewed the RRT activation records of our institution to determine the clinical characteristics and predictors of survival of Korean patients who required an RRT activation. Materials and Methods 1) Subjects We retrospectively reviewed the RRT activation records of Seoul St. Mary’s Hospital from June 2013 to December 2016. Seoul St. Mary’s Hospital is a tertiary teaching hospital in Korea. With 22 upper floors, six basements, and 1,320 beds (including 119 beds in the ICU), it is the largest Korean hospital contained within a single building. Approximately 50,000 adult patients admitted to the hospital per year. The RRT of the hospital, called the St. Mary’s Advanced Life Support Team (SALT), was implemented in June 2013 for hospital quality improvement. During the first 2 months of testing, the RRT targeted patients in the Department of Orthopedic Surgery. After successful testing, the range of RRT activations was gradually expanded to include other departments. Since June 2016, the RRT has covered all departments except those for pediatric patients. The RRT can be activated by both phone calls and a screening system. Approval was obtained from the institutional review board of Seoul St. Mary’s Hospital. The requirement for informed consent was waived by the ethical review board.

ce June 2016, the RRT has covered all departments except those for pediatric patients. The RRT can be activated by both phone calls and a screening system. Approval was obtained from the institutional review board of Seoul St. Mary’s Hospital. The requirement for informed consent was waived by the ethical review board. 2) RRT description From June 2013 to May 2016, the duty hours of the RRT were from 8 ᴀᴍ to 10 ᴩᴍ on weekdays and from 8 ᴀᴍ to 4 ᴩᴍ on weekends and holidays. During this period, the RRT comprised three experienced RRT nurses, three to four pulmonologists (intensivists in charge of the medical ICU), two to three surgeons (intensivists in charge of the surgical ICU), and two to three residents training in the ICU. Intensivists were on duty from 8 ᴀᴍ to 5 ᴩᴍ on weekdays, while residents covered the RRT after 5 ᴩᴍ, and on weekends and holidays. However, after May 2016, the RRT was changed to a 24-hour system and a greater number of medical staff (drawn from internal medicine, cardiothoracic surgery, general surgery, and neurosurgery) joined the RRT for overnight duty. The number of nurses in the RRT was also increased to nine.

2) RRT description From June 2013 to May 2016, the duty hours of the RRT were from 8 ᴀᴍ to 10 ᴩᴍ on weekdays and from 8 ᴀᴍ to 4 ᴩᴍ on weekends and holidays. During this period, the RRT comprised three experienced RRT nurses, three to four pulmonologists (intensivists in charge of the medical ICU), two to three surgeons (intensivists in charge of the surgical ICU), and two to three residents training in the ICU. Intensivists were on duty from 8 ᴀᴍ to 5 ᴩᴍ on weekdays, while residents covered the RRT after 5 ᴩᴍ, and on weekends and holidays. However, after May 2016, the RRT was changed to a 24-hour system and a greater number of medical staff (drawn from internal medicine, cardiothoracic surgery, general surgery, and neurosurgery) joined the RRT for overnight duty. The number of nurses in the RRT was also increased to nine. 3) RRT activation and screening The criteria for RRT activation are shown in Figure 1. There are 12 activation criteria, including a direct phone call for any serious concerns about overall deterioration as detected by a physician, nurse, or caregivers at the bedside. Exclusion criteria were patients in the department of pediatrics, patients with a do not resuscitate status or in cardiopulmonary resuscitation. RRT activation was defined as the process when intensivists or residents arrive at the patient bedside and apply medical treatment, give consultation, or make decisions about ICU transfer. A portable multi-monitor and the enterprise point-of-care blood analysis system (Epoc; Alere, Waltham, MA, USA) were carried by the RRT at each activation. The Epoc system is a handheld, wireless solution that provides blood gas, electrolyte, and metabolite results at the patient’s bedside within approximately 30 seconds of the introduction of a sample. Patients are also protected by an electronic medical record (EMR)-based screening system. If the clinical value in the EMR satisfies the RRT activation criteria, patients are automatically screened. The RRT charge nurse checks the condition of patients and recommends RRT activation to ward staff or nurses, as necessary. Postoperative and post-ICU monitoring are also in operation for particular departments. In 2016, 26,783 patients were screened (18,037 EMR-screened patients, 8,041 postoperative patients, and 707 post-ICU patients).

se checks the condition of patients and recommends RRT activation to ward staff or nurses, as necessary. Postoperative and post-ICU monitoring are also in operation for particular departments. In 2016, 26,783 patients were screened (18,037 EMR-screened patients, 8,041 postoperative patients, and 707 post-ICU patients). 4) Data collection Basic patient characteristics data, including sex, age, body mass index, admission route, department, admission ward, and past medical history (including malignancies), were collected from the EMR database. We also checked the status of the operation, number of postoperative days, activation method, response time, Acute Physiology and Chronic Health Evaluation-II (APACHE-II) score, modified early warning score (MEWS), position of activation, and RRT activation day, time, and reason. APACHE-II scores were collected from all patients with RRT activation regardless of ICU transfer, except the patients with RRT activation in 2013. Vital signs, and laboratory test and arterial blood gas analysis results at the time of RRT activation were also collected. Survival, length of hospital stay after RRT activation, and ICU transfer status (at the time of RRT activation and 24 hours after RRT activation) were examined as outcomes.

RRT activation in 2013. Vital signs, and laboratory test and arterial blood gas analysis results at the time of RRT activation were also collected. Survival, length of hospital stay after RRT activation, and ICU transfer status (at the time of RRT activation and 24 hours after RRT activation) were examined as outcomes. 5) Statistical analysis The mean and standard deviation were computed for normally distributed continuous variables, whereas medians and the interquartile range (IQR, 25th to 75th percentile) were used for non-normally distributed continuous data. Categorical data are described as numbers (%). Student t-test was performed for normally distributed data, and the Mann-Whitney U-test was used for non-normally distributed data to compare clinical characteristics between subgroups. Categorical variables were compared using the chi-square and Fisher exact tests, as appropriate. Missing values were excluded from the analyses. Logistic regression analyses were performed to estimate the associations between survival and clinical characteristics. Clinical parameters with a P-value of 0.2 in the univariate logistic regression were included in the multivariate logistic regression. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated. Statistical analyses were performed using R software ver. 3.1.1 (https://cran.r-project.org/). A P-value <0.05 was considered to be statistically significant.

in the univariate logistic regression were included in the multivariate logistic regression. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated. Statistical analyses were performed using R software ver. 3.1.1 (https://cran.r-project.org/). A P-value <0.05 was considered to be statistically significant. Results 1) Epidemiology There were 1.4, 1.5, and 2.0 RRT activations/1,000 admissions in 2014, 2015, and 2016, respectively. Figure 1 shows the reasons for RRT activation. There are 12 activation criteria, and each criterion can be duplicated. The most frequent reason for activation was “serious concerns about overall deterioration.” “O2 saturation ≤90% for more than 5 minutes with prior oxygen therapy,” “sudden alteration of consciousness,” and “systolic blood pressure ≤85 mmHg with correlated symptoms or signs” were the most common reasons for RRT activation. RRT activation by phone call was mainly done by residents (77 cases, 35.8%) and nurses (127 cases, 59.1%) (Figure 2A). Orthopedics (49.1%), obstetrics and gynecology (13.6%), urology (10.8%), and plastic surgery (9.8%) were the most frequently activated departments (Figure 2B). We also analyzed the number of RRT activations by the specific day of the week according to surgical status and activation method (Figure 2C and D). Although there was no significant group difference, the postoperative patients tended to be activated on weekdays. In contrast, the non-operation group showed only marginal variation over the week. Similarly, RRT activation by phone call showed a decreasing trend at the weekend. However, RRT activation with screening showed only marginal variation over the week.

erence, the postoperative patients tended to be activated on weekdays. In contrast, the non-operation group showed only marginal variation over the week. Similarly, RRT activation by phone call showed a decreasing trend at the weekend. However, RRT activation with screening showed only marginal variation over the week. 2) Clinical characteristics Table 1 lists the basic characteristics of the RRT activation cases. Of the 287 RRT activation cases, 43.2% were for male patients and their median age was 70.0 years (IQR, 58.0 to 78.0 years). The percentage of admissions via the outpatients department and emergency room was 56.1% and 43.9%, respectively. The median body mass index, presence of malignancy, and a postoperative status were 23.0 kg/m2 (IQR, 20.1 to 25.8 kg/m2 ), 34.8%, and 69.3%, respectively. For the postoperative patients, the RRT was activated on a median of 2.0 days (IQR, 1.0 to 5.0 days). The median response time, MEWS, and APACHE-II score were 8.6 minutes (IQR, 5.6 to 11.6 minutes), 5.0 points (IQR, 4.0 to 7.0 points), and 14.0 points (IQR, 10.0 to 18.0 points), respectively. Approximately 80% of RRT activations occurred on weekdays, not including holidays, and the median number of days between admission and RRT activation was 6.0 days (IQR, 3.0 to 13.0 days). We also compared patients according to survival and activation method. The survival group showed a significantly lower rate of malignancy (P < 0.001), higher postoperative rate (P = 0.041), lower MEWS (P < 0.001), and lower APACHE-II score (P < 0.001) compared to the expired group. In addition, the survival group showed a significantly lower number of days between admission and RRT activation (P < 0.001). When only analyzing the postoperative patients, the survival group showed a lower number of postoperative days on which the RRT was activated (P = 0.012). Comparing the groups according to phone and screening activation methods, there was no significant difference in the basic characteristics.

T activation (P < 0.001). When only analyzing the postoperative patients, the survival group showed a lower number of postoperative days on which the RRT was activated (P = 0.012). Comparing the groups according to phone and screening activation methods, there was no significant difference in the basic characteristics. 3) Outcome After RRT activation, about 30% of patients were moved to the ICU for further treatment (Table 2). The survival rate after RRT activation was 83.6%, and 6.3% of patients required endotracheal intubation. The median hospital stay after RRT activation was 12.0 days (IQR, 6.0 to 25.5 days). The survival group showed a significantly lower intubation rate (P = 0.019) and longer hospital stay after activation (P = 0.008). Shorter hospital stay after activation in the expired group compared with survival group was due to early death after RRT activation. The phone-activated group showed a significantly higher rate of ICU admissions (P = 0.017) and longer hospital stay (P = 0.022) after RRT activation compared to the screening group.

08). Shorter hospital stay after activation in the expired group compared with survival group was due to early death after RRT activation. The phone-activated group showed a significantly higher rate of ICU admissions (P = 0.017) and longer hospital stay (P = 0.022) after RRT activation compared to the screening group. 4) Clinical parameters associated with survival On univariate logistic regression analysis, malignancy, a postoperative status, MEWS, APACHE-II score, number of days between admission and activation, and intubation were significantly associated with survival (P < 0.001, P = 0.029, P < 0.001, P < 0.001, P < 0.001, and P = 0.012, respectively) (Table 3). In multivariate logistic regression analysis, malignancy, APACHE-II score, and number of days between admission and activation were the clinical parameters found to be significantly associated with survival (P < 0.001, P = 0.009, and P = 0.001, respectively), with an OR of 7.47 (95% CI, 3.00 to 20.53), 1.10 (95% CI, 1.03 to 1.19), and 1.04 (95% CI, 1.02 to 1.07), respectively.

PACHE-II score, and number of days between admission and activation were the clinical parameters found to be significantly associated with survival (P < 0.001, P = 0.009, and P = 0.001, respectively), with an OR of 7.47 (95% CI, 3.00 to 20.53), 1.10 (95% CI, 1.03 to 1.19), and 1.04 (95% CI, 1.02 to 1.07), respectively. Discussion In this study, we retrospectively analyzed the RRT activation records of a single tertiary medical center. Over the course of 3.5 years, there were 287 RRT activations by both phone and screening. In our medical center, around 70% of RRT activations were for postoperative patients. After RRT activation, the survival rate of patients was 83.6%, and approximately 30% of patients were moved to the ICU for further treatment. The presence of malignancy was the most important factor in survival. In addition, RRT activation with screening showed a favorable outcome compared with RRT activation by phone.

After RRT activation, the survival rate of patients was 83.6%, and approximately 30% of patients were moved to the ICU for further treatment. The presence of malignancy was the most important factor in survival. In addition, RRT activation with screening showed a favorable outcome compared with RRT activation by phone. The afferent limb associated with RRT activation is one of the most important components to improve outcomes [13]. In our medical center, frequent promotion of RRT and education of the medical staff was performed along with RRT implementation. In addition, our RRT can be easily activated by a phone call to an RRT nurse. As a result, residents and nurses were the most frequent RRT activators, according to the data. Due to easy access to an RRT nurse, residents or nurses may feel less intimidated by making an RRT activation, and the percentage of early calls could thus increase [1]. Patient screening is also performed by an experienced RRT nurse for early detection and response. The current study showed that screening can lead to a favorable outcome, such as with respect to the ICU transfer rate and length of hospital stay after activation, compared with a phone call. Huh et al. [13] also demonstrated a better activation outcome by screening rather than by phone call. However, there were no significant differences in basic characteristics including MEWS and APACHE-II score between phone-activated group and screening group in our study. There are possibilities that patients were in different clinical severity with same MEWS or APACHE-II score. In addition, further studies with large population are needed to evaluate the clinical status more precisely.

cs including MEWS and APACHE-II score between phone-activated group and screening group in our study. There are possibilities that patients were in different clinical severity with same MEWS or APACHE-II score. In addition, further studies with large population are needed to evaluate the clinical status more precisely. In the efferent limb, our median response time, which is the time between RRT activation and the arrival of a physician, was 8.6 minutes. We did not include the type of intervention in this study due to a number of overlaps, inaccurate records, and difference in the quality. However, interventions were commonly for (1) managing intravenous fluid, (2) administering diuretics, (3) modifying antibiotics, (4) supplying oxygen and nebulizers, (5) ordering diagnostic studies, and (6) recommending consultation with other departments. About 30% of patients were moved to the ICU for further treatment.

ever, interventions were commonly for (1) managing intravenous fluid, (2) administering diuretics, (3) modifying antibiotics, (4) supplying oxygen and nebulizers, (5) ordering diagnostic studies, and (6) recommending consultation with other departments. About 30% of patients were moved to the ICU for further treatment. Interestingly, about 70% of patients with an RRT activation were in postoperative care. In addition, departments related to surgery, such as orthopedics, obstetrics and gynecology, urology, and plastic surgery, were the main activators of the RRT in our medical center. From our results, patients in acute postoperative care tend to require a greater degree of RRT activation. In these surgical departments, a shortage of surgeons working on the general ward, and inexperience with medical emergencies, were the main reasons for an RRT call. However, we also experienced adverse outcomes, such as a decreased sense of responsibility, desensitization to emergencies for general ward physicians, and a heavy burden on RRT staff in cases where a high degree of intervention, rather than a quick second opinion, was expected [3,10]. In contrast, the proportion of internal medicine patients requiring RRT activation was lower than in other medical centers [11]. Some internists still questioned the effectiveness of the RRT and prefer to make decisions within their own subdivision. Although there were frequent screening detections from internal medicine patients, self-management was preferred in most cases rather than RRT activation. Greater discussion and education are needed to expand the range of RRT activities.

ffectiveness of the RRT and prefer to make decisions within their own subdivision. Although there were frequent screening detections from internal medicine patients, self-management was preferred in most cases rather than RRT activation. Greater discussion and education are needed to expand the range of RRT activities. In an Australian study, the overall in-hospital mortality of RRT patients was about 25%, compared with 15% in those not limited to medical therapy [10,20]. Another international prospective study of RRT showed a ward mortality rate of 11% [21]. In addition, age, national early warning score, and care limitations were significant predictors of mortality in a multivariable logistic regression. In our study, the overall mortality rate was 16.4%. The presence of malignancy, APACHE-II score, and number of days between admission and RRT activation were significant factors in survival.

tional early warning score, and care limitations were significant predictors of mortality in a multivariable logistic regression. In our study, the overall mortality rate was 16.4%. The presence of malignancy, APACHE-II score, and number of days between admission and RRT activation were significant factors in survival. There were some limitations to this study. First, it used a retrospective, single center design. It is therefore difficult to generalize our results. However, our study provides more information on RRTs and could serve as a useful reference for modifying the RRT to fit the particular situation of each hospital. Second, although the RRT dose in our center is increasing annually, the RRT dose was lower than expected. A successful RRT system requires more than 25 calls per 1,000 admissions and a low call rate is known to be a key reason for failure [3,10,22]. To solve this problem, our RRT system was changed to a 24-hour system, and the range of RRT activations was expanded to all departments except pediatrics. As a result, the total number of RRT activation rose 57% compared with the same period a year earlier. Third, we categorized the patients with the presence of malignancy and status of postoperation. However, these classification may lead to selection bias due to diversity of type and severity. Finally, our institution’s RRT records mostly concern postoperative patients. Further cooperation is needed with the subdivisions of the Department of Internal Medicine to improve the system.

ncy and status of postoperation. However, these classification may lead to selection bias due to diversity of type and severity. Finally, our institution’s RRT records mostly concern postoperative patients. Further cooperation is needed with the subdivisions of the Department of Internal Medicine to improve the system. In conclusion, this study delineates our experience of RRT implementation in a single tertiary medical center. Patient screening before a severe adverse event was an important factor in the outcome. The clinical parameters that were related to survival in this study can be used in RRT risk assessments. Further studies and efforts are needed to improve the quality of the current RRT system and achieve greater benefits. No potential conflict of interest relevant to this article was reported. Figure 1. Reasons for rapid response team activation. There are 12 activation criteria, including a direct phone call for serious concerns about overall deterioration as detected by a physician, nurse, or caregivers at the bedside. Criteria can be duplicated. Figure 2. Epidemiology of rapid response team (RRT) activations. RRT activation according to (A) position, (B) department, (C) RRT activation by the specific day of the week according to the surgical status and (D) activation methods. OS: orthopedics; OBGY: obstetrics and gynecology; URO: urology; PS: plastic surgery; IM: internal medicine; NP: neuropsychiatry; ENT: otorhinolaryngology; DT: dentistry; GS: general surgery. Table 1. Basic characteristics of RRT activation

Figure 2. Epidemiology of rapid response team (RRT) activations. RRT activation according to (A) position, (B) department, (C) RRT activation by the specific day of the week according to the surgical status and (D) activation methods. OS: orthopedics; OBGY: obstetrics and gynecology; URO: urology; PS: plastic surgery; IM: internal medicine; NP: neuropsychiatry; ENT: otorhinolaryngology; DT: dentistry; GS: general surgery. Table 1. Basic characteristics of RRT activation Characteristic Total (n = 287) Survival Activation method Survived (n = 240) Expired (n = 47) P-value Phone call (n = 215) Screening (n = 72) P-value Male sex 124 (43.2) 105 (43.8) 19 (40.4) 0.795 97 (45.1) 27 (37.5) 0.321 Age (yr) 70.0 (58.0–78.0) 71.0 (58.0–78.0) 64.0 (54.5–74.5) 0.088 69.0 (58.5–78.0) 71.0 (56.0–77.5) 0.799 BMI (kg/m2) 23.0 (20.1–25.8) 23.0 (20.1–25.8) 22.9 (20.5–24.9) 0.784 23.0 (20.0–25.8) 22.9 (20.6–25.7) 0.495 Admission route 0.129 0.398 OPD 160 (56.1) 129 (54.0) 31 (67.4) 116 (54.5) 44 (61.1) ER 125 (43.9) 110 (46.0) 15 (32.6) 97 (45.5) 28 (38.9) Malignancy 100 (34.8) 67 (28.0) 33 (70.2) <0.001a 76 (35.5) 24 (33.3) 0.847 Status of postoperation 199 (69.3) 172 (71.7) 26 (55.3) 0.041a 145 (67.4) 53 (73.6) 0.405 Postoperation days at RRT activationb 2.0 (1.0–5.0) 2.0 (1.0–5.0) 3.0 (2.0–8.0) 0.012a 3.0 (1.0–5.5) 2.0 (1.0–3.0) 0.223 Response time (min) 8.6 (5.6–11.6) 9.0 (6.0–11.0) 10.0 (6.5–14.0) 0.375 9.0 (6.0–12.0) 10.5 (0.0–21.0) 0.927 MEWS 5.0 (4.0–7.0) 5.0 (4.0–7.0) 7.0 (5.0–10.0) <0.001a 5.0 (4.0–8.0) 5.0 (4.0–7.0) 0.656 APACHE-II scorec 14.0 (10.0–18.0) 13.0 (9.0–18.0) 16.5 (13.0–23.0) <0.001a 14.0 (10.0–18.0) 13.0 (8.0–18.0) 0.208 RRT activation during weekday except holidays 229 (79.8) 195 (81.2) 34 (72.3) 0.233 175 (81.4) 54 (75.0) 0.317 Days between admission to RRT activation 6.0 (3.0–13.0) 5.0 (3.0–11.0) 11.0 (5.0–25.5) <0.001a 6.0 (3.0–13.0) 5.0 (2.0–13.5) 0.622 Values are presented as number (%) or median (interquartile range).

8.0) 0.208 RRT activation during weekday except holidays 229 (79.8) 195 (81.2) 34 (72.3) 0.233 175 (81.4) 54 (75.0) 0.317 Days between admission to RRT activation 6.0 (3.0–13.0) 5.0 (3.0–11.0) 11.0 (5.0–25.5) <0.001a 6.0 (3.0–13.0) 5.0 (2.0–13.5) 0.622 Values are presented as number (%) or median (interquartile range). RRT: rapid response team; BMI: body mass index; OPD: outpatient department; ER: emergency room; MEWS: modified early warning score; APACHE-II: Acute Physiology and Chronic Health Evaluation-II. a P-values are significant at the 0.05 level; b Postoperation days were calculated with patients after operation; c Missing data (n = 41) was excluded. Table 2. Outcome after RRT activation Characteristic Total (n = 287) Survival Activation method Survived (n = 240) Expired (n = 47) P-value Phone call (n = 215) Screening (n = 72) P-value At RRT activation 0.965 0.123 GW 275 (95.8) 230 (95.8) 45 (95.7) 203 (94.4) 72 (100.0) ICU 7 (2.4) 6 (2.5) 1 (2.1) 7 (3.3) 0 Other 5 (1.7) 4 (1.6) 1 (2.1) 5 (2.3) 0 After RRT activation 0.093 0.017a GW 193 (67.2) 167 (69.6) 26 (55.3) 135 (62.8) 58 (80.6) ICU 91 (31.7) 70 (29.2) 21 (44.7) 77 (35.8) 14 (19.4) Other 3 (1.0) 3 (1.2) 0 3 (1.4) 0 Intubation 18 (6.3) 11 (4.6) 7 (14.9) 0.019a 17 (7.9) 1 (1.4) 0.090 Survival 240 (83.6) - - - 179 (83.3) 61 (84.7) 0.915 Hospital stay days after activation 12.0 (6.0–25.5) 12.5 (7.0–26.5) 7.0 (1.0–22.0) 0.008a 13.0 (7.0–31.0) 11.0 (5.5–16.0) 0.022a Values are presented as number (%) or median (interquartile range). RRT: rapid response team; GW: general ward; ICU: intensive care unit.

Characteristic Total (n = 287) Survival Activation method Survived (n = 240) Expired (n = 47) P-value Phone call (n = 215) Screening (n = 72) P-value At RRT activation 0.965 0.123 GW 275 (95.8) 230 (95.8) 45 (95.7) 203 (94.4) 72 (100.0) ICU 7 (2.4) 6 (2.5) 1 (2.1) 7 (3.3) 0 Other 5 (1.7) 4 (1.6) 1 (2.1) 5 (2.3) 0 After RRT activation 0.093 0.017a GW 193 (67.2) 167 (69.6) 26 (55.3) 135 (62.8) 58 (80.6) ICU 91 (31.7) 70 (29.2) 21 (44.7) 77 (35.8) 14 (19.4) Other 3 (1.0) 3 (1.2) 0 3 (1.4) 0 Intubation 18 (6.3) 11 (4.6) 7 (14.9) 0.019a 17 (7.9) 1 (1.4) 0.090 Survival 240 (83.6) - - - 179 (83.3) 61 (84.7) 0.915 Hospital stay days after activation 12.0 (6.0–25.5) 12.5 (7.0–26.5) 7.0 (1.0–22.0) 0.008a 13.0 (7.0–31.0) 11.0 (5.5–16.0) 0.022a Values are presented as number (%) or median (interquartile range). RRT: rapid response team; GW: general ward; ICU: intensive care unit. a P-values are significant at the 0.05 level. Table 3. Clinical parameters associated with survival

Characteristic Total (n = 287) Survival Activation method Survived (n = 240) Expired (n = 47) P-value Phone call (n = 215) Screening (n = 72) P-value At RRT activation 0.965 0.123 GW 275 (95.8) 230 (95.8) 45 (95.7) 203 (94.4) 72 (100.0) ICU 7 (2.4) 6 (2.5) 1 (2.1) 7 (3.3) 0 Other 5 (1.7) 4 (1.6) 1 (2.1) 5 (2.3) 0 After RRT activation 0.093 0.017a GW 193 (67.2) 167 (69.6) 26 (55.3) 135 (62.8) 58 (80.6) ICU 91 (31.7) 70 (29.2) 21 (44.7) 77 (35.8) 14 (19.4) Other 3 (1.0) 3 (1.2) 0 3 (1.4) 0 Intubation 18 (6.3) 11 (4.6) 7 (14.9) 0.019a 17 (7.9) 1 (1.4) 0.090 Survival 240 (83.6) - - - 179 (83.3) 61 (84.7) 0.915 Hospital stay days after activation 12.0 (6.0–25.5) 12.5 (7.0–26.5) 7.0 (1.0–22.0) 0.008a 13.0 (7.0–31.0) 11.0 (5.5–16.0) 0.022a Values are presented as number (%) or median (interquartile range). RRT: rapid response team; GW: general ward; ICU: intensive care unit. a P-values are significant at the 0.05 level. Table 3. Clinical parameters associated with survival Variable Univariate analysis Multivariate analysisa Odds ratio (95% CI) P-value Odds ratio (95% CI) P-value Admission route 1.76 (0.92–3.51) 0.096 1.06 (0.46–2.49) 0.889 Malignancy 6.05 (3.11–12.35) <0.001b 7.47 (3.00–20.53) <0.001b Status of postoperation 0.49 (0.26–0.93) 0.029b 0.93 (0.39–2.27) 0.874 MEWS 1.27 (1.13–1.43) <0.001b 0.95 (0.77–1.15) 0.583 APACHE-II score 1.11 (1.05–1.17) <0.001b 1.10 (1.03–1.19) 0.009b RRT activation during weekday except holiday 1.66 (0.79–3.33) 0.167 0.91 (0.31–2.44) 0.862 Days between admission to RRT activation 1.04 (1.02–1.06) <0.001b 1.04 (1.02–1.07) 0.001b Type of ward after activation 1.65 (0.90–2.98) 0.102 1.96 (0.79–4.79) 0.143 Intubation 3.64 (1.27–9.82) 0.012b 1.35 (0.21–7.03) 0.731 CI: confidence interval; MEWS: modified early warning score; APACHE-II: Acute Physiology and Chronic Health Evaluation-II; RRT: rapid response team.

1.06) <0.001b 1.04 (1.02–1.07) 0.001b Type of ward after activation 1.65 (0.90–2.98) 0.102 1.96 (0.79–4.79) 0.143 Intubation 3.64 (1.27–9.82) 0.012b 1.35 (0.21–7.03) 0.731 CI: confidence interval; MEWS: modified early warning score; APACHE-II: Acute Physiology and Chronic Health Evaluation-II; RRT: rapid response team. a Clinical parameters which showed P-value, 0.2 at univariate logistic regression were included for multivariate logistic regression;level. b P-values are significant at the 0.05 level.

Dear Editor: I have read the article entitled “A Pilot Survey of Difficult Intubation and Cannot Intubate, Cannot Ventilate Situations in Korea” published by Kim et al. [1] in the Korean Journal of Critical Care Medicine in August 2016, with great interest. An official survey on difficult intubations is still a very meaningful pilot study in Korea. The authors suggested that the video laryngoscope is the most preferred modality among Korean anesthesiologists and intensivists for “Cannot Intubate, Cannot Ventilate (CICV)” and difficult intubation conditions. This preference reflects the results from a 2013 report of a survey performed in Canada [2]. I believe that these findings are valuable and should be actively applied in special conditions, such as in the intensive care unit (ICU).

ists for “Cannot Intubate, Cannot Ventilate (CICV)” and difficult intubation conditions. This preference reflects the results from a 2013 report of a survey performed in Canada [2]. I believe that these findings are valuable and should be actively applied in special conditions, such as in the intensive care unit (ICU). In general, patients in the ICU exhibit signs and symptoms of acute respiratory distress syndrome and sepsis. These patients lack a physiologic reserve when compared to other patients, and are often facing life-threatening conditions. Moreover, a difficult airway occurs more often outside the operating room, while the rate of incidence is 11% to 22% in critically ill patients [3,4]. Therefore, when compared to patients who undergo tracheal intubation for elective surgery in the operating room [1], it is more important to accurately predict a difficult airway and be successful on the first attempt at intubation in ICU patients [5]. Thus, if a difficult airway condition is predicted in ICU patients, it is necessary to proactively use the video laryngoscope on the first intubation attempt [6]. A prior prospective study reports that the use of the C-MAC® video laryngoscope (Karl Storz, Tuttlingen, Germany), over the Macintosh blade, has increased the success rate of the first attempt at tracheal intubation in ICU patients suspected of having a difficult airway from 55% to 79% [7]. In light of these trends, intensive care staff in Korea should consider proactively assessing for difficult airways and using a video laryngoscope before a CICV situation ensues.

ased the success rate of the first attempt at tracheal intubation in ICU patients suspected of having a difficult airway from 55% to 79% [7]. In light of these trends, intensive care staff in Korea should consider proactively assessing for difficult airways and using a video laryngoscope before a CICV situation ensues. The first point to consider is that the survey in the study by Kim et al. [1] was based on the clinicians’ individual experiences and preferences. A prospective study, hence, is needed for achieving more objective outcomes. In addition, the outcomes of this study were limited to unanticipated difficult airway and CICV situations, not anticipated difficult airways. The results could be different when applying them to anticipated difficult airways, and this should be considered in future studies.

eeded for achieving more objective outcomes. In addition, the outcomes of this study were limited to unanticipated difficult airway and CICV situations, not anticipated difficult airways. The results could be different when applying them to anticipated difficult airways, and this should be considered in future studies. Furthermore, there is no comment on whether the clinical situation was managed using a team approach. A team approach is an important issue because when attempting tracheal intubation in critically ill patients suspected of having a difficult airway, cooperation among multiple experts, including intensivists and surgeons, is needed [5]. Despite the ongoing debate about the actual effectiveness of the rapid response team in management of difficult airway patients, a recent report focused on the introduction of the difficult airway response team (DART), which is comprised of anesthesiologists, otolaryngologists, trauma surgeons, and emergency medicine physicians, and their role in securing the airway in difficult airway patients [8]. More specifically, the role of otolaryngologists in the DART is becoming more significant and these changes are likely to occur in Korea in the near future [9]. When faced with situations such as dealing with a difficult airway or CICV, a team approach provides the clinicians with opportunities to make judgments that may be different from what they would have made working alone.

is becoming more significant and these changes are likely to occur in Korea in the near future [9]. When faced with situations such as dealing with a difficult airway or CICV, a team approach provides the clinicians with opportunities to make judgments that may be different from what they would have made working alone. In conclusion, the survey by Kim et al. [1] is a valuable study that shows, for the first time, the preference of Korean anesthesiologists and intensivists for CICV and difficult airway patients. However, as mentioned before, there were limitations, including the retrospective nature of the study and difficulties in immediately applying the findings clinically. Therefore, in the near future, a multicenter retrospective study with a large sample size in a clinical setting in Korea, as well as a well-designed prospective study, are needed. Furthermore, additional research on programs such as DART, which directly predicts the possibility of a difficult airway in a critically ill patient and allows experts of multiple disciplines to collaborate to solve a problem, should be performed. No potential conflict of interest relevant to this article was reported.

Mycoplasma pneumoniae (MP) is a common respiratory pathogen in children. MP infections usually have a mild clinical course, and pneumonia is the most prominent clinical manifestation. Several types of neurologic disorders have been reported to be extrapulmonary manifestations in 0.1% of mycoplasma infections, including meningo-encephalitis, transverse myelitis, seizures, acute disseminated encephalomyelopathy, Guillain-Barre syndrome, cerebral ataxia, and stroke [1]. To date, only 12 cases of MP-associated strokes in children have been reported in the literature [2-12]. In most of these cases, the patients partially or fully recovered after treatment with antibiotics, intravenous immunoglobulin, steroids, or aspirin. In two of the cases, the children died despite medical treatment. Herein, we report the case of an extensive and progressive acute cerebral infarction that occurred after MP infection; the patient not only received medical managements but also underwent a decompressive craniectomy.

Mycoplasma pneumoniae (MP) is a common respiratory pathogen in children. MP infections usually have a mild clinical course, and pneumonia is the most prominent clinical manifestation. Several types of neurologic disorders have been reported to be extrapulmonary manifestations in 0.1% of mycoplasma infections, including meningo-encephalitis, transverse myelitis, seizures, acute disseminated encephalomyelopathy, Guillain-Barre syndrome, cerebral ataxia, and stroke [1]. To date, only 12 cases of MP-associated strokes in children have been reported in the literature [2-12]. In most of these cases, the patients partially or fully recovered after treatment with antibiotics, intravenous immunoglobulin, steroids, or aspirin. In two of the cases, the children died despite medical treatment. Herein, we report the case of an extensive and progressive acute cerebral infarction that occurred after MP infection; the patient not only received medical managements but also underwent a decompressive craniectomy. Case Report A previously healthy 5-year-old boy presented with a productive cough and a fever for 3 days; he was found to have pneumonia and admitted to the local hospital. On the basis of the positive result of the mycoplasma-specific IgM antibody, he was treated with intravenous clarithromycin for 3 days, starting on the day of admission. His symptoms and chest radiographic findings progressively worsened despite an additional 5 days of antibiotic treatment with cefotaxime and vancomycin. During the 4 days before his transfer to our hospital, he was treated with methyl-prednisolone (2 mg/kg/d), and his fever disappeared and his chest radiography findings improved slightly. On the day of his transfer, he became drowsy, and exhibited dysarthria, limited eye movement, and irritability. He experienced a generalized tonic seizure that lasted for 3 hours despite receiving treatment with anticonvulsants including intravenous diazepam, fosphenytoin, and phenobarbital. Because of newly developed fever and neurologic symptoms, a cerebrospinal fluid (CSF) analysis and brain magnetic resonance imaging (MRI) were conducted. An acute cerebral infarction was suspected, and the patient was transferred to our hospital while receiving treatment for the management of increased intracranial pressure (IICP), including dexamethasone, mannitol, and continuous midazolam infusion. His medical and family histories were unremarkable.

I) were conducted. An acute cerebral infarction was suspected, and the patient was transferred to our hospital while receiving treatment for the management of increased intracranial pressure (IICP), including dexamethasone, mannitol, and continuous midazolam infusion. His medical and family histories were unremarkable. When the patient arrived at our hospital, he had a temperature of 37.8℃, respiratory rate of 27 breaths/min, pulse rate of 114 beats/min, and blood pressure of 108/64 mmHg. He was receiving 4 L/min of oxygen through a nasal cannula and his oxygen saturation was 97%. His breath sounds were coarse but he had no signs of respiratory distress. He was stuporous; however, his pupils were reactive to light and isocoric. His deep tendon reflexes were normal, and he had no neck rigidity. His motor and sensory functions were not evaluated because he was sedated.

s oxygen saturation was 97%. His breath sounds were coarse but he had no signs of respiratory distress. He was stuporous; however, his pupils were reactive to light and isocoric. His deep tendon reflexes were normal, and he had no neck rigidity. His motor and sensory functions were not evaluated because he was sedated. The laboratory examinations revealed a hemoglobin level of 10.6 g/dl, leukocyte counts of 1,260/mm3 with 82.3% neutrophils and 11.9% lymphocytes, and platelet count of 115,000/mm3. His liver function tests revealed a slightly decreased albumin level of 3.2 g/dl, an elevated aspartate aminotransferase level of 53 IU/L, and an alanine aminotransferase level of 94 IU/L. His serum biochemistry results were all within the normal limits except for an increased serum lactate dehydrogenase level of 747 IU/L (reference range, 100 to 225 IU/L). The coagulation studies revealed a normal prothrombin time and activated partial thromboplastin time, mildly decreased fibrinogen level of 129 mg/dl (reference range, 180 to 380 mg/dl), and elevated D-dimer level of 6.84 μg/ml (reference, <0.4 μg/ml). The arterial blood gas results were as follows: pH, 7.43; PaO2, 90 mmHg; and PaCO2, 37 mmHg. The C-reactive protein level was 0.79 mg/dl (reference, <0.5 mg/dl), erythrocyte sedimentation rate 2 mm/h (reference < 9 mm/h), and procalcitonin level 0.124 ng/ml (reference < 0.5 ng/ml). The additional coagulation studies revealed that the patient’s protein C and antithrombin III levels were nearly normal; however, his protein S activity was decreased to 59% (reference range, 73% to 150%). Test results for anticardiolipin, antiglycoprotein, antinuclear, anti-DNA, anti-neutrophil cytoplasmic antibodies, as well as rheumatic factor were negative. His serum homocysteine concentration was 3.4 μM (reference range, 5 to 15 μM). Polymerase chain reaction (PCR) analysis of the methylenetetrahydrofolate reductase (MTHFR) gene showed a heterozygous MTHFR A1298C mutation. Serum Mycoplasma antibody levels were elevated to 1:20,480 on indirect particle agglutination, and PCR of the respiratory specimen was positive for Mycoplasma.

ange, 5 to 15 μM). Polymerase chain reaction (PCR) analysis of the methylenetetrahydrofolate reductase (MTHFR) gene showed a heterozygous MTHFR A1298C mutation. Serum Mycoplasma antibody levels were elevated to 1:20,480 on indirect particle agglutination, and PCR of the respiratory specimen was positive for Mycoplasma. The patient’s chest radiogram demonstrated consolidation and atelectasis in the right upper and left lower lungs (Figure 1). The electrocardiography and echocardiography findings were normal. MRI performed at the previous hospital was suspicious for bilateral acute ischemic stroke in posterior circulation (Figure 2). The right vertebral and basilar arteries were not visualized on magnetic resonance angiography (MRA) (Figure 3).

s (Figure 1). The electrocardiography and echocardiography findings were normal. MRI performed at the previous hospital was suspicious for bilateral acute ischemic stroke in posterior circulation (Figure 2). The right vertebral and basilar arteries were not visualized on magnetic resonance angiography (MRA) (Figure 3). On the day of admission, mannitol and dexamethasone were continuously infused for IICP management. The antibiotic regimen was changed to vancomycin, cefotaxime, and levofloxacin and acyclovir was added. On the second hospital day after transfer to our facility, the patient’s seizures worsened despite increased doses of midazolam. He suddenly became apneic and intubation and mechanical ventilator support were required. Anti-inflammatory therapy included methylprednisolone at 30 mg/kg for 3 days and intravenous immunoglobulin (IVIG) at 0.5 g/kg for 4 days. Because of the deterioration in his neurologic condition, he underwent repeat brain MRI and MRA, which revealed further multifocal narrowing in the bilateral distal internal carotid arteries and progressive swelling of the right cerebellum with brain stem compression. To avert impending herniation, the patient underwent craniotomy, bilateral suboccipital craniectomy, and insertion of external ventricular drainage for IICP control. Analysis of the CSF showed a leukocyte count of 0/μL, glucose level of 105 mg/dl, and protein level of 13 mg/dl. The PCR of the CSF was negative for Mycoplasma. On the 5th day after the transfer to our hospital, the patient’s neurologic status was stable; however, the follow-up brain MRI and MRA revealed a slight interval progression of the T2 high signal intensity in the right pons and a multifocal narrowing in the proximal middle cerebral artery (Figure 4). As a result, enoxaparin was added to control the progression of the brain lesion. Seven days after admission, the consolidation was shown to be resolved on a follow-up chest radiography. His neurologic status improved, and he started breathing spontaneously, including spontaneous eye blinking and withdrawal of painful stimuli. He received bedside physiotherapy for rehabilitation. On the 13th day of admission, a tracheostomy was performed because his cough reflex remained weak. On the 25th hospital day, enoxaparin was changed to warfarin, and the patient was transferred to the general ward. Despite active rehabilitation, he remained bed-ridden when he was discharged to home 2 months later.

tion. On the 13th day of admission, a tracheostomy was performed because his cough reflex remained weak. On the 25th hospital day, enoxaparin was changed to warfarin, and the patient was transferred to the general ward. Despite active rehabilitation, he remained bed-ridden when he was discharged to home 2 months later. Discussion We report the case of a child with extensive and progressive acute cerebral artery occlusions after an MP infection, requiring decrompressive craniectomy. MP is an important pathogen of respiratory disease and has been reported in 10-40% of cases of community-acquired pneumonia in children [13]. Since the 2000s, a high prevalence of macrolide-resistant MP (MRMP) strains has been detected: as high as 90% in Asia, Europe, and the United States [13,14]. Tetracyclines and fluoroquinolones can be used as alternative to macrolides in the treatment of MRMP; however, their use in children is limited because of their adverse effects on bone and cartilage development [15]. In our case report, on the basis of the suspicion that the patient had an MRMP infection, clarithromycin was changed to levofloxacin; no adverse effects of the drug were observed.

in the treatment of MRMP; however, their use in children is limited because of their adverse effects on bone and cartilage development [15]. In our case report, on the basis of the suspicion that the patient had an MRMP infection, clarithromycin was changed to levofloxacin; no adverse effects of the drug were observed. The pathogenesis of refractory or severe MP infections is considered to be closely associated with an excessive immune response, such as a highly activated cellmediated immune response and the expression of proinflammatory cytokines such as interleukin (IL)-1, IL-2, IL-4, IL-6, interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α) [16]. Furthermore, several authors, including Miyashita et al. [17], determined that immunosuppressive therapy has a profound beneficial effect by reducing the immune-mediated pulmonary injury seen in MP infection. After the 5-year-old patient was transferred to our hospital, the antibiotics regimen was changed and he received high-dose methyl-prednisolone. The combination of the appropriate antibiotics and the immunemodulating therapies may have improved the pneumonia. As MRMP cases have been increasing rapidly in recent years worldwide, especially in younger children, appropriate treatment with antibiotics and immune-modulating agents should be considered, and further investigation are needed to assess for the adverse effects of drugs.

ting therapies may have improved the pneumonia. As MRMP cases have been increasing rapidly in recent years worldwide, especially in younger children, appropriate treatment with antibiotics and immune-modulating agents should be considered, and further investigation are needed to assess for the adverse effects of drugs. Various extrapulmonary manifestations have been reported in association with MP infection. Neurological impairments are the most common manifestations, including meningoencephalitis, acute disseminated encephalomyelopathy, transverse myelitis, seizures, peripheral nerve involvement, and stroke [1]. To date, a total of 12 children with MP-associated strokes have been reported in the literature [2-12]. Ten of the 12 cases involved an arterial occlusion in the anterior circulation of brain, and one case involved a posterior circulation stroke [10]. Occlusions of both the anterior and posterior circulation after an MP infection, like in our case, were reported in one 4-year-old patient, which resulted in death [5]. However, there are no previously reported cases of a rapidly progressing cerebral infarction, regardless of the specific brain regions, after an MP infection. Unlike the previously reported severe cases, we found that the stroke associated with the MP infection proceeded rapidly because of vascular narrowing and brain swelling; these findings were confirmed on consecutive MRI and MRA imaging of the brain.

tion, regardless of the specific brain regions, after an MP infection. Unlike the previously reported severe cases, we found that the stroke associated with the MP infection proceeded rapidly because of vascular narrowing and brain swelling; these findings were confirmed on consecutive MRI and MRA imaging of the brain. The mechanisms of stroke associated with MP infections are well understood. Direct neurological invasion, hypercoagulable or thrombotic states, and vasculitis have all been suggested as possible etiologies for these strokes [1,5]. Assessing MP-specific antibodies and performing PCR of the CSF are useful for detecting direct central nervous system invasion. In our case, the PCR analyses of the CSF and brain tissue were negative for MP, and antibodies were not examined. A total of 6 of the 12 reported cases of stroke associated with MP infection had hypercoagulable and thrombotic conditions confirmed with laboratory tests [3,5-7,10,11]. Among these cases, one patient had a genetic defect in the MTHFR gene with hyperhomocysteinemia, which caused thrombosis, and one child had a sickle cell trait. Our patient was hypercoagulable with decreased protein S activity and an elevated D-dimer level. Although the heterozygous A1298C mutation of the MTHFR gene was discovered in our case, the causality between MTHFR gene polymorphism and cerebral infarction is weak. Compared with those in previous studies [18], our patient showed low levels of homocysteine and only one heterozygosity of C677T and A1298C mutations was found. The last possible mechanism is vasculitis, which results from direct invasion of the vascular wall with an autoimmune process or vascular inflammation by pro-inflammatory cytokines such as IL-6, IFN-γ, and TNF [19]. Brain biopsy is considered the gold standard for the diagnosis of central nervous system vasculitis; however, it is not always feasible. Our patient did not show elevation of acute-phase reactants, and compatible pathologic findings. However, we suspected that the anterior circulation infarction was due to vasculitis because multiple and bilateral stenotic vascular lesions were found on MRA. Furthermore, as both the neurologic condition and pneumonia were simultaneously deteriorated and improved, a systemic proinflammatory process was suggested to have affected the clinical course including progression of the cerebral infarction.

s because multiple and bilateral stenotic vascular lesions were found on MRA. Furthermore, as both the neurologic condition and pneumonia were simultaneously deteriorated and improved, a systemic proinflammatory process was suggested to have affected the clinical course including progression of the cerebral infarction. Finally, in our case, the cerebral infarction appeared to be the result of the combined effects of hypercoagulability and vascular inflammatory injury.

s because multiple and bilateral stenotic vascular lesions were found on MRA. Furthermore, as both the neurologic condition and pneumonia were simultaneously deteriorated and improved, a systemic proinflammatory process was suggested to have affected the clinical course including progression of the cerebral infarction. Finally, in our case, the cerebral infarction appeared to be the result of the combined effects of hypercoagulability and vascular inflammatory injury. The management of MP-associated stroke remains controversial. In the previously reported cases, the patients received antibiotics (macrolides in most cases), anti-inflammatory medications such as steroids and IVIG, or aspirin [2-12]. To date, there has been no consensus about the optimal treatment strategy. Among the 12 prior stroke cases, the patient with extensive cerebral involvement received antibiotics and low-dose aspirin based on the results of the hypercoagulability laboratory tests [5]. However, the patient did not receive anti-inflammatory medications, and ultimately died. Our patient received a combination of antibiotics, anticoagulants, and anti-inflammatory medications including methyl-prednisolone and IVIG. IVIG has not been used extensively in cerebral vasculitis; however, IVIG was infused for its anti-inflammatory effect through the modulation of cytokine antagonists and cytokine production, as previously reported [20]. We cannot specify which treatment affected the progression of the disease and improved his clinical condition, nor can we exclude the possibility that the cause of the stroke improved naturally. However, as several potential mechanisms could cause a stroke associated with MP infection, treatment with anti-inflammatory agents and antibiotics should be considered early, and should be based on the results of imaging and laboratory tests.

we exclude the possibility that the cause of the stroke improved naturally. However, as several potential mechanisms could cause a stroke associated with MP infection, treatment with anti-inflammatory agents and antibiotics should be considered early, and should be based on the results of imaging and laboratory tests. To summarize, we report a case of an extensive and rapidly progressive cerebral infarction after MP infection that possibly resulted from acquired hypercoagulability and cytokine-induced vascular inflammation. Clinicians should be cautious of the neurologic signs or symptoms when MP infections are especially severe and refractory, and multidisciplinary management may be necessary to improve the pulmonary and neurologic outcomes of MP infection that could otherwise proceed to death. No potential conflict of interest relevant to this article was reported. Figure 1. Chest radiogram on the day of admission shows consolidation and atelectasis in the right upper and left lower lung. Figure 2. Magnetic resonance imaging (MRI) performed at the previous hospital on the day of admission. Axial section of diffusion-weighted MRI showing an area of restricted diffusion in the left thalamus (A), pons, and bilateral cerebellum (arrow) (B). Axial section of the apparent diffusion map showing a corresponding low apparent diffusion coefficient value in the left thalamus (arrow) (C), pons, and bilateral cerebellum (arrow) (D).

diffusion-weighted MRI showing an area of restricted diffusion in the left thalamus (A), pons, and bilateral cerebellum (arrow) (B). Axial section of the apparent diffusion map showing a corresponding low apparent diffusion coefficient value in the left thalamus (arrow) (C), pons, and bilateral cerebellum (arrow) (D). Figure 3. Magnetic resonance angiography (MRA) performed at the previous hospital on the day of admission. MRA image showing a filling defect (A) in the right vertebral artery (arrow) and (B) basilar artery (arrow) suggestive of thrombosis. Figure 4. Magnetic resonance angiography image showing multifocal narrowing in both the middle cerebral artery and internal carotid artery (arrow) suggestive of vasculitis.

Introduction Mucormycosis is an emerging infectious disease caused by fungi of the order Mucorales and represents the second leading cause of invasive fungal infection, following aspergillosis [1,2]. It is known to develop mainly in patients with diabetes, hematologic malignancy, organ transplantation, and those receiving immunosuppressive therapy [3]. The clinical presentation commonly results from involvement of rhinocerebral disease or pulmonary infection [4]. It was reported that Mucorales fungi most commonly affects the lungs in patients with hematological malignancies [5]. Pulmonary mucormycosis (PM) is a life-threatening infection with high mortality of over 70% [6]. PM has recently emerged as patients with hematological malignancies or transplant patients have been on antifungal prophylaxis with Aspergillus-active but Mucorales-inactive agents [7]. Although its incidence was rare, reported to be 1.7 cases per million people/year by a population-based study [8], the incidence of PM is expected to increase because more individuals may be susceptible to invasive fungal infections resulting from either immunosuppressive disease or therapy with advances in modern medicine and because diagnosis of PM is difficult and underreported [4]. In these situations, more attention to PM may be needed in patients receiving intensive care, particularly those with immunocompromised status.

ve fungal infections resulting from either immunosuppressive disease or therapy with advances in modern medicine and because diagnosis of PM is difficult and underreported [4]. In these situations, more attention to PM may be needed in patients receiving intensive care, particularly those with immunocompromised status. Unlike pulmonary aspergillosis, the clinical outcomes of PM have not improved significantly over the last decade, mainly because early diagnosis of PM is difficult and antifungal agents against Mucorales show limited activity [7]. Moreover, there are limited data differentiating computed tomography (CT) findings for PM from invasive pulmonary aspergillosis [9]. Although diagnostic methods for PM, such as fiber optic bronchoscopy with bronchoalveolar lavage or image-guided fine needle aspiration cytology, have been reported [10,11], the accurate diagnosis of PM in patients in the intensive care unit remains a challenge. At present, early and accurate diagnosis of PM is a very important issue for improving the therapeutic effect. We evaluated the clinical manifestations, imaging features, diagnosis, treatment, and prognosis of PM in a Korean tertiary hospital and identified the role of transbronchial lung biopsy (TBLB) in the diagnosis of PM in patients admitted to the intensive care unit. Materials and Methods This study was performed at the a tertiary-care teaching hospital in South Korea. The medical records of adult patients (aged 16 years and older) who met the criteria for proven or probable PM were retrospectively reviewed from January 2003 to December 2013.

At present, early and accurate diagnosis of PM is a very important issue for improving the therapeutic effect. We evaluated the clinical manifestations, imaging features, diagnosis, treatment, and prognosis of PM in a Korean tertiary hospital and identified the role of transbronchial lung biopsy (TBLB) in the diagnosis of PM in patients admitted to the intensive care unit. Materials and Methods This study was performed at the a tertiary-care teaching hospital in South Korea. The medical records of adult patients (aged 16 years and older) who met the criteria for proven or probable PM were retrospectively reviewed from January 2003 to December 2013. Proven and probable PM were defined as described by the revised criteria of the European Organization for Research and Treatment of Cancer/Mycosis Study Group [12]. Proven PM was defined by histological evidence of non-septate, right-angle branching filamentous fungi invading tissue plus culture positive for Mucorales species from pulmonary tissue or immunohistochemical staining positive for anti-Rhizopus arrhizus monoclonal antibody (LSBio, Seattle, WA, USA). Probable PM was defined as the presence of host factors plus one or more clinical indications. The host factors included uncontrolled diabetes, hematological malignancy, organ transplantation, and immunosuppressive therapy. Furthermore, the clinical indications included Mucorales growth in sputum or bronchoalveolar lavage fluid culture or CT finding such as dense well-circumscribed lesions with or without a halo sign or an air-crescent sign or cavity.

iabetes, hematological malignancy, organ transplantation, and immunosuppressive therapy. Furthermore, the clinical indications included Mucorales growth in sputum or bronchoalveolar lavage fluid culture or CT finding such as dense well-circumscribed lesions with or without a halo sign or an air-crescent sign or cavity. TBLB was performed on the lobe in which CT findings suspected involvement of PM in the intensive care unit via portable bronchoscopy. A chest X-ray and medical records after TBLB were reviewed to exclude pneumothorax and bleeding. Evaluation in each patient included (1) an underlying condition of immunocompromised state, (2) clinical presentations, (3) CT findings (unilateral or bilateral involvement, nodule, mass, consolidation, cavitation, ground glass opacity, air-fluid level, pleural effusion, and halo sign), (4) cytomorphological features and a final diagnosis of specimen, (5) complications following the diagnostic procedure, and (6) treatment response. The data are presented as medians (range) for continuous variables and as percentages for discrete variables. These variables were analyzed by using simple descriptive statistics. All patients gave their written informed consent for bronchoscopy and TBLB. Institutional review board of Samsung Medical Center approved the analyses of the clinical and TBLB data. Ethical approval for this study was exempted by the institutional review board (SMC 2017-02-023-002).

The data are presented as medians (range) for continuous variables and as percentages for discrete variables. These variables were analyzed by using simple descriptive statistics. All patients gave their written informed consent for bronchoscopy and TBLB. Institutional review board of Samsung Medical Center approved the analyses of the clinical and TBLB data. Ethical approval for this study was exempted by the institutional review board (SMC 2017-02-023-002). Results Characteristics of PM patients and their underlying diseases are summarized in Table 1. Of the nine patients, four were male. The median age was 64 years (range, 12 to 73 years). The underlying diseases at diagnosis of PM were hematologic malignancy (seven cases) and solid cancer (two cases). Three patients had diabetes mellitus, and nine patients had received chemotherapy prior to diagnosis of PM. PM was proven and probable in seven and two cases, respectively. CT findings of PM are summarized in Table 2. Unilateral involvement was shown in eight cases (89%), consolidation in eight cases (89%), and ground glass opacity in four cases. However, reverse halo sign was only shown in one case (11%). Fungal infections, such as invasive pulmonary aspergillosis or PM, suspected on CT by radiologists were present in three cases (33%). Four cases (44%) were suspected to have bacterial pneumonia on CT.

eight cases (89%), and ground glass opacity in four cases. However, reverse halo sign was only shown in one case (11%). Fungal infections, such as invasive pulmonary aspergillosis or PM, suspected on CT by radiologists were present in three cases (33%). Four cases (44%) were suspected to have bacterial pneumonia on CT. The mortality rate was 56% (five of nine cases) (Table 3). Pathologic findings of lung specimens by each diagnostic procedure showed seven cases (78%) compatible with mucormycosis and two cases (22%) suggesting fungal infections of invasive aspergillosis or mucormycosis. The median treatment period was 35 days (range, 3 to 154 days). Six of nine cases (67%) were diagnosed with PM from TBLB via portable bronchoscopy. There were no complications such as pneumothorax and hemorrhage after TBLB on chest radiography (CXR) and medical records. One case was diagnosed by percutaneous lung aspiration/biopsy, and two cases were diagnosed by surgical biopsy and resection. The condition of the two PM cases receiving surgical resection improved following administration of antifungal agents of liposomal amphotericin. None of the six cases diagnosed by TBLB could receive surgical treatment due to poor general condition, only one case (16.7%) improved after medical treatment. Discussion The present study shows that PM is associated with high mortality rates, which might be improved in cases receiving surgical treatment. Moreover, TBLB might be an early and feasible tool for diagnosing PM in suspected critical patients.

Six of nine cases (67%) were diagnosed with PM from TBLB via portable bronchoscopy. There were no complications such as pneumothorax and hemorrhage after TBLB on chest radiography (CXR) and medical records. One case was diagnosed by percutaneous lung aspiration/biopsy, and two cases were diagnosed by surgical biopsy and resection. The condition of the two PM cases receiving surgical resection improved following administration of antifungal agents of liposomal amphotericin. None of the six cases diagnosed by TBLB could receive surgical treatment due to poor general condition, only one case (16.7%) improved after medical treatment. Discussion The present study shows that PM is associated with high mortality rates, which might be improved in cases receiving surgical treatment. Moreover, TBLB might be an early and feasible tool for diagnosing PM in suspected critical patients. Little is known about the clinical presentations of PM. Most of the data showing clinical features of mucormycosis, including PM, has originated from case series and small-sized studies on specific patient groups such as patients with hematologic malignancies [5,10]. Known clinical manifestations of PM are summarized as follows: (1) underlying host risk factors were neutropenia, induction chemotherapy, hematopoietic stem cell transplantation with graft-versus-host disease, and lung transplantation [7]; (2) the pathogenesis of the disease state was hyphal invasion of pulmonary blood vessels, which could result in hemorrhage, thrombosis, ischemia, and infarction of distal tissue [7,13]; (3) clinical characteristics were prolonged high-grade fever (>38°C), nonproductive cough, airway obstruction from endobronchial or tracheal lesions, massive hemoptysis; and (4) the mortality rate was 66% or higher, depending on the level of immunosuppression [6,14]. To date, however, there is no clinical history specific for the diagnosis of PM.

were prolonged high-grade fever (>38°C), nonproductive cough, airway obstruction from endobronchial or tracheal lesions, massive hemoptysis; and (4) the mortality rate was 66% or higher, depending on the level of immunosuppression [6,14]. To date, however, there is no clinical history specific for the diagnosis of PM. It is typically difficult to suspect PM and make a diagnosis early because PM is characterized by non-specific clinical features and radiologic findings, a low culture positive rate, and a lack of biologic markers [4]. There were several reports that CT findings useful for the diagnosis of PM were progressive, homogeneous lobar or multilobar consolidation; nodules or mass-like or wedge-shaped consolidation; and a halo sign of ground glass opacity surrounding a pulmonary nodule [15,16]. On the other hand, Jung et al. [9] showed that the reverse halo sign helps the clinician to suspect PM and differentiate it from invasive pulmonary aspergillosis. A fungal culture of specimens obtained from a sputum culture, bronchoalveolar lavage culture, or percutaneous needle aspirate may be more helpful in the diagnosis of PM. Although the diagnosis of mucormycosis was recommended as proven, probable, or possible according to a host factor, a clinical criterion, and a mycological criterion [12], a definitive diagnosis of PM requires histologic identification of the organism, seen as mucoraceous hyphae in affected lung tissues [7,17]. Bronchoalveolar lavage, percutaneous needle aspiration, open lung biopsy, and pleural fluid culture are considered as diagnostic techniques to determine the affected lung. Of these techniques, the risk of surgery and the low incidence of pleural involvement limits the diagnostic value of open lung biopsy and pleural fluid culture. Moreover, it was recommended that isolation of the organism only from sputum or bronchial aspirate from bronchoalveolar lavage was not sufficient for a definite diagnosis [12,17], although it had been reported that analysis of an adequate bronchoalveolar lavage specimen from fiber optic bronchoscopy was a useful diagnostic method [10]. Recently, Sharma et al. [11] evaluated the role of fine needle aspiration cytology in the diagnosis of pulmonary infections in 42 immunocompromised patients and showed that it was a relatively reliable, safe, and rapid method of diagnosing pulmonary infection, including PM, in immunocompromised patients.

c method [10]. Recently, Sharma et al. [11] evaluated the role of fine needle aspiration cytology in the diagnosis of pulmonary infections in 42 immunocompromised patients and showed that it was a relatively reliable, safe, and rapid method of diagnosing pulmonary infection, including PM, in immunocompromised patients. In this study, we showed that TBLB might also be a relatively reliable and safe method of diagnosing PM in immunocompromised patients. Especially in the intensive care unit, TBLB could be a useful tool for diagnosing PM in suspected patients because TBLB is easier and more feasible than conducting portable bronchoscopy at the bedside.

e showed that TBLB might also be a relatively reliable and safe method of diagnosing PM in immunocompromised patients. Especially in the intensive care unit, TBLB could be a useful tool for diagnosing PM in suspected patients because TBLB is easier and more feasible than conducting portable bronchoscopy at the bedside. Treatment of PM is complicated because the organisms, Mucorales, are inherently resistant to antifungal agents [18,19], and PM demonstrates rapid clinical progression [7]. There are also limited data for antifungal susceptibility and available results of minimum inhibitory concentration testing. Moreover, it is difficult to use empirical antifungal agents for suspected patients with invasive fungal infections because Mucorales may express increased virulence following exposure to voriconazole [20], a widely used treatment for invasive pulmonary aspergillosis, which has a clinical presentation similar to that of PM. Therefore, at present, an early and accurate diagnosis is paramount in the treatment of PM. The main treatment strategy for PM is immediate surgical resection of the infected lung, such as wedge resection, lobectomy, or pneumonectomy, in combination with antifungal agents. These treatments have been associated with lower mortality rates in many reported cases [21-23]. Additionally, for medical therapy for PM, the only recommended antifungal agents are members of the polyene class, including amphotericin B deoxycholate and its lipid derivatives. Triazoles, posaconazole, isavuconazole, echinocandins, and combination therapy could be also considered [4]. In our study, two patients receiving surgical resection were alive after treatment. However, it may be considered that patients in good condition are more likely to receive surgery.

te and its lipid derivatives. Triazoles, posaconazole, isavuconazole, echinocandins, and combination therapy could be also considered [4]. In our study, two patients receiving surgical resection were alive after treatment. However, it may be considered that patients in good condition are more likely to receive surgery. This report has limitations due to its retrospective and non-controlled nature. Long-term observational and interventional controlled studies are needed for more advanced knowledge of PM. Moreover, clinicians should be concerned with complications of TBLB, though there were no cases of pneumothorax nor hemorrhage in this study. PM is an emerging invasive infection and is often fatal, and its successful treatment relies on a timely diagnosis. TBLB can be an easy and useful technique for diagnosing PM in the intensive care unit. No potential conflict of interest relevant to this article was reported. Table 1. Baseline characteristics and underlying disease of patients with pulmonary mucormycosis Characteristic Value Age (yr) 64 (12–73) Sex (M/F) 4/5 Underlying disease Malignancy 9 Hematologic malignancy 7 (77.8) Solid cancer 2 (22.2) Diabetes mellitus 3 Proven/probable pulmonary mucormycosis 7/2 Diagnostic method Transbronchial lung biopsy 6 Surgical biopsy 2 Percutaneous needle aspiration 1 Values are presented as mean (range), number, or number (%). Table 2. CT findings of pulmonary mucormycosis

Characteristic Value Age (yr) 64 (12–73) Sex (M/F) 4/5 Underlying disease Malignancy 9 Hematologic malignancy 7 (77.8) Solid cancer 2 (22.2) Diabetes mellitus 3 Proven/probable pulmonary mucormycosis 7/2 Diagnostic method Transbronchial lung biopsy 6 Surgical biopsy 2 Percutaneous needle aspiration 1 Values are presented as mean (range), number, or number (%). Table 2. CT findings of pulmonary mucormycosis Variable No. (%) Unilateral/bilateral 8 (89)/1 (11) Nodule 2 (22) Consolidation 8 (89) Mass 3 (33) Cavitation 3 (33) Ground glass opacity 4 (44) Air-fluid level 0 Pleural effusion 2 (22) Reverse halo sign 1 (11) Suggestion of pulmonary infection by radiologist on CT findings Bacterial pneumonia 4 (44) Atypical pneumonia 1 (11) Actinomycosis 1 (11) Fungal infection 3 (33) CT: computed tomography. Table 3. Treatment and outcomes of pulmonary mucormycosis

Variable No. (%) Unilateral/bilateral 8 (89)/1 (11) Nodule 2 (22) Consolidation 8 (89) Mass 3 (33) Cavitation 3 (33) Ground glass opacity 4 (44) Air-fluid level 0 Pleural effusion 2 (22) Reverse halo sign 1 (11) Suggestion of pulmonary infection by radiologist on CT findings Bacterial pneumonia 4 (44) Atypical pneumonia 1 (11) Actinomycosis 1 (11) Fungal infection 3 (33) CT: computed tomography. Table 3. Treatment and outcomes of pulmonary mucormycosis Case Age (yr) Sex Underlying disease ANC Diagnostic method CT diagnosis Pathology Drug Period (d)a Outcome 1 69 F AML 290 TBLB BP or AP R/O Amphotericin B 3 Death 2 17 M NHL 160 TBLB Fungal infection C/W Amphotericin B 50 Alive 3 73 F PNH 10,720 TBLB BP C/W L-amphotericin 9 Death 4 55 M AML 1,440 TBLB BP C/W Amphotericin 21 Death 5 66 M Multiple cancer 2,920 PCNA Actinomycosis R/O Itraconazole 84 Alive 6 12 M LCH 1,460 VATs Fungal infection C/W L-amphotericin 154 Alive 7 70 F RCC 6,060 TBLB BP C/W Amphotericin B 15 Death 8 64 F AML 1,980 TBLB BP C/W L-amphotericin 35 Death 9 14 F AML 300 VATs Fungal infection C/W L-amphotericin 144 Alive ANC: absolute neutrophil count at diagnosis of pulmonary mucormycosis; CT: computed tomography; AML: acute myeloid leukemia; TBLB: transbronchial lung biopsy; BP: bacterial pneumonia; AP: atypical pneumonia; R/O: further need differentiate between aspergillosis and mucormycosis; NHL: non-Hodgkin’s lymphoma; C/W: compatible with mucormycosis; PNH: paroxysmal nocturnal hemoglobinuria; L-amphotericin: liposomal amphotericin; PCNA: percutaneous needle aspiration; LCH: langerhans cell histiocytosis; VATs: video assisted thoracoscopic surgery; RCC: renal cell carcinoma.

pergillosis and mucormycosis; NHL: non-Hodgkin’s lymphoma; C/W: compatible with mucormycosis; PNH: paroxysmal nocturnal hemoglobinuria; L-amphotericin: liposomal amphotericin; PCNA: percutaneous needle aspiration; LCH: langerhans cell histiocytosis; VATs: video assisted thoracoscopic surgery; RCC: renal cell carcinoma. a Treatment periods with antifungal agents.

Chronic obstructive pulmonary disease (COPD), which is marked by a fixed obstruction of the airway, is a progressive disease including emphysema and chronic bronchitis. According to data from the fourth Korea National Health and Nutrition Survey, the prevalence of COPD is 13.4% in South Korea [1]. Intensive care unit (ICU) admission is required in more than 25% of patients with COPD [2]. ICU stays of patients with lung hyperinflation are longer compared to patients without lung hyperinflation [3]. In an acute exacerbation of COPD, airway resistance rises, positive end expiratory pressure (PEEP) rises, and hyperinflation of lungs occurs. Increased lung volume by hyperinflation compresses inferior vena cava and right ventricle, therefore decreases cardiac output and blood pressure. Moreover, asymmetric lung compliance aggravates unilateral lung hyperinflation which is found in unilateral lung transplantation, fibrosis or pneumonia of a single lung [4,5]. Severe obstruction increases work of breathing and fatigue of respiratory muscles [6]. With the understanding about respiratory dynamics, the better strategies will be discussed. The uneven distribution of volume could be reduced by reducing the diameter of the airway [7]. Measurement of lung hyperinflation is integral to the assessment of physiological impairment in individuals with COPD and can effectively be targeted for treatment [8]. Recognition and successful management of the unilateral lung hyperinflation may avoid complications such as barotrauma and hypotension associated with the presence of intrinsic PEEP [9].

With the understanding about respiratory dynamics, the better strategies will be discussed. The uneven distribution of volume could be reduced by reducing the diameter of the airway [7]. Measurement of lung hyperinflation is integral to the assessment of physiological impairment in individuals with COPD and can effectively be targeted for treatment [8]. Recognition and successful management of the unilateral lung hyperinflation may avoid complications such as barotrauma and hypotension associated with the presence of intrinsic PEEP [9]. No potential conflict of interest relevant to this article was reported.

Introduction Acute respiratory distress syndrome (ARDS) is a refractory hypoxemic respiratory failure with bilateral lung infiltrates, in the absence of left atrial hypertension or evidence of volume overload [1,2]. The ARDS is associated with high mortality rates, which vary widely from 30% to 70% in several reports [3,4]. Although the annual ARDS mortality rates have shown improvement in several studies [5,6], ARDS remains a life-threatening disease with high mortality. One of major pathophysiologic mechanisms in the early phase of ARDS is known to be endothelial injury and alveolar epithelial damage from proinflammatory cascade [7] , leading to the process of proliferation and fibrosis [8]. Inflammatory biomarkers such as tumor necrosis factor-alpha, interleukin 6, platelet activating factor are known to involve in ARDS pathogenesis [9-11]. In the advanced stage, activated alveolar fibrocytes, fibroblasts and myofibroblasts may drive fibroproliferation, although the process remains unclear [12]. These inflammatory cytokines and increased pulmonary fibrosis predict unfavorable outcome and higher mortality [13,14]. Therefore, the anti-inflammatory therapy or prevention of pulmonary fibrosis in ARDS could be targets of therapy to improve survival.

ive fibroproliferation, although the process remains unclear [12]. These inflammatory cytokines and increased pulmonary fibrosis predict unfavorable outcome and higher mortality [13,14]. Therefore, the anti-inflammatory therapy or prevention of pulmonary fibrosis in ARDS could be targets of therapy to improve survival. Angiotensin II exerts proinflammatory responses through activating nuclear factor-κB in monocytes [15]. The renin-angiotensin system (RAS) inhibitors, such as angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB) reduces nuclear factor-κB activation and lipopolysaccharide induced lung neutrophil recruitment [15-17]. Rats with acute lung injury showed lower lung injury score or alveolar collagen deposit when treated with captopril and they suggested inhibition of ACE could offer protective effects on acute lung injury [18-21]. The mechanism is supposed that captopril attenuates lung fibrosis by abrogating apoptosis in lung epithelial cells. ACE inhibitor is also well known to have a protective effect against fibrosis and remodeling, notably in the liver, kidneys and heart [22-24]. However, there was little evidence that ACE inhibitor played a protective role in human ARDS. We hypothesized that ACE inhibitor could provide beneficial effects for ARDS patients by protecting them from prolonged inflammation and fibrosis. Therefore, we conducted this study to determine whether there is a difference in clinical outcomes including mortality rates in ARDS patients according to the use of ACE inhibitor or ARB.

d that ACE inhibitor could provide beneficial effects for ARDS patients by protecting them from prolonged inflammation and fibrosis. Therefore, we conducted this study to determine whether there is a difference in clinical outcomes including mortality rates in ARDS patients according to the use of ACE inhibitor or ARB. Materials and Methods 1) Patients population We retrospectively reviewed the medical records and radiographs of patients admitted to the medical intensive care unit (MICU) at a tertiary care hospital for mechanical ventilation support between January 2005 and December 2010. Eligibility criteria were that patients were: (1) over 20 years of age, and (2) fulfilled the Berlin definition: acute onset within 1 week of a known clinical insult or new or worsening respiratory symptoms; bilateral opacities not fully explained by effusions, lobar/lung collapse, or nodules; respiratory failure not fully explained by cardiac failure or fluid overload; arterial hypoxemia with PaO2/FiO2 lower than 300 (positive end-expiratory pressure [PEEP] or continuous positive airway pressure, ≥5 cmH2O) [25]. We excluded patients who were not intubated within 24 hours of ICU transfer or did not stay in the ICU for more than 48 hours following admission. Patients who were intubated and transferred from outside were also excluded.

(positive end-expiratory pressure [PEEP] or continuous positive airway pressure, ≥5 cmH2O) [25]. We excluded patients who were not intubated within 24 hours of ICU transfer or did not stay in the ICU for more than 48 hours following admission. Patients who were intubated and transferred from outside were also excluded. 2) Study design and data collection Patients were classified into two groups according to whether or not they took ACE inhibitor/ARB after ICU admission (RAS inhibitor group and non-RAS inhibitor group, respectively). The study was approved by the institutional review board and ethics committee (No. H-1206-090-414). No consent from the participants was needed. Demographic data (age, gender), risk factors for ARDS and etiology of ARDS were evaluated. We also calculated the PaO2/FiO2 ratio, Sequential Organ Failure Assessment (SOFA) score (at admission, 48 hours and 96 hours after admission) and Acute Physiology and Chronic Health Evaluation (APACHE) II score (within 24 hours after admission). Medication histories were reviewed to find the patients who took ACE inhibitor or ARB after ICU admission, and to collect data about the generic name, dose and duration of medication. The primary outcome was ICU mortality, and secondary outcomes we measured by length of stay in ICU, ICU readmission rate, reintubation rate, radiologic improvement (decreased consolidation or infiltrates on chest radiography at ICU discharge compared to ICU admission) weaning failure rate (rate of reintubation within 48 hours after extubation) and 28- and 90-day in-hospital mortality.

asured by length of stay in ICU, ICU readmission rate, reintubation rate, radiologic improvement (decreased consolidation or infiltrates on chest radiography at ICU discharge compared to ICU admission) weaning failure rate (rate of reintubation within 48 hours after extubation) and 28- and 90-day in-hospital mortality. 3) Statistical analysis Quantitative variables are expressed as mean ± standard deviation, and qualitative variables are expressed as numbers and percentages. Differences between independent groups were assessed by paired t-tests for quantitative data and chi-square tests for qualitative variables. We used propensity score matching to reduce the bias caused by confounding variables. The propensity score covariates included age, gender, preexisting disease (hypertension, diabetes, coronary artery disease, chronic obstructive pulmonary disease, tuberculosis, chronic liver disease, chronic kidney disease, and malignancy) and clinical disease severity scores (SOFA score and APACHE II score). Patients with similar propensity scores in each group were matched with 1 to 1 ratio. Outcomes were compared in both unmatched and matched cohort. In the ICU and hospital, mortalities were estimated with the Kaplan-Meier analysis and log-rank test using propensity score matching analysis. All statistical analyses were performed using Stata version 13.0 (Stata Corp., College Station, TX, USA). A P-value <0.05 was considered to indicate a significant difference.

t. In the ICU and hospital, mortalities were estimated with the Kaplan-Meier analysis and log-rank test using propensity score matching analysis. All statistical analyses were performed using Stata version 13.0 (Stata Corp., College Station, TX, USA). A P-value <0.05 was considered to indicate a significant difference. Results We reviewed the records of 3,309 patients admitted to the ICU between January 2005 and December 2010. Of the 294 patients who satisfied the inclusion criteria, 112 patients were excluded, leaving 182 patients for the final analysis. After the propensity score matching, 34 patients in each group were analyzed (Figure 1). Of the 182 patients, 134 (74%) were male and the mean age was 64.1 ± 13.7 years. Thirty-seven patients received ACE inhibitor and ARB during their ICU stay. Table 1 presents the patients’ demographic characteristics and history of underlying disease. History of hypertension and chronic kidney disease were significantly higher in the RAS inhibitor group. Though history of malignancy was higher in patients in the non-RAS inhibitor group, patients with lung cancer were higher in RAS inhibitor group. (P = 0.590) No difference in the proportion of direct lung injury was observed between the two groups.

ic kidney disease were significantly higher in the RAS inhibitor group. Though history of malignancy was higher in patients in the non-RAS inhibitor group, patients with lung cancer were higher in RAS inhibitor group. (P = 0.590) No difference in the proportion of direct lung injury was observed between the two groups. Eighty-nine patients (48.9%) were classified as severe ARDS according to the Berlin definition. Proportion of patients with severe ARDS was higher in the RAS inhibitor group (54.3%) than non-RAS inhibitor group (47.6%) without statistical significance (P = 0.431). Proportions of patients with mild ARDS were 4.8% and 8.6%, those with moderate ARDS were 47.6% and 37.1% in non-RAS inhibitor group and RAS inhibitor group, respectively. We analyzed the PaO2/FiO2 ratio (initial, 24 hours, 72 hours), SOFA score (initial, 48 hours, 96 hours), and APACHE II score and found no differences between the two groups. The average PEEP levels were about 6.9 cmH2O in both groups (P = 0.681). Tidal volumes were also similar, those values were 7.4 ± 1.9 ml/kg and 7.5 ± 1.9 ml/kg, respectively in non-RAS inhibitor group and RAS inhibitor group (P = 0.774). We constructed the propensity score matched cohort, which consisted of 34 patients in each group. Demographic characteristics, comorbidities and mechanical ventilator parameters were similar between two groups in the propensity score matched cohort.

tively in non-RAS inhibitor group and RAS inhibitor group (P = 0.774). We constructed the propensity score matched cohort, which consisted of 34 patients in each group. Demographic characteristics, comorbidities and mechanical ventilator parameters were similar between two groups in the propensity score matched cohort. About 60% of the patients in the RAS inhibitor group had taken ACE inhibitor or ARB within 1 month before MICU admission (Table 2). In the RAS inhibitor group, 10 patients had been given ACE inhibitor and 27 patients were given ARB. For the patients receiving ACE inhibitor/ARB, ICU mortality was 45.9% (17 of 37 patients), which was lower than the 58.9% mortality rate of patients without medication, though not significantly (P = 0.166) (Table 3). Inhospital mortality was also higher in the non-RAS inhibitor group and the day 28 mortality in non-RAS inhibitor group was almost twice as high without statistical significance (18.9% vs. 35.2%, P = 0.058). In the propensity score matched cohort, in hospital mortality at day 28 was also higher in non-RAS inhibitor group (29.4%) than RAS inhibitor group (8.8%, P = 0.031). However, other parameters of outcome including in hospital mortality at day 90 showed no significant differences between two groups.

. 35.2%, P = 0.058). In the propensity score matched cohort, in hospital mortality at day 28 was also higher in non-RAS inhibitor group (29.4%) than RAS inhibitor group (8.8%, P = 0.031). However, other parameters of outcome including in hospital mortality at day 90 showed no significant differences between two groups. ICU-free day at 28 days were about 5 days, similar in both groups (P = 0.822). ICU-free day at 90 days was longer in RAS inhibitor group (mean, 34.2 days) than non-RAS inhibitor group (mean, 29.4 days; P = 0.481). There was also no significant difference in hospital-free day at 90 days and ventilator-free day at 28 days, between two groups. The Kaplan-Meier estimate of survival showed a higher mortality rate in non-RAS inhibitor patients (P < 0.001) (Figure 2). Survival analysis after adjustment for propensity score showed similar results in each group (P = 0.002). Discussion Although many studies have been conducted to establish pharmacologic management of patients with ARDS, there is no proven pharmacologic therapy to improve the survival and prognosis of ARDS patients [26]. Clinical trials using well-known anti-inflammatory agents, including corticosteroids, 3-hydroxy-3-methylglutaryl-coenzyme-reductase inhibitors, heparins and aspirin, have not demonstrated clear clinical benefits in ARDS [27-30].

o proven pharmacologic therapy to improve the survival and prognosis of ARDS patients [26]. Clinical trials using well-known anti-inflammatory agents, including corticosteroids, 3-hydroxy-3-methylglutaryl-coenzyme-reductase inhibitors, heparins and aspirin, have not demonstrated clear clinical benefits in ARDS [27-30]. Interestingly, several studies conducted to support anti-inflammatory effect of ACE inhibitor in vascular disease [31-33] and also showed protective effect against pneumonia [34,35]. In addition, some animal models have revealed that administration of ACE inhibitor or ARB inhibits fibrosis in the kidneys, heart and liver [36-39]. We therefore tried to evaluate the effect of anti-inflammatory and antifibrotic properties of ACE inhibitor and ARB on outcomes in patients with ARDS. In this study, radiologic improvement was more frequently detected in the RAS inhibitor group, although this difference was not significant. The mean PaO2/FiO2 ratio on admission was lower in the RAS inhibitor group than in the non-RAS inhibitor group (103.1 ± 51.7 vs. 112.2 ± 45.4, P = 0.297). However, the ratio increased over time than non-RAS inhibitor group (187.8 ± 113.6 vs. 173.1 ± 90.9, P = 0.408). Clinical improvement in the late phase in regards to chest radiology and oxygenation, might reflect lower degrees in pulmonary sequelae including fibrosis in RAS inhibitor group.

s. 112.2 ± 45.4, P = 0.297). However, the ratio increased over time than non-RAS inhibitor group (187.8 ± 113.6 vs. 173.1 ± 90.9, P = 0.408). Clinical improvement in the late phase in regards to chest radiology and oxygenation, might reflect lower degrees in pulmonary sequelae including fibrosis in RAS inhibitor group. An interesting finding was that the ACE inhibitor group showed greater ICU stay duration and length of mechanical ventilation. Previous studies that conducted survival analysis suggested that surviving patients stayed longer in the ICU [31,40,41]. This could explain the differences between the two groups. Considering that no differences in ICU-free day or ventilator-free day, longer duration of ICU or mechanical ventilation could not be affirmed as poor clinical outcome. More patients in the RAS inhibitor group were transferred to general wards, supported by portable bi-level positive airway pressure ventilators (18.9% vs. 5.5%, P = 0.008). Considering the similar results in terms of PaO2/FiO2 ratio and radiologic improvement on discharge, higher dependence on the bi-level positive airway pressure ventilator might be mainly caused by muscle weakness [42].

rds, supported by portable bi-level positive airway pressure ventilators (18.9% vs. 5.5%, P = 0.008). Considering the similar results in terms of PaO2/FiO2 ratio and radiologic improvement on discharge, higher dependence on the bi-level positive airway pressure ventilator might be mainly caused by muscle weakness [42]. The mortality rates, both in the ICU and in the hospital, were not different in both groups. The overall mortality rate was high compared to other cohorts, which estimated mortality rates ranged from 28% to 58% [31,43,44]. It might reflect high disease severity considering high APACHE II score and high proportion of cancer patients in our study [45]. Among the patients who died, the duration of ICU stay was longer (P = 0.001) in the RAS inhibitor group than in the non-RAS inhibitor group, which might explain the lower 28-day mortality rates. Survival analysis showed significantly better survival rates in the RAS inhibitor group. We adjusted confounding factors, including disease severity, using the propensity score, and found the same results as above. The likelihood of survival of patients with ARDS could be improved by the administration of RAS inhibitor.

urvival analysis showed significantly better survival rates in the RAS inhibitor group. We adjusted confounding factors, including disease severity, using the propensity score, and found the same results as above. The likelihood of survival of patients with ARDS could be improved by the administration of RAS inhibitor. This study has several limitations. First, ARDS is a complicated syndrome, and it is difficult to distinguish ARDS from other diseases that might cause radiologic abnormality and respiratory failure. However, we did our best to overcome the weakness of the retrospective design. We reviewed all medical records, including echocardiography results, daily body weight changes and vital sign changes, to evaluate volume status. Records of physical examination and assessment by medical attendance were also reviewed to exclude other diseases. Second, ACE inhibitors or ARB administered to patients did not have uniform component or dosage. However, for the treatment of hypertension, ACE inhibitor and ARB showed similar efficacy in lowering blood pressure [46,47]. Though the mechanisms of action between ACE inhibitor and ARB are different, there was no relevant difference in clinical outcomes for renal protection [48]. Further researches are needed to compare efficacy in anti-inflammatory effect. Third, the sample size was relatively small, and a considerable number of patients had many comorbidities. These confounding factors might have influenced the outcome [49].

relevant difference in clinical outcomes for renal protection [48]. Further researches are needed to compare efficacy in anti-inflammatory effect. Third, the sample size was relatively small, and a considerable number of patients had many comorbidities. These confounding factors might have influenced the outcome [49]. In conclusion, in this study we found that ACE inhibitor or ARB may have beneficial effect on ARDS patients. We need large prospective clinical study to validate this finding. No potential conflict of interest relevant to this article was reported. Figure 1. Enrollment and analysis of patients. ICU: intensive care unit; SNUH: Seoul National University Hospital; ARDS: acute respiratory distress syndrome. Figure 2. Kaplan-Meier survival analysis. ACE: angiotensinconverting enzyme; ARB: angiotensin receptor blocker; ICU: intensive care unit. Table 1. Baseline characteristics

Figure 1. Enrollment and analysis of patients. ICU: intensive care unit; SNUH: Seoul National University Hospital; ARDS: acute respiratory distress syndrome. Figure 2. Kaplan-Meier survival analysis. ACE: angiotensinconverting enzyme; ARB: angiotensin receptor blocker; ICU: intensive care unit. Table 1. Baseline characteristics Variable Unmatched cohort Propensity matched cohort Non-RAS inhibitor group (n = 145) RAS inhibitor group (n = 37) P-value Non-RAS inhibitor group (n = 34) RAS inhibitor group (n = 34) P-value Age (yr) 63.3 ± 13.7 67.4 ± 13.5 0.104 70.1 ± 13.7 68.1 ± 12.2 0.514 Male sex 106 (73.1) 28 (75.7) 0.751 27 (79.4) 25 (73.5) 0.567 Comorbidity Hypertension 34 (23.5) 20 (54.1) <0.001 14 (41.2) 15 (44.1) 0.806 Diabetes 39 (26.9) 13 (35.1) 0.322 10 (29.4) 13 (38.2) 0.442 Coronary artery disease 11 (7.6) 5 (13.5) 0.256 5 (14.7) 5 (14.7) 1.000 COPD 9 (6.2) 2 (5.4) 0.855 1 (2.9) 2 (5.9) 0.555 Tuberculosis 25 (17.2) 4 (10.8) 0.340 7 (20.6) 4 (11.8) 0.323 Chronic liver disease 17 (11.7) 2 (5.4) 0.262 1 (2.9) 2 (5.9) 0.555 Chronic kidney disease 8 (5.5) 12 (32.4) <0.001 8 (23.5) 11 (32.4) 0.417 Cerebral vascular accident 7 (4.8) 4 (10.8) 0.173 4 (11.8) 3 (8.8) 0.690 Malignancy 92 (63.5) 12 (35.3) 0.003 10 (29.4) 12 (35.3) 0.604 Pulmonary 17 (18.5) 3 (25.0) 0.590 3 (30.0) 3 (25.0) 0.793 Extrapulmonary 75 (81.5) 9 (75) - 7 (70.0) 9 (75.0) - Injury mechanism 0.789 1.000 Direct 127 (87.6) 33 (89.2) 30 (88.2) 30 (88.2) Indirect 18 (12.4) 4 (10.8) 4 (11.8) 4 (11.8) ARDS severity 0.431 0.200 Mild 7 (4.8) 3 (8.6) 0 3 (8.8) Moderate 70 (47.6) 13 (37.1) 12 (35.3) 12 (35.3) Severe 70 (47.6) 19 (54.3) 22 (64.7) 22 (64.7) PF ratio Initial 112.2 ± 45.4 103.1 ± 51.7 0.297 93.9 ± 33.4 103.0 ± 53.6 0.407 24 hr after admission 144.5 ± 73.0 144.9 ± 69.9 0.974 133.4 ± 68.2 149.2 ± 71.1 0.353 72 hr after admission 173.1 ± 90.9 187.8 ± 113.6 0.408 150.1 ± 81.8 193.8 ± 116.1 0.081 Discharge day 141.5 ± 114.3 164.6 ± 124.2 0.282 132.9 ± 113.4 173.3 ± 125.8 0.168 APACHE II score 31.5 ± 6.2 30.3 ± 5.2 0.316 29.5 ± 6.8 30.2 ± 5.0 0.612 SOFA score Initial 11.1 ± 2.9 10.8 ± 2.5 0.542 10.1 ± 2.7 10.6 ± 2.5 0.462 48 hr after admission 10.9 ± 3.8 9.5 ± 2.5 0.031 11.1 ± 4.1 9.4 ± 2.3 0.051 96 hr after admission 10.2 ± 3.8 9.3 ± 2.8 0.190 10.1 ± 3.9 9.1 ± 2.7 0.246 Mechanical ventilation PEEP (cmH2O) 6.9 ± 2.6 6.9 ± 2.6 0.681 7.6 ± 2.9 6.7 ± 2.6 0.184 Pressure above PEEP (cmH2O) 18.9 ± 4.7 18.8 ± 4.2 0.845 18.8 ± 4.2 18.4 ± 3.9 0.639 FiO2 0.71 ± 0.21 0.71 ± 0.19 0.928 0.75 ± 0.19 0.71 ± 0.20 0.351 Tidal volume (ml

er admission 10.2 ± 3.8 9.3 ± 2.8 0.190 10.1 ± 3.9 9.1 ± 2.7 0.246 Mechanical ventilation PEEP (cmH2O) 6.9 ± 2.6 6.9 ± 2.6 0.681 7.6 ± 2.9 6.7 ± 2.6 0.184 Pressure above PEEP (cmH2O) 18.9 ± 4.7 18.8 ± 4.2 0.845 18.8 ± 4.2 18.4 ± 3.9 0.639 FiO2 0.71 ± 0.21 0.71 ± 0.19 0.928 0.75 ± 0.19 0.71 ± 0.20 0.351 Tidal volume (ml /kg) 7.4 ± 1.9 7.5 ± 1.9 0.774 7.3 ± 2.0 7.5 ± 1.9 0.612 Minute volume (L) 9.8 ± 2.9 9.5 ± 3.0 0.603 8.8 ± 2.8 9.5 ± 3.1 0.382 Steroid 89 (61.4) 21 (56.8) 0.608 14 (41.2) 19 (55.9) 0.225 CRRT 34 (23.5) 7 (18.9) 0.556 7 (20.6) 6 (17.7) 0.758 NO gas 56 (38.6) 17 (46.0) 0.417 13 (38.2) 15 (44.1) 0.622 Inotropics 131 (90.3) 29 (78.4) 0.046 30 (88.2) 26 (76.5) 0.203 Values are presented as mean ± standard deviation or number (%). RAS: renin-angiotensin system; COPD: chronic obstructive pulmonary disease; ARDS: acute respiratory distress syndrome; PF: PaO2/FiO2; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; PEEP: positive end-expiratory pressure; CRRT: continuous renal replacement therapy. Table 2. Cause of ICU admission Cause Total (n = 182) Non-RAS inhibitor group (n = 145) RAS inhibitor group (n = 37) Pulmonary Pneumonia 136 (74.7) 112 (76.2) 24 (68.6) Hemorrhage 10 (5.5) 6 (4.1) 4 (11.4) Exacerbation of ILD 4 (2.8) 3 (2.0) 2 (5.7) Extrapulmonary Bacteremia 11 (6.0) 10 (6.8) 1 (2.9) Postoperative 5 (2.75) 2 (1.4) 3 (8.6) Transfusion related 4 (2.2) 4 (2.2) 0 Drug related 2 (1.1) 2 (1.4) 0 Radiation related 1 (0.6) 1 (0.7) 0 Unknown 8 (4.4) 7 (4.8) 1 (2.9) Values are presented as number (%).

.1) 4 (11.4) Exacerbation of ILD 4 (2.8) 3 (2.0) 2 (5.7) Extrapulmonary Bacteremia 11 (6.0) 10 (6.8) 1 (2.9) Postoperative 5 (2.75) 2 (1.4) 3 (8.6) Transfusion related 4 (2.2) 4 (2.2) 0 Drug related 2 (1.1) 2 (1.4) 0 Radiation related 1 (0.6) 1 (0.7) 0 Unknown 8 (4.4) 7 (4.8) 1 (2.9) Values are presented as number (%). ICU: intensive care unit; RAS: renin-angiotensin system; ILD: interstitial lung disease. Table 3. Primary and secondary outcomes according to study group

.1) 4 (11.4) Exacerbation of ILD 4 (2.8) 3 (2.0) 2 (5.7) Extrapulmonary Bacteremia 11 (6.0) 10 (6.8) 1 (2.9) Postoperative 5 (2.75) 2 (1.4) 3 (8.6) Transfusion related 4 (2.2) 4 (2.2) 0 Drug related 2 (1.1) 2 (1.4) 0 Radiation related 1 (0.6) 1 (0.7) 0 Unknown 8 (4.4) 7 (4.8) 1 (2.9) Values are presented as number (%). ICU: intensive care unit; RAS: renin-angiotensin system; ILD: interstitial lung disease. Table 3. Primary and secondary outcomes according to study group Variable Unmatched cohort Propensity matched cohort Non-RAS inhibitor group (n = 145) RAS inhibitor group (n = 37) P-value Non-RAS inhibitor group (n = 34) RAS inhibitor group (n = 34) P-value Duration of mechanical ventilation (d) 19.5 ± 21.9 29.5 ± 20.8 0.013 16.4 ± 8.6 30.2 ± 21.5 <0.001 Radiologic improvement 69 (47.6) 22 (59.5) 0.197 17 (50.0) 21 (61.8) 0.329 ICU readmission 4 (2.8) 2 (5.4) 0.421 2 (5.9) 4 (11.8) 0.393 ICU day 20.2 ± 15.7 32.1 ± 21.7 <0.001 19.9 ± 10.9 33.6 ± 23.1 0.003 Hospital day 58.0 ± 90.1 89.8 ± 111.2 0.070 43.8 ± 31.2 94.7 ± 114.7 0.015 ICU-free day, 28 days 5.2 ± 8.1 4.8 ± 7.7 0.822 5.1 ± 7.6 4.9 ± 7.8 0.962 ICU-free day, 90 days 29.4 ± 36.4 34.2 ± 33.8 0.481 32.4 ± 37.3 35.2 ± 33.8 0.747 Hospital-free day, 90 days 16.2 ± 25.9 14.2 ± 24.9 0.685 16.6 ± 25.3 14.6 ± 25.2 0.749 Ventilator-free day, 28 days 6.4 ± 8.9 6.1 ± 8.9 0.898 7.1 ± 8.9 6.3 ± 8.9 0.737 Mortality In ICU mortality 85 (58.6) 17 (46.0) 0.166 19 (55.9) 14 (41.2) 0.225 In hospital mortality 95 (65.5) 21 (56.8) 0.322 20 (58.8) 18 (52.9) 0.625 At day 28 51 (35.2) 7 (18.9) 0.058 10 (29.4) 3 (8.8) 0.031 At day 90 82 (56.6) 16 (43.2) 0.147 18 (52.9) 13 (38.2) 0.223 Tracheostomy 41 (28.3) 17 (46.0) 0.040 12 (35.3) 16 (47.1) 0.324 Change to portable BiPAP 8 (5.5) 7 (18.9) 0.008 2 (5.9) 7 (20.6) 0.074 Reintubation 25 (17.2) 7 (18.9) 0.735 6 (17.7) 6 (17.7) 1.000 Values are presented as mean ± standard deviation or number (%).

ay 90 82 (56.6) 16 (43.2) 0.147 18 (52.9) 13 (38.2) 0.223 Tracheostomy 41 (28.3) 17 (46.0) 0.040 12 (35.3) 16 (47.1) 0.324 Change to portable BiPAP 8 (5.5) 7 (18.9) 0.008 2 (5.9) 7 (20.6) 0.074 Reintubation 25 (17.2) 7 (18.9) 0.735 6 (17.7) 6 (17.7) 1.000 Values are presented as mean ± standard deviation or number (%). RAS: renin-angiotensin system; ICU: intensive care unit; BiPAP: bilevel positive airway pressure.

ys [range, 10 to 63 days] and 47 days [range, 14 to 163 days], respectively, P = 0.013) (Figure 1). In addition, the median number of days ventilator-free [6] was significantly higher in the early than the late tracheostomy group (4.5 days [range, 0 to 24 days] and 0 day [range, 0 to 14 days], respectively, P = 0.040). Discussion Of the 38 patients who underwent ST while on ECMO, the overall rate of ST-related clinically relevant complications was 10.6% and the rate of major bleeding complications was 5.3%. Our results suggest that ST can be safely used by experienced operators on adult patients undergoing ECMO, with careful optimization of the coagulation status. Utilization of ECMO in the awake patient in the ICU has gained widespread acceptance [2]. However, despite the inherent benefits, extubation during ECMO is possible only for some patients and under the control of experienced personnel. Camporota et al. [1] reported that very few centers extubated patients during ECMO (almost 90% of centers extubated <25% of their patients). Only 1.5% of the ECMO centers reviewed extubated more than 75% of their patients while on ECMO. Currently, most clinicians prefer to keep patients intubated or perform an early tracheostomy to allow self-ventilation.

Introduction Most centers carrying out extracorporeal membrane oxygenation (ECMO) find difficulty in performing extubation while patients receive ECMO [1]. Use of ECMO with the patient awake could avoid several disadvantage related to sedation, intubation, and mechanical ventilation [2]. However, given the complexity of patient‒ECMO interaction, in practical terms, this method would be applicable only for some patients at centers where personnel are sufficiently experienced [3]. Thus, tracheostomy is recommended when prolonged mechanical ventilation is presumed [4]. However, the risks of bleeding during ECMO may deter many centers from performing surgical tracheostomy (ST). Therefore, it is clinically important to evaluate safety issues through a review of adverse events after ST during the use of ECMO. We reviewed and analyzed the clinical correlation of preoperative coagulation status and bleeding complications after ST during ECMO utilization in critically ill patients.

cheostomy (ST). Therefore, it is clinically important to evaluate safety issues through a review of adverse events after ST during the use of ECMO. We reviewed and analyzed the clinical correlation of preoperative coagulation status and bleeding complications after ST during ECMO utilization in critically ill patients. Materials and Methods We retrospectively identified all patients who underwent ST while on ECMO support, from April 2012 to March 2016. The following routine clinical baseline data were collected: age, sex, diagnosis and reason for ECMO support, Acute Physiology and Chronic Health Evaluation II score, type of ECMO, time from commencement of mechanical ventilation to start of ECMO support, and time from start of ECMO to tracheostomy. In addition, the following hematological and coagulatory parameters were measured: hemoglobin level, platelet count, international normalized ratio (INR), activated partial thromboplastin time (aPTT) prior to ST, and amount and type of blood products administrated 24 hours before and after ST. The institutional review board of our hospital approved this study and waived the requirement for written informed consent.

l, platelet count, international normalized ratio (INR), activated partial thromboplastin time (aPTT) prior to ST, and amount and type of blood products administrated 24 hours before and after ST. The institutional review board of our hospital approved this study and waived the requirement for written informed consent. 1) ECMO protocol and routine care We applied the ECMO protocol based on the Extracorporeal Life Support Organization guidelines [5]. The ECMO system consisted of a polymethylpentene fiber oxygenator system (QUADROX PLS; Maquet Inc., Hirrlingen, Germany) with simplified Bioline-coated circuits (Maquet Inc., Rastatt, Germany). All patients were supported with centrifugal pumps (Maquet Inc.), and were cannulated peripherally through the femoral artery and vein or internal jugular vein using the Seldinger technique. Patients received an initial unfractionated heparin (UFH) bolus of 50 units/kg body weight when the cannula was placed, and UFH was continuously infused during ECMO. Exceptionally, patients who have bleeding complications supported ECMO without anticoagulation therapy. Based on assessment by the attending physician and intensivist of the individual benefits and risks of the procedure, most patients on ECMO who were undergoing, or expected to undergo, a prolonged weaning process underwent the ST procedure.

who have bleeding complications supported ECMO without anticoagulation therapy. Based on assessment by the attending physician and intensivist of the individual benefits and risks of the procedure, most patients on ECMO who were undergoing, or expected to undergo, a prolonged weaning process underwent the ST procedure. 2) Anticoagulation and tracheostomy protocol Our anticoagulation protocol involved stopping the administration of intravenous UFH 4 hours before the planned tracheostomy. However, we did not stop heparin infusion in patients at high risk of a thromboembolic event. ST was performed at the bedside in the intensive care unit (ICU) under sedation with propofol and remifentanil. The patient’s shoulders were elevated and the head extended (unless cervical disease or injury was present), thereby elevating the upper trachea. Sterile drapes were placed, creating a surgical window from the top of the larynx to the suprasternal notch. After administration of local anesthesia, a skin incision (2–3 cm) was made at two fingers’ width from the upper sternal notch. Sharp dissection was then made to cut through the platysma muscle, with bleeding controlled by hemostats and electrocautery. Blunt dissection parallel to the long axis of the trachea was then used to separate the submuscular tissues until the thyroid isthmus was identified. The isthmus must be mobilized and a small incision made to create a space for the tracheostomy procedure. Blunt dissection was continued longitudinally through the pretracheal fascia, and the target tracheal ring (usually the third) was identified. Complete removal of the anterior part of this ring, according to tube size, was performed to create the stoma. After removal of the intubation tube, the tracheostomy tube was inserted with an attached ventilator. With continuous control of bleeding, the tracheostomy was completed. Two hours following the end of the tracheostomy procedure, heparin administration was initiated at the previous rate and adjusted according to subsequent coagulation test results. After surgery, the patients were maintained in a state of wakefulness without sedation.

control of bleeding, the tracheostomy was completed. Two hours following the end of the tracheostomy procedure, heparin administration was initiated at the previous rate and adjusted according to subsequent coagulation test results. After surgery, the patients were maintained in a state of wakefulness without sedation. 3) Outcome definition Complications associated with the tracheostomy procedure were graded as either major or minor, and were followed up until either death or hospital discharge. Major complications were defined as any of the following: procedure-related death, acute hypotension, tracheal wall injury, pneumothorax, major bleeding requiring surgical repair, and stoma infection. Minor complications included localized minor bleeding, which was defined as self-limiting bleeding or bleeding successfully treated by local compression. In general, in patients undergoing tracheostomy with ECMO support, oozing-type bleeding frequently occurs, with no special postoperative monitoring focus. We considered this oozing-type bleeding as major if an additional hemostatic procedure, such as electrocauterization, was required. Any clotting complications related to ECMO were recorded, including any documented evidence of visible new clot formation within the circuit or a sudden malfunction of the ECMO within 24 hours of the ST procedure.

-type bleeding as major if an additional hemostatic procedure, such as electrocauterization, was required. Any clotting complications related to ECMO were recorded, including any documented evidence of visible new clot formation within the circuit or a sudden malfunction of the ECMO within 24 hours of the ST procedure. 4) Statistical analysis All statistical analyses were performed using R software version 3.0.1 (http://www.R-project.org). Continuous variables were described as medians and ranges (minimum to maximum) and compared using the nonparametric Wilcoxon rank sum test. Categorical variables were described as numbers (%) and compared using Fisher exact test or chi-squared test. All tests were two-sided, and P-values less than 0.050 were considered statistically significant.

escribed as medians and ranges (minimum to maximum) and compared using the nonparametric Wilcoxon rank sum test. Categorical variables were described as numbers (%) and compared using Fisher exact test or chi-squared test. All tests were two-sided, and P-values less than 0.050 were considered statistically significant. Results 1) Baseline characteristics Between April 1, 2012 and March 31, 2016, 38 patients underwent ST while on ECMO. Their median age was 59 years (range, 20 to 83 years), and 26 (68.4%) were male. Thirty-three patients (86.8%) were on venoveno (VV) ECMO; three (7.9%) were on venovenoarterial (VVA) ECMO; and two (5.3%) were on venoarterial (VA) ECMO. The main indication for ECMO was severe acute respiratory distress syndrome, which was present in 31 patients (81.6%). The main primary lung-related diagnosis was pneumonia, which was present in 13 patients (34.2%). Twenty-three patients (60.5%) were administered UFH for anticoagulation during ECMO, while 15 patients were supported by ECMO without anticoagulation due to bleeding complications prior to ST. The most common site of bleeding was retroperitoneal (5/15), followed by gastrointestinal (4/15), intrapulmonary (4/15), and cannula site (2/15). ST was performed on a median of 9.5 days after intubation (range, 1 to 25 days) and on a median of 9 days after ECMO initiation (range, 1 to 18 days). Early tracheostomy (<8 days after intubation) was performed in 12 cases (31.6%) (Table 1).

ed by gastrointestinal (4/15), intrapulmonary (4/15), and cannula site (2/15). ST was performed on a median of 9.5 days after intubation (range, 1 to 25 days) and on a median of 9 days after ECMO initiation (range, 1 to 18 days). Early tracheostomy (<8 days after intubation) was performed in 12 cases (31.6%) (Table 1). 2) Coagulation characteristics The median values for hemoglobin, platelet count, INR, and aPTT before ST were 9.6 g/dl (range, 7.5 to 12.5 g/dl), 126 ×109 /L (range, 46 to 434 ×109 /L), 1.2 (range, 1 to 2.3), and 62 seconds (27 to 114.2 seconds), respectively. The median values for hemoglobin in the 24 hours following ST were similar to those during the previous 24 hours (median, 9.4 g/dl; range, 6.4 to 12.7 g/dl). In 13 out of the 23 patients administered anticoagulation therapy, heparin infusion was paused before ST, while in 10 it was maintained during the procedure. We classified patients into three groups in terms of pre-tracheostomy heparin management (group 1, heparin stopped within 4 hours before ST; group 2, heparin maintained during ST; group 3, no heparin administered before ST). There were no significant differences among the three groups regarding the median values for hemoglobin and platelet count before ST (P = 0.796, P = 0.579). However, there were significant differences among the three groups regarding the median values for INR and aPTT before ST (P = 0.020, P = 0.014). In addition, there were no significant differences among the three groups regarding the median values for hemoglobin 24 hours after ST (P = 0.166) (Table 2).

796, P = 0.579). However, there were significant differences among the three groups regarding the median values for INR and aPTT before ST (P = 0.020, P = 0.014). In addition, there were no significant differences among the three groups regarding the median values for hemoglobin 24 hours after ST (P = 0.166) (Table 2). 3) Complications There were no ST procedure-related deaths or periprocedural thrombotic complications related to ECMO. Two patients (5.3%) suffered a major bleeding complication: one patient on VV ECMO, and for whom the heparin infusion was stopped before ST, developed bleeding at the tracheostomy site, which required electrocauterization; the second patient, supported by VA ECMO and not administered heparin due to retroperitoneal bleeding, developed delayed postoperative bleeding from a small vessel arising from the innominate artery. This was due to vascular erosion of the friable mucosa as a result of friction from the tracheostomy tube, and required surgical repair. All other complications were classified as minor bleeding (n = 2, 5.3%) and were treated with local compression. Regarding the incidence of major or minor bleeding, there was no significant difference found among three groups according to pre-tracheostomy heparin management (P = 0.723). The total bleeding incidence after ST in groups 1, 2, and 3 was 15.4%, 0%, and 13.3%, respectively (Table 3).

3) Complications There were no ST procedure-related deaths or periprocedural thrombotic complications related to ECMO. Two patients (5.3%) suffered a major bleeding complication: one patient on VV ECMO, and for whom the heparin infusion was stopped before ST, developed bleeding at the tracheostomy site, which required electrocauterization; the second patient, supported by VA ECMO and not administered heparin due to retroperitoneal bleeding, developed delayed postoperative bleeding from a small vessel arising from the innominate artery. This was due to vascular erosion of the friable mucosa as a result of friction from the tracheostomy tube, and required surgical repair. All other complications were classified as minor bleeding (n = 2, 5.3%) and were treated with local compression. Regarding the incidence of major or minor bleeding, there was no significant difference found among three groups according to pre-tracheostomy heparin management (P = 0.723). The total bleeding incidence after ST in groups 1, 2, and 3 was 15.4%, 0%, and 13.3%, respectively (Table 3). 4) Clinical course and further outcomes The successful rates of weaning off ECMO and 6-month survival were 65.8% and 44.7%, respectively. The total durations of mechanical ventilation, ECMO support, and ICU/hospital length of stay are presented in Table 1. There were no significant differences between the early (<8 days of intubation) and late tracheostomy groups regarding both 6-month survival rates (41.7% and 46.2%, respectively; P = 0.796) and median hospital length of stay (57.5 days [range, 14 to 149 days] and 65 [range, 16 to 228], respectively, P = 0.343). However, early tracheostomy significantly reduced the median number of days on mechanical ventilation (23.5 days [range, 4 to 63 days] and 38.5 days [range, 14 to 76 days], respectively, P = 0.007) and median ICU length of stay (27 days [range, 10 to 63 days] and 47 days [range, 14 to 163 days], respectively, P = 0.013) (Figure 1). In addition, the median number of days ventilator-free [6] was significantly higher in the early than the late tracheostomy group (4.5 days [range, 0 to 24 days] and 0 day [range, 0 to 14 days], respectively, P = 0.040).

ry few centers extubated patients during ECMO (almost 90% of centers extubated <25% of their patients). Only 1.5% of the ECMO centers reviewed extubated more than 75% of their patients while on ECMO. Currently, most clinicians prefer to keep patients intubated or perform an early tracheostomy to allow self-ventilation. The most important clinical benefit of tracheostomy is to allow early physical therapy and active rehabilitation by avoiding sedation and paralytics. It also facilitates pulmonary toilet and early ventilator weaning, and intermittent mechanical ventilation to help prevent atelectasis, especially when the patient is fatigued after active rehabilitation [7,8]. Despite the fact that tracheostomy is required in many patients supported by ECMO for respiratory support, almost 30% of ECMO centers never perform tracheostomy on patients with ECMO, in view of the increased risk of bleeding [1]. Hemorrhage is a potential complication that can be significantly exacerbated by the use of anticoagulation which is required with ECMO, and may thus have a marked impact on the patient’s further clinical course [3,9]. The degree of bleeding may be increased either during or after tracheostomy in patients supported by ECMO, due to the activation of coagulation enzymes, or to platelet depletion or anticoagulant medication [10].

ich is required with ECMO, and may thus have a marked impact on the patient’s further clinical course [3,9]. The degree of bleeding may be increased either during or after tracheostomy in patients supported by ECMO, due to the activation of coagulation enzymes, or to platelet depletion or anticoagulant medication [10]. The rates of major bleeding in our study are similar to those recorded in studies on ST in critically ill patients [11-13]. Costa et al. [13] reported major complications in 12.2% of 56 patients undergoing ST in the emergency department. Bleeding was the most frequent complication, although no deaths were related to the procedure. Schmidt et al. [3] reported no fatal complications directly related to ST in patients supported by ECMO. In that study, bleeding was the most frequent complication of ST and no significant difference was noted regarding the type of anticoagulation management, e.g., heparin cessation or maintenance. It remains to be established whether the use of anticoagulation increases the incidence of bleeding during tracheostomy in patients on ECMO. Pasin et al. [14] reported a minor bleeding incidence of 19% for tracheostomy, and an incidence of 6% for major bleeding requiring that transfusion patients be administered anticoagulation. However, in our experience, bleeding complications were not seen in the group administered heparin. This finding may be related to the fact that, in our patients in the heparin group, the therapeutic levels were maintained well by following a low-dose anticoagulation protocol. Previously, we reported that the low-dose heparin strategy is a safe and effective protocol to reduce bleeding in adults undergoing ECMO [15]. Therefore, for the safe performance of tracheostomy, we recommend the use of low-dose heparin during ECMO and for stopping the administration of intravenous UFH 4 hours before the planned tracheostomy. Heparin administration, at the previous rate, should be initiated within 2 hours of completion of the tracheostomy, and adjusted according to subsequent coagulation test results. Exceptionally, in a patient at high risk of a thromboembolic event, ST can be performed while anticoagulation therapy is administered. In such a case, the surgeon(s) must be skilled in performing proper and accurate hemostasis to prevent postoperative bleeding.

adjusted according to subsequent coagulation test results. Exceptionally, in a patient at high risk of a thromboembolic event, ST can be performed while anticoagulation therapy is administered. In such a case, the surgeon(s) must be skilled in performing proper and accurate hemostasis to prevent postoperative bleeding. Recently, percutaneous dilatational tracheostomy (PDT) has replaced ST in many ICUs, with comparable safety and convenience in the hands of the experienced operator [16,17]. Although PDT is more accessible and less timedemanding compared to ST, it has several weaknesses [18-21]. PDT is associated with a high risk of both morbidity and mortality in certain patients, and includes the requirement for high positive end-expiratory pressure or fraction of inspired oxygen (FiO2), difficult anatomy (e.g., obesity, thick short neck, or tracheal deviation), coagulopathy, emergency procedures, and hemodynamic instability [22,23]. The need for better visualization of local neck anatomy and improved control of oxygenic/hemodynamic stability establishes ST as the safest method in critical situations such as ECMO. Although the bleeding incidence is comparable in ST and PDT, the former method is more effective in the prevention and control of bleeding complications [19,24]. Several recent reports have been published on the safety of thoracic surgical procedures involving ECMO, but none to date on the safety aspects of ST in patients supported by ECMO [25,26]. In our experience there were minimal complications, which was easily corrected. ST was well tolerated and enhanced the delivery of care in the period prior to lung recovery.

on the safety of thoracic surgical procedures involving ECMO, but none to date on the safety aspects of ST in patients supported by ECMO [25,26]. In our experience there were minimal complications, which was easily corrected. ST was well tolerated and enhanced the delivery of care in the period prior to lung recovery. This study has several limitations. The interpretation of the result is limited by potential bias introduced by the retrospective study design. The available data are limited to those recorded in the medical record, which may have been missed or underestimated. However, we recorded all severe complications, and these were detected by the detailed review. In addition, the small sample size reduces the power of this study. Therefore, further larger studies are needed to confirm these preliminary results and bleeding risk stratification for scheduled ST in patients supported by ECMO. In conclusion, the complication rates of ST in patients supported by ECMO were low and comparable to those of other critically ill patients. Therefore, ST during ECMO support is a relatively safe procedure, which should be performed by experienced operators under careful optimization of the coagulation status. The use of ST with ECMO should therefore not be dismissed on account of the potential bleeding complications caused by the administration of anticoagulation. No potential conflict of interest relevant to this article was reported.

In conclusion, the complication rates of ST in patients supported by ECMO were low and comparable to those of other critically ill patients. Therefore, ST during ECMO support is a relatively safe procedure, which should be performed by experienced operators under careful optimization of the coagulation status. The use of ST with ECMO should therefore not be dismissed on account of the potential bleeding complications caused by the administration of anticoagulation. No potential conflict of interest relevant to this article was reported. Figure 1. Differences in clinical outcomes between early and late tracheostomy in patients on extracorporeal membrane oxygenation. Values are presented as the median day. MV: mechanical ventilator; ICU: intensive care unit. Table 1. Baseline characteristics of patients with ST on ECMO

No potential conflict of interest relevant to this article was reported. Figure 1. Differences in clinical outcomes between early and late tracheostomy in patients on extracorporeal membrane oxygenation. Values are presented as the median day. MV: mechanical ventilator; ICU: intensive care unit. Table 1. Baseline characteristics of patients with ST on ECMO Variable Value (n = 38) Male sex 26 (68.4) Age (yr) 59 (20–83) Reason for ECMO ARDS 28 (73.7) ARDS with RVF 3 (7.9) Primary cardiac failure 7 (18.4) Primary diagnosis Pneumonia 13 (34.2) Acute exacerbation of COPD 1 (2.6) Acute exacerbation of ILD 9 (23.7) Trauma 1 (2.6) Postop ARDS 9 (23.7) Septic cardiomyopathy 3 (7.9) Acute coronary syndrome 2 (5.3) APACHE II score 15.5 (6–39) Anticoagulation Heparin 23 (60.5) No heparin 15 (39.5) ECMO mode VV 33 (86.8) VVA 3 (7.9) VA 2 (5.3) ST on days after intubation 9.5 (1–25) ST on days after ECMO initiation 9 (1–18) Early tracheostomy (<8 days of intubation) 12 (31.6) Mechanical ventilation day 33.5 (4-76) Hospital day 62.5 (14–228) ICU day 38 (10–163) Successful weaning of ECMO 25 (65.8) 6-Month survival 17 (44.7) Values are presented as number (%) or median (range). ECMO: extracorporeal membrane oxygenation; ARDS: acute respiratory distress syndrome; RVF: right ventricular failure; COPD: chronic obstructive pulmonary disease; ILD: interstitial lung disease; APACHE: Acute Physiology and Chronic Health Evaluation; VV: venovenous; VVA: venovenoarterial; VA: venoarterial; ST: surgical tracheostomy; ICU: intensive care unit. Table 2. Coagulatory parameters before and after ST

ECMO: extracorporeal membrane oxygenation; ARDS: acute respiratory distress syndrome; RVF: right ventricular failure; COPD: chronic obstructive pulmonary disease; ILD: interstitial lung disease; APACHE: Acute Physiology and Chronic Health Evaluation; VV: venovenous; VVA: venovenoarterial; VA: venoarterial; ST: surgical tracheostomy; ICU: intensive care unit. Table 2. Coagulatory parameters before and after ST Variable Total (n = 38) Heparin (n = 23) No-heparin (n = 15) Stop the heparin before ST (n = 13) Maintain the heparin before ST (n = 10) Prior to ST Hemoglobin (g/dl) 9.6 (7.5–12.5) 10.4 (7.5–12.2) 9.6 (8.0–12.5) 9.5 (7.5–11.9) Platelets (×109/L) 126 (46–434) 127 (65–167) 150 (46–352) 121 (46–434) INR 1.2 (1–2.3) 1.1 (1.1–1.4) 1.1 (1–1.3) 1.3 (1.1–2.3) aPTT (s) 62 (27–114.2) 66.7 (37.3–104.5) 74.2 (46.2–114.2) 49.4 (27–96.4) Within 24 h post-ST Hemoglobin (g/dl) 9.4 (6.4–12.7) 9.7 (6.4–12.2) 9.5 (8.0–12.5) 9.2 (7.8–12.3) Values are presented as median (range). ST: surgical tracheostomy; INR: international normalized ratio; aPTT: activated partial thromboplastin time. Table 3. ST-related complications Complication Value ST-related death 0 Acute hypotension 0 Major bleeding 2 Minor bleeding 2 Pneumothorax 0 Tracheal wall injury 0 Stromal infection 0 ECMO clotting post-ST 0 ST: surgical tracheostomy; ECMO: extracorporeal membrane oxygenation.

Introduction Complications after aneurysmal subarachnoid hemorrhage (aSAH) account for significant patient morbidity and mortality; fevers, in particular, have been associated with poor outcomes [1,2]. In aSAH patients, fever increases cerebral edema and intracranial pressure, exacerbates ischemia, decreases the arterio-jugular oxygen content difference, and alters levels of consciousness [3,4]. Infection can be identified as the cause of fever in approximately 75% of febrile aSAH patients [5] and is most frequently due to pneumonia, wound infection, sepsis, or urinary tract infection [3]; fever may also be due to thrombophlebitis, atelectasis, or allergic reactions. Determining the source of the febrile illness allows prompt therapy and improves the patient’s prognosis. The incidence of acalculous cholecystitis has been reported to be 0.5%–5% [6] in critically ill patients, and cerebrovascular disease is a risk factor for acute cholecystitis (AC) [7,8]. Glenn and Becker [9] also described one case of aSAH among 15 acalculous cholecystitis patients. However, little is known about the frequency of cholecystitis in patients with aSAH, and abdominal evaluations are not typically performed for febrile patients who have recently undergone aSAH surgeries. In this study, we discuss our experiences with febrile aSAH patients who were eventually diagnosed with AC.

atients. However, little is known about the frequency of cholecystitis in patients with aSAH, and abdominal evaluations are not typically performed for febrile patients who have recently undergone aSAH surgeries. In this study, we discuss our experiences with febrile aSAH patients who were eventually diagnosed with AC. Materials and Methods We retrospectively reviewed the medical records and radiological findings of 192 consecutive patients hospitalized for aSAH, between January 2009 and December 2012. We evaluated the patient characteristics, neurologic and radiologic statuses (using the Hunt and Hess grade and Fisher classification), vital signs and laboratory data during hospitalization, pathological data, lengths of fasting prior to diagnosis, and hospital courses for these patients. Patients were evaluated for fever if they had a body temperature of >38.3°C, according to the Society of Critical Care Medicine guidelines [10]. Based on physical examinations, laboratory findings, and radiologic image studies, all patients with fever were categorized according to the fever’s causes: lung problems, including pneumonia, atelectasis, and pleural effusion; catheter-related infections; operation-related infections, including cerebrospinal fluid infection and local wound infection; thrombophlebitis; drug fever; cholecystitis; inflammation due to hemorrhage without infectious causes; pseudomembranous colitis bacteremia; and urinary tract infections. Laboratory studies, including white blood cell (WBC) counts, erythrocyte sedimentation rate, and serum levels of C-reactive protein (CRP), aspartame aminotransferase (AST), alanine aminotransferase, total bilirubin, and direct bilirubin were investigated. Patients with fever due to AC underwent percutaneous gallbladder drainage (PTGBD) or laparoscopic cholecystectomy (LC). PTGBD was performed by experienced interventional radiologists, and LC was performed by experienced general surgeons. AC diagnoses and treatment modalities were based on the decisions of the general surgeons after considering each patient’s condition.

s gallbladder drainage (PTGBD) or laparoscopic cholecystectomy (LC). PTGBD was performed by experienced interventional radiologists, and LC was performed by experienced general surgeons. AC diagnoses and treatment modalities were based on the decisions of the general surgeons after considering each patient’s condition. The characteristics of patients with and without AC were compared using the Student t-test. A one-way analysis of variance was used to compare the laboratory findings for each fever cause, and a forward conditional logistic regression analysis was used to determine predictors of AC in patients with aSAH. All statistical analyses were performed using SPSS version 20.0 (IBM Corp., Armonk, NY, USA) software package. Results Of the 192 aSAH patients admitted during the study period, including two with histories of cholecystectomies, eight (4.2%; mean age, 66.29 ± 11.59 years [range, 57 to 79 years]) were eventually diagnosed with AC. Two patients underwent PTGBD, and six underwent LC. During the study period, the incidence of fever was 79.7% (n = 153) in the aSAH patients. Lung problems were the most common causes of fever (43.1%). The incidence of non-infectious fever was 22.2%. AC was the cause of fever in 5.2% of the 153 patients with fever (Table 1). In the analysis based on the cause of fever, WBC counts, AST levels, and CRP levels were significantly different between patients with and without AC (Table 2).

st common causes of fever (43.1%). The incidence of non-infectious fever was 22.2%. AC was the cause of fever in 5.2% of the 153 patients with fever (Table 1). In the analysis based on the cause of fever, WBC counts, AST levels, and CRP levels were significantly different between patients with and without AC (Table 2). For febrile patients without AC, the mean first day of fever was hospital day 3.57 ± 3.62. The mean first day of fever for febrile patients with AC was hospital day 6.13 ± 2.17 (P = 0.049). The mean initial consecutive fasting times for patients with and without AC were hospital day 5.38 ± 2.78 and hospital day 3.3 ± 2.31, respectively (P = 0.014). The predictors of AC following aSAH were diabetes mellitus and the initial consecutive fasting time (Table 3). Pathologic findings for the six patients who underwent LC are shown in Table 4. Their mean hospital stay was 54.3 ± 24.5 days (range, 30 to 89 days), and the mean intensive care unit stay was 26.8 ± 8.5 days (range, 17 to 37 days). On average, the patients underwent cholecystectomy after 24.5 ± 12.4 days (range, 16 to 47 days) of hospitalization. Each patient had a fever and leukocytosis at the time of diagnosis, and two had abdominal pain. Four patients demonstrated findings consistent with chronic cholecystitis, in addition to AC, during their pathologic examinations, and two patients had necrotic gallbladders evident during surgery and the pathological examinations.

ient had a fever and leukocytosis at the time of diagnosis, and two had abdominal pain. Four patients demonstrated findings consistent with chronic cholecystitis, in addition to AC, during their pathologic examinations, and two patients had necrotic gallbladders evident during surgery and the pathological examinations. Discussion Fever is a source of concern in aSAH patients, and it is associated with increased morbidity and mortality [1,2]. For eight patients in the present study, AC was ultimately determined to be the cause of fever. Previously, rare case reports have described the occurrence of AC in aSAH patients [11], including in one aSAH patient among 15 acalculous cholecystitis patients [9]. Additionally, some studies have suggested that cerebrovascular disease might be a risk factor for the occurrence of AC [7,8]. To the best of our knowledge, this is the first study to evaluate AC as a medical complication during the treatment of aSAH.

g in one aSAH patient among 15 acalculous cholecystitis patients [9]. Additionally, some studies have suggested that cerebrovascular disease might be a risk factor for the occurrence of AC [7,8]. To the best of our knowledge, this is the first study to evaluate AC as a medical complication during the treatment of aSAH. In the present study, four of the six AC patients (66.6%) pathologically proven had acalculous cholecystitis, whereas acalculous cholecystitis generally accounts for only 2%–15% of cholecystitis cases [12]. All eight of our patients had favorable condition grades and were diagnosed with AC following an evaluation for fever, leukocytosis, and abdominal pain during hospitalization in the neurological intensive care unit. Calculous cholecystitis is the result of gallstones blocking the flow of bile in the biliary tree, which leads to gallbladder inflammation [13]. The pathogenesis of acalculous cholecystitis is similar, but may be caused by bile stasis [14]. Bile stasis changes the chemical composition of bile and leads to gallbladder mucosal injury. Increases in bile lysophosphatidylcholine levels strongly influence functional water transport, and may cause gallbladder mucosa inflammation [15]. Other bile components, such as beta-glucuronidase, may also be involved in the pathogenesis of acute acalculous cholecystitis [16]. In our AC patients, the mean initial consecutive fasting period was 5.38 ± 2.78 days, suggesting that bile congestion occurred during the period leading to gallbladder distension and resulting in acute inflammation and cholecystitis [17].

may also be involved in the pathogenesis of acute acalculous cholecystitis [16]. In our AC patients, the mean initial consecutive fasting period was 5.38 ± 2.78 days, suggesting that bile congestion occurred during the period leading to gallbladder distension and resulting in acute inflammation and cholecystitis [17]. Another mechanism leading to acalculous cholecystitis is the intense injury of blood vessels in the gallbladder muscularis and serosa [9] due to massive activation of the sympathetic nervous system in aSAH patients. This, then, leads to catecholamine-mediated injury [18,19]. At high concentrations, catecholamines directly damage the vascular endothelium, resulting in endothelial cell swelling, necrosis, and progressive de-endothelization [20,21]. Therefore, massive activation of the sympathetic nervous system, a feature of aSAH, could lead to acute acalculous cholecystitis.

injury [18,19]. At high concentrations, catecholamines directly damage the vascular endothelium, resulting in endothelial cell swelling, necrosis, and progressive de-endothelization [20,21]. Therefore, massive activation of the sympathetic nervous system, a feature of aSAH, could lead to acute acalculous cholecystitis. Early enteral nutritional support for intensive care unit patients is known to improve their clinical outcomes [22] and is also known to improve the survival of aSAH patients [23]. Frequently, intracranial pressure-induced vomiting, persistent diarrhea, drug-related gastrointestinal mucosal inflammation, and ileus result in prolonged parenteral feeding of these patients. Considering that the reported incidence of acute acalculous cholecystitis during long-term total parenteral nutrition is 30% [24], acute acalculous cholecystitis may have occurred in our patients during their prolonged fasting period. In our study, the initial consecutive fasting time and the presence of diabetes mellitus were predictors for AC following aSAH. Furthermore, opioid analgesics [14] increase intraluminal bile duct pressure by causing sphincter of Oddi spasms. Therefore, the opioid analgesics administered to the aSAH patients may have also contributed to the onset of AC.

ive fasting time and the presence of diabetes mellitus were predictors for AC following aSAH. Furthermore, opioid analgesics [14] increase intraluminal bile duct pressure by causing sphincter of Oddi spasms. Therefore, the opioid analgesics administered to the aSAH patients may have also contributed to the onset of AC. Persistent fever during this time could result in cerebral edema, increased intracranial pressure, exacerbation of ischemia, decreased arterio-jugular differences in oxygen content, and alteration in the level of consciousness that, in turn, would delay recovery [3,4]. Further, all of our patients had persistent fevers despite the use of antibiotics for >3 days. Because of the risks associated with persistent fevers in aSAH patients, we maintained a high degree of suspicion of cholecystitis, allowing prompt diagnoses and the provision of appropriate therapy. Computed tomography was found to be the most appropriate diagnostic tool for examining patients in the intensive care unit because it is not as operator-dependent as abdominal ultrasonography and requires less patient cooperation. Laparoscopic cholecystectomy was performed within 24 hours of diagnosis, in consultation with a general surgeon. PTGBD could have been performed instead of cholecystectomy, and two patients underwent PTGBD. These courses, chosen in consultation with a general surgeon, are supported by studies suggesting that early cholecystectomy is associated with decreased morbidity and cost [25]. In addition, two of our patients demonstrated necrotic gallbladder changes, making cholecystectomy the most appropriate treatment.

ent PTGBD. These courses, chosen in consultation with a general surgeon, are supported by studies suggesting that early cholecystectomy is associated with decreased morbidity and cost [25]. In addition, two of our patients demonstrated necrotic gallbladder changes, making cholecystectomy the most appropriate treatment. Although infections have been identified in 75% of febrile aSAH patients [5], our experience suggests infectious fever is the case of 71.3% of fevers in febrile aSAH patients. AC was the sixth leading cause of fever and the fourth leading cause of infectious fever in our study (Table 1). Thus, an evaluation for possible AC is indicated in any aSAH patient with a persistent fever that does not have an identified source. WBC counts, AST levels, and CRP levels are helpful for identifying cases with suspected AC, according to our results (Table 2). Prompt diagnosis and treatment may significantly affect the patient’s prognosis. Our study is limited by its retrospective design and single-center results. Nonetheless, it suggests the importance of recognizing AC as a cause of fever in aSAH patients. Because fever is a major medical complication for these patients, prompt determination of the source of the fever may reduce patient morbidity and mortality, and improve prognosis.

ospective design and single-center results. Nonetheless, it suggests the importance of recognizing AC as a cause of fever in aSAH patients. Because fever is a major medical complication for these patients, prompt determination of the source of the fever may reduce patient morbidity and mortality, and improve prognosis. In conclusion, our study demonstrated the importance of maintaining a high degree of suspicion for the presence of AC as a cause of fever in aSAH patients. The presence of diabetes mellitus and the initial consecutive fasting time are predictors of AC in aSAH. Elevated WBC counts, AST levels, and CRP levels can help in distinguishing the cause of fever as AC. In addition, we suggest that early enteral nutrition and the avoidance of opioid analgesics may help prevent AC in aSAH patients. No potential conflict of interest relevant to this article was reported. Table 1. Causes of fever in patients with aneurysmal subarachnoid hemorrhage Cause Incidence Lung problema 66 (43.1) Inflammation due to non-infectious causeb 34 (22.2) Catheter-related infection 17 (11.1) Operation-related infectionc 11 (7.2) Drug fever 10 (6.5) Acute cholecystitis 8 (5.2) Thrombophlebitis 4 (2.6) Othersd 3 (2.0) Values are presented as number (%). a Include pneumonia, atelectasis, pleural effusion, and pulmonary edema; b Inflammation due to non-infectious causes may be due to intracranial hemorrhage; c Include cerebrospinal fluid infection and local wound infection; d Include urinary tract infections, pseudomembranous colitis, and bacteremia.

Cause Incidence Lung problema 66 (43.1) Inflammation due to non-infectious causeb 34 (22.2) Catheter-related infection 17 (11.1) Operation-related infectionc 11 (7.2) Drug fever 10 (6.5) Acute cholecystitis 8 (5.2) Thrombophlebitis 4 (2.6) Othersd 3 (2.0) Values are presented as number (%). a Include pneumonia, atelectasis, pleural effusion, and pulmonary edema; b Inflammation due to non-infectious causes may be due to intracranial hemorrhage; c Include cerebrospinal fluid infection and local wound infection; d Include urinary tract infections, pseudomembranous colitis, and bacteremia. Table 2. Comparison of laboratory findings between patients with and without acute cholecystitis

a Include pneumonia, atelectasis, pleural effusion, and pulmonary edema; b Inflammation due to non-infectious causes may be due to intracranial hemorrhage; c Include cerebrospinal fluid infection and local wound infection; d Include urinary tract infections, pseudomembranous colitis, and bacteremia. Table 2. Comparison of laboratory findings between patients with and without acute cholecystitis Variable Acute cholecystitis Mean ± SD Fever due to other causes Mean ± SD P-value WBC 18,497.50 ± 9,426.01 Lung problem 11,651.23 ± 3,766.27 0.001 Catheter-related infection 13,259.41 ± 4,468.37 0.104 Non-infectious cause 11,282.06 ± 4,335.04 0.001 Operation-related infection 12,018.18 ± 3,188.42 0.036 Drug fever 9,703.00 ± 2,611.47 0.001 Phlebitis 12,477.50 ± 2,326.83 0.328 Others 9,000.00 ± 7,014.98 0.034 AST 61.63 ± 63.81 Lung problem 31.36 ± 18.97 0.017 Catheter-related infection 33.06 ± 17.85 0.090 Non-infectious cause 29.39 ± 17.42 0.017 Operation-related infection 31.40 ± 14.82 0.122 Drug fever 21.00 ± 5.66 0.022 Phlebitis 37.00 ± 36.04 0.671 Others 21.00 ± 13.08 0.175 CRP 9.25 ± 10.87 Lung problem 5.43 ± 4.86 0.460 Catheter-related infection 3.60 ± 3.85 0.167 Non-infectious cause 3.00 ± 3.09 0.043 Operation-related infection 4.80 ± 3.46 0.563 Drug fever 4.20 ± 3.88 0.395 Phlebitis 8.25 ± 4.57 1.000 Others 12.00 ± 14.14 0.997 SD: standard deviation; WBC: white blood cell counts; AST: aspartame aminotransferase; CRP: C-reactive protein. Table 3. Logistic regression analysis results for predicting acute cholecystitis in aneurysmal subarachnoid hemorrhage

Variable Acute cholecystitis Mean ± SD Fever due to other causes Mean ± SD P-value WBC 18,497.50 ± 9,426.01 Lung problem 11,651.23 ± 3,766.27 0.001 Catheter-related infection 13,259.41 ± 4,468.37 0.104 Non-infectious cause 11,282.06 ± 4,335.04 0.001 Operation-related infection 12,018.18 ± 3,188.42 0.036 Drug fever 9,703.00 ± 2,611.47 0.001 Phlebitis 12,477.50 ± 2,326.83 0.328 Others 9,000.00 ± 7,014.98 0.034 AST 61.63 ± 63.81 Lung problem 31.36 ± 18.97 0.017 Catheter-related infection 33.06 ± 17.85 0.090 Non-infectious cause 29.39 ± 17.42 0.017 Operation-related infection 31.40 ± 14.82 0.122 Drug fever 21.00 ± 5.66 0.022 Phlebitis 37.00 ± 36.04 0.671 Others 21.00 ± 13.08 0.175 CRP 9.25 ± 10.87 Lung problem 5.43 ± 4.86 0.460 Catheter-related infection 3.60 ± 3.85 0.167 Non-infectious cause 3.00 ± 3.09 0.043 Operation-related infection 4.80 ± 3.46 0.563 Drug fever 4.20 ± 3.88 0.395 Phlebitis 8.25 ± 4.57 1.000 Others 12.00 ± 14.14 0.997 SD: standard deviation; WBC: white blood cell counts; AST: aspartame aminotransferase; CRP: C-reactive protein. Table 3. Logistic regression analysis results for predicting acute cholecystitis in aneurysmal subarachnoid hemorrhage Variable β Coefficient SE β Wald statistic OR P-value Initial consecutive fasting time 0.282 0.125 5.078 1.325 0.024 Diabetes mellitus 2.170 1.016 4.564 8.758 0.033 SE: standard error; OR: odds ratio. Table 4. Pathologic findings of patients who underwent laparoscopic cholecystectomy

Table 3. Logistic regression analysis results for predicting acute cholecystitis in aneurysmal subarachnoid hemorrhage Variable β Coefficient SE β Wald statistic OR P-value Initial consecutive fasting time 0.282 0.125 5.078 1.325 0.024 Diabetes mellitus 2.170 1.016 4.564 8.758 0.033 SE: standard error; OR: odds ratio. Table 4. Pathologic findings of patients who underwent laparoscopic cholecystectomy Number Age (yr) Sex Pathologic findings 1 76 F Chronic inflammation, erosion, gall stone (–) 2 58 F Acute inflammation, chronic, gall stone (+) 3 61 F Chronic inflammation, gall stone (–) 4 57 F Acute inflammation, edema, fibrosis, gall stone (–) 5 62 F Acute and chronic inflammation, necrotizing, gall stone (+) 6 79 F Acute and chronic inflammation, mucosa necrotic appearance, gall stone (–)

Dear Editor: We report a case of 20-year-old female patient with ofloxacin-induced anaphylaxis mediated by IgG4 antibody. One of second-generation fluoroquinolone, ofloxacin is prescribed widely to treat bacterial infections. Reports of serious hypersensitivity reactions to quinolone are increasing due to high consumption worldwide [1]. The pathogenesis of anaphylaxis caused by ofloxacin is not yet fully understood, as it is not commonly reported. There has been a report demonstrating high serum specific IgE levels to ofloxacin-human serum albumin (HSA) conjugate using enzymelinked immunosorbent assay (ELISA), enzyme-linked immunospecific assay [2]; however, this is the first report to suggest an IgG4-mediated but not IgE-mediated mechanism in a patient with ofloxacin-induced anaphylaxis.

onstrating high serum specific IgE levels to ofloxacin-human serum albumin (HSA) conjugate using enzymelinked immunosorbent assay (ELISA), enzyme-linked immunospecific assay [2]; however, this is the first report to suggest an IgG4-mediated but not IgE-mediated mechanism in a patient with ofloxacin-induced anaphylaxis. A 20-year-old female with allergic rhinitis was referred to our department for the evaluation of anaphylaxis because she had experienced generalized urticaria and dyspnea with wheezing, and hoarseness within 1 hour after oral ingestion of ofloxacin 100 mg. It was her first ingestion of ofloxacin. The patient had a history of acute, generalized urticaria and drug allergy to nonsteroidal anti-inflammatory drugs. Her mother had a history of bronchial asthma. At the initial visit, no abnormal findings were noted on physical examination and radiography/spirometry results. Her serum total IgE level was 42 kU/L. Allergy skin prick tests showed positive responses to Dermatophagoides farinae and D. pteronyssinus, but negative responses to ofloxacin at concentrations of 0.1−10 mg/ml, with a positive control being a mean wheal size of 4 mm to histamine.

ion and radiography/spirometry results. Her serum total IgE level was 42 kU/L. Allergy skin prick tests showed positive responses to Dermatophagoides farinae and D. pteronyssinus, but negative responses to ofloxacin at concentrations of 0.1−10 mg/ml, with a positive control being a mean wheal size of 4 mm to histamine. Ofloxacin-HSA conjugate was conducted to investigate immunologic mechanisms to detect serum specific IgE antibody to ofloxacin-HSA conjugate using ELISA as previously described [2]. When the positive cutoff value was determined from the mean +3 standard deviation absorbance values of 20 nonatopic healthy controls that never experienced any other drug allergy according to medical records, serum specific IgE antibody was not detected in the patient’s serum, while high serum specific IgG4 antibody was detected as shown in Figure 1. A basophil activation test (BAT) was done with addition of ofloxacin and anti-IgG4 antibody using peripheral basophils from the patient and three healthy controls to confirm specific IgG4-mediated mechanisms. A significant upregulation of CD203c, a marker of activated basophils, was noted with serial additions of ofloxacin-HSA (from 6.35% to 9.7%), while no changes were noted in three healthy controls (Figure 2). Until now, anaphylaxis cases caused by ciprofloxacin, one of the most frequently prescribed oral fluoroquinolones in the U.S. and worldwide, have occasionally been reported [3,4]. The frequency of ofloxacin prescription as an alternative therapy is expected to increase due to a high incidence of drug-resistant tuberculosis in Asia.

aphylaxis cases caused by ciprofloxacin, one of the most frequently prescribed oral fluoroquinolones in the U.S. and worldwide, have occasionally been reported [3,4]. The frequency of ofloxacin prescription as an alternative therapy is expected to increase due to a high incidence of drug-resistant tuberculosis in Asia. Until now, immediate reactions such as urticaria and anaphylaxis were reported as a phenotype of hypersensitivity reactions to quinolone, although frequencies have been reported to be less than 2% [1]. An IgE-mediated reaction is known as a major pathogenetic mechanism of immediate hypersensitivity to quinolone. Using sepharoseradioimmunoassay for the determination of specific IgE in the serum, a previous study reported an IgE-mediated mechanism in 12 of 38 (31.5%) patients with anaphylaxis and urticaria as well as serum specific IgE to ciprofloxacin, moxifloxacin, and levofloxacin [5]. Another study showed in 30 of 55 (54.5%) patients with immediate reactions to eight quinolones including ofloxacin detected by serum specific IgE [6]. In cases of ofloxacin-induced anaphylaxis, our previous study reported four of five (80%) patients with serum specific IgE to ofloxacin-HSA conjugate using ELISA [2].

]. Another study showed in 30 of 55 (54.5%) patients with immediate reactions to eight quinolones including ofloxacin detected by serum specific IgE [6]. In cases of ofloxacin-induced anaphylaxis, our previous study reported four of five (80%) patients with serum specific IgE to ofloxacin-HSA conjugate using ELISA [2]. Previous studies have suggested involvement of non-IgE mediated mechanisms in patients with quinolone-induced immediate hypersensitivity reactions [1,7]. Nam et al. [8] reported one patient with cefotetan-induced anaphylaxis that was mediated by IgG antibody. The non-IgE mediated mechanisms of anaphylaxis are not well understood in cases of quinolone-induced anaphylaxis. Since IgG-mediated immediate anaphylaxis through the low-affinity IgG receptor Fcγ-receptor was introduced in a murine model [9], IgG-mediated alternative pathways are suggested, as basophil to release platelet-activating factor upon stimulation with allergen-IgG complexes [10]. It is possible that alternative pathways could be mediated by basophils, IgG, IgG receptor, and platelet-activating factor, particularly a high level of serum IgG, but not specific IgE to relevant allergens. The BAT is a useful diagnostic method for immediate-type drug allergy based upon different activation markers, being mostly CD203c and CD63 involving IgE-dependent and IgE-independent basophil-mediated reactions [11-13]. The sensitivity of the BAT to quinolone has recently been reported to 71.1% [5]. In the present study, we presumed that specific IgG4 antibody may play a role in the activation of basophil as a significant mediator through an alternative mechanism in that there were a high serum specific IgG4 and no serum specific IgE to ofloxacin-HSA conjugate and a significant upregulation of basophils in response to ofloxacin and anti-IgG4 antibody.

specific IgG4 antibody may play a role in the activation of basophil as a significant mediator through an alternative mechanism in that there were a high serum specific IgG4 and no serum specific IgE to ofloxacin-HSA conjugate and a significant upregulation of basophils in response to ofloxacin and anti-IgG4 antibody. Anaphylaxis is a serious reaction that may lead to death and prompt management is crucial [14]. The mainstay of anaphylaxis treatment is to restore and maintain vital signs by early administration of intramuscular epinephrine along with antihistamine and corticosteroid. Regardless of allergen or mechanism of anaphylaxis, mast cell and basophil are activated to commence and amplify allergic reactions, and release various inflammatory mediators such as tryptase and histamine to cause anaphylaxis reactions [15]. In conclusion, this is a case of ofloxacin-induced anaphylaxis through an IgG4-mediated but not IgE-medicated basophil activation mechanism. No potential conflict of interest relevant to this article was reported. Figure 1. Detection of serum specific IgG4 antibodies to ofloxacin-human serum albumin (HSA) conjugate by ELISA from the patient (●) and 20 non-atopic normal controls (○). OD: optical density; ELISA: enzyme-linked immunosorbent assay. Figure 2. Results of basophil activation test with addition of ofloxacin-human serum albumin (HSA) conjugate (A) and anti-IgG4 antibodies (B) in the patient (●) and three healthy controls (◇, △, ∇).

Introduction Acute kidney injury (AKI) is a common complication in patients admitted to the intensive care unit (ICU); it is estimated to occur in 30%–50% of admissions [1]. Compared to individuals with stable kidney function, AKI patients have poor outcomes, particularly those receiving renal replacement therapy [1,2]. Among modalities of renal replacement therapy, continuous renal replacement therapy (CRRT) is preferred in the ICU setting because it is best suited for hemodynamically unstable patients with multiple organ failure. Previous studies have investigated the impact of CRRT timing on outcomes, but no significant results were reported [3,4]. In contrast, volume status in AKI patients appears to have significant effects on outcome [5-7]. Fluid accumulation can result in multiple organ failure and death; thus, aggressive volume control using diuretics or renal replacement therapy is suggested in critically ill patients with AKI.

results were reported [3,4]. In contrast, volume status in AKI patients appears to have significant effects on outcome [5-7]. Fluid accumulation can result in multiple organ failure and death; thus, aggressive volume control using diuretics or renal replacement therapy is suggested in critically ill patients with AKI. Precise assessment of volume status is necessary to maintain euvolemic status. Approaches such as monitoring changes in body weight or central venous pressure are frequently used, but have poor sensitivity [8,9]. Therefore, researchers have attempted to find an ideal method to determine volume status. Laboratory tests including assessment of brain natriuretic peptide (BNP) and N-terminal pro-BNP levels, which reflect cardiac reactions to volume load, have been used to assess hydration status [10]. However, various studies have challenged the utility of BNP for assessment of volume status and predicting outcomes [11-13]. Bioelectrical impedance analysis (BIA) is a promising noninvasive and convenient method for assessing volume status [14,15]. Previous studies have suggested that, in hemodialysis patients, volume status management using BIA could be achieved, and blood pressure might be improved [14,15]. Nevertheless, few studies have investigated the usefulness of BIA in critically ill patients, and more data are needed to demonstrate the associations between parameters obtained from BIA and outcomes for clinical application.

status management using BIA could be achieved, and blood pressure might be improved [14,15]. Nevertheless, few studies have investigated the usefulness of BIA in critically ill patients, and more data are needed to demonstrate the associations between parameters obtained from BIA and outcomes for clinical application. In this prospective observational study, we identified AKI patients who required CRRT and evaluated the degree of volume overload as the ratio of extracellular water to total body water (ECW/TBW) obtained from a segmental, multifrequency BIA device, which seems to be more accurate than whole-body BIA [16]. We then investigated the utility of ECW/TBW for predicting outcomes in CRRT patients in comparison with other variables. Materials and Methods 1) Patients Patients who needed CRRT for AKI were screened between February and November in 2014 in a 750-bed tertiary referral center in Seoul, Korea. AKI was defined as an abrupt loss of kidney function requiring CRRT that developed within 7 days. The study included adult patients (aged ≥18 years) who were admitted for medical illness. We excluded patients who had received kidney transplantation or who had been treated with hemodialysis or peritoneal dialysis for end-stage renal disease. A total of 73 patients were eligible. Of those, 31 patients were included in the study; 42 patients who did not wish to participate in this study or for whom BIA was not performed were excluded.

received kidney transplantation or who had been treated with hemodialysis or peritoneal dialysis for end-stage renal disease. A total of 73 patients were eligible. Of those, 31 patients were included in the study; 42 patients who did not wish to participate in this study or for whom BIA was not performed were excluded. The study was performed according to the Helsinki Declaration and was approved by the institutional review board (No. C2012195[890]) of Chung-Ang University Hospital where the study was performed. Written informed consent was obtained from all eligible patients or next of kin before enrollment. Because this study was observational, the treatment protocol, including CRRT and medication, was not changed regardless of enrollment. 2) CRRT protocol CRRT was initiated by decision of the nephrologist if patients who could not bear intermittent hemodialysis due to unstable vital signs, refractory pulmonary edema, intractable hyperkalemia or metabolic acidosis, uremic symptoms including pericarditis and encephalopathy, or oliguria with progressive azotemia. Continuous hemodiafiltration was carried out using a Prisma (Baxter, Lund, Sweden) or Prismaflex machine with a high flux hemofilter (ST100, Baxter) and dialysate and replacement fluids (Hemosol B0, Baxter). Anticoagulation was conducted with nafamostat mesilate (SK Chemicals, Seoul, Korea). The target dose of CRRT was 40 ml/kg/h.

emodiafiltration was carried out using a Prisma (Baxter, Lund, Sweden) or Prismaflex machine with a high flux hemofilter (ST100, Baxter) and dialysate and replacement fluids (Hemosol B0, Baxter). Anticoagulation was conducted with nafamostat mesilate (SK Chemicals, Seoul, Korea). The target dose of CRRT was 40 ml/kg/h. 3) Volume status analysis Body composition was assessed using a segmental and multifrequency BIA device (Inbody S10; Biospace, Seoul, Korea) before CRRT initiation. Eight electrodes were placed on the surface of the thumb, fingers of the hand, and ball of the foot and heel with the patient in the supine position. We collected ECW and TBW data from BIA and then calculated ECW/TBW to estimate the volume status of subjects.

(Inbody S10; Biospace, Seoul, Korea) before CRRT initiation. Eight electrodes were placed on the surface of the thumb, fingers of the hand, and ball of the foot and heel with the patient in the supine position. We collected ECW and TBW data from BIA and then calculated ECW/TBW to estimate the volume status of subjects. 4) Data collection All data were collected from electronic medical records. Baseline demographic and clinical data included age, sex, comorbidities, causes of admission, reasons for AKI, and CRRT indications. Body weight was measured by a bed scale before conducting BIA and CRRT. Laboratory findings at CRRT initiation included hematocrit, albumin, C-reactive protein, lactate, BNP, and total carbon dioxide (CO2). To calculate the Sequential Organ Failure Assessment (SOFA) score, related parameters were also collected [17]. All available intake and output data between ICU admission and CRRT initiation were collected. Intake was composed of oral and parenteral fluid administered, and output included urine, gastrointestinal losses, and drains. Using these data, cumulative and daily fluid balance was calculated from ICU admission to CRRT initiation. Oliguria was defined if urine output for 24 hours before CRRT was under 400 ml/d.

ted. Intake was composed of oral and parenteral fluid administered, and output included urine, gastrointestinal losses, and drains. Using these data, cumulative and daily fluid balance was calculated from ICU admission to CRRT initiation. Oliguria was defined if urine output for 24 hours before CRRT was under 400 ml/d. 5) Statistical analysis Continuous variables were expressed as median (interquartile range) and were compared using the Wilcoxon rank-sum test. Categorical variables were expressed as number (percentage) and were compared between groups using the chi-square test. To investigate whether volume status can influence outcomes, patients were divided into two groups using an ECW/TBW of 0.41 as the cutpoint, which was the approximate median value. Cumulative incidence of death from initiation of CRRT was estimated by the Kaplan-Meier method and was compared between groups using the log-rank test. Cox regression analyses were conducted to investigate the associations between variables and mortality. In addition, we evaluated the receiver operating characteristics (ROC) curve for predicting mortality and determined the area under the curve (AUC). All statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). A two-sided P-value <0.05 was considered to be significant.

s and mortality. In addition, we evaluated the receiver operating characteristics (ROC) curve for predicting mortality and determined the area under the curve (AUC). All statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA). A two-sided P-value <0.05 was considered to be significant. Results 1) Baseline characteristics A total of 31 ICU patients receiving CRRT were analyzed in this study. There were 18 men (58.1%), and the median age was 67 years (interquartile range, 51 to 78 years). Of the included patients, 14 (45.2%) died within 28 days after CRRT initiation. We compared characteristics between 17 survivors and 14 non-survivors to determine factors associated with mortality (Table 1). There were no differences in age, sex, comorbidities, causes of admission, etiologies of AKI, or CRRT indications between these two groups. Total SOFA score, daily fluid balance, and serum albumin level differed between survivors and non-survivors (P = 0.040, P = 0.048, and P = 0.007, respectively). In addition, ECW/TBW was lower in survivors than in non-survivors (0.40; interquartile range, 0.40 to 0.42 vs. 0.42; interquartile range, 0.41 to 0.44; P = 0.044). 2) Relationship between BIA-assessed volume status and mortality Survival rate was compared between the two groups. Figure 1 shows the results of survival analysis. As shown, less hydrated patients had a better survival rate than the more hydrated patients (P = 0.044). The 28-day survival rate was 73.3% in patients with ECW/TBW <0.41 and 36.0% in those with ECW/TBW ≥0.41.

mortality Survival rate was compared between the two groups. Figure 1 shows the results of survival analysis. As shown, less hydrated patients had a better survival rate than the more hydrated patients (P = 0.044). The 28-day survival rate was 73.3% in patients with ECW/TBW <0.41 and 36.0% in those with ECW/TBW ≥0.41. To determine the hazard ratio (HR) for mortality, several variables, including ECW/TBW, were analyzed by Cox regression analysis (Table 2). Age, sex, and oliguria were not associated with mortality (P = 0.577, P = 0.810, and P = 0.440, respectively). Total SOFA score and daily fluid balance were also not related to mortality, despite weak associations (HR, 1.2; 95% confidence interval [CI], 1.0 to 1.6; P = 0.079 and HR, 1.3; 95% CI, 1.0 to 1.7; P = 0.072). Serum level of lactate and BNP at the time of CRRT initiation did not predict mortality (P = 0.060 and P = 0.331, respectively). Patients with ECW/TBW ≥0.41 had decreased survival rates, but the difference was not statistically significant (HR, 3.0; 95% CI, 0.9 to 9.8; P = 0.061).

, 1.3; 95% CI, 1.0 to 1.7; P = 0.072). Serum level of lactate and BNP at the time of CRRT initiation did not predict mortality (P = 0.060 and P = 0.331, respectively). Patients with ECW/TBW ≥0.41 had decreased survival rates, but the difference was not statistically significant (HR, 3.0; 95% CI, 0.9 to 9.8; P = 0.061). 3) Using ECW/TBW to predict mortality We evaluated the ROC curve to identify variables that predicted mortality. Among various factors, total SOFA score, daily fluid balance, and serum level of lactate predicted mortality. The AUC of these was 0.73 (95% CI, 0.53 to 0.92; P = 0.039), 0.75 (95% CI, 0.56 to 0.93; P = 0.025), and 0.75 (95% CI, 0.57 to 0.93; P = 0.024), respectively. In addition, the ROC of ECW/TBW at the time of CRRT initiation was estimated (Figure 2). The AUC of ECW/TBW was 0.73 (95% CI, 0.54 to 0.92) for mortality (P = 0.037). An optimal cutoff was 0.413, which showed a sensitivity and specificity of 71.4% and 70.6%, respectively.

95% CI, 0.57 to 0.93; P = 0.024), respectively. In addition, the ROC of ECW/TBW at the time of CRRT initiation was estimated (Figure 2). The AUC of ECW/TBW was 0.73 (95% CI, 0.54 to 0.92) for mortality (P = 0.037). An optimal cutoff was 0.413, which showed a sensitivity and specificity of 71.4% and 70.6%, respectively. Discussion In this prospective observational study, we investigated the utility of using BIA to assess volume status as the ratio of ECW/TBW in critically ill patients receiving CRRT. Non-survivors had a relatively higher ECW/TBW than survivors. In addition, we evaluated the ability of ECW/TBW to predict mortality, and patients who had a higher ECW/TBW had increased risk of 28-day death, and the HR for death was considerable (HR, 3.0; 95% CI, 0.9 to 9.8), albeit without statistical significance. Last, we explored the ROC curve and found that the AUC of ECW/TBW was 0.73 (95% CI, 0.54 to 0.92) for predicting 28-day mortality in patients receiving CRRT.

a higher ECW/TBW had increased risk of 28-day death, and the HR for death was considerable (HR, 3.0; 95% CI, 0.9 to 9.8), albeit without statistical significance. Last, we explored the ROC curve and found that the AUC of ECW/TBW was 0.73 (95% CI, 0.54 to 0.92) for predicting 28-day mortality in patients receiving CRRT. Fluid accumulation has been identified as an independent risk factor for death in patients with AKI [5,6,18,19]. Furthermore, a previous study showed that an excessively positive fluid balance prior to CRRT can result in organ failure and death in critically ill patients who receive CRRT [7]. Volume status needs to be monitored before deciding on fluid therapy. However, isotope dilution, the gold-standard method to determine volume status, is difficult to perform in clinical practice [20]. In addition, traditional clinical approaches such as monitoring body weight or central venous pressure are known to have poor sensitivity [8,9]. This study comprised 22 patients whose central venous pressure was measured, but we found that these values did not predict mortality in the ROC curve (AUC, 0.72; P = 0.082). Fluid balance between the input and output is widely used in practice. Although this study again confirmed that fluid balance was associated with mortality, an assessment of fluid balance using input and output data could be incorrect, especially in ICU patients [21,22]. Thus, various other methods have been explored. Level of BNP has been used to assess volume status [10], but previous reports found that BNP level did not predict outcomes [11-13], similar to our findings. BIA is a promising method to assess fluid status because it allows analysis of body composition and has the advantages of safety, convenience, noninvasiveness, and low cost [14,15,23-25]. Nevertheless, its role in critically ill patients has not been widely explored [26,27]. In this study, we investigated the utility of BIA to assess volume status in AKI patients who receive CRRT.

ows analysis of body composition and has the advantages of safety, convenience, noninvasiveness, and low cost [14,15,23-25]. Nevertheless, its role in critically ill patients has not been widely explored [26,27]. In this study, we investigated the utility of BIA to assess volume status in AKI patients who receive CRRT. The principle underlying BIA is the analysis of the vectors of resistance and reactance resulting from the passage of electric currents of low amplitude and low and high frequencies through an organism [28,29]. Several parameters including intracellular water, ECW, and TBW are derived from analysis of these resistance and reactance vectors. ECW/TBW is a simple-to-measure, straightforward parameter that allows determination of the degree of hydration because excess volume accumulates primarily as ECW [30]. The present study randomly used the approximate median value, ECW/TBW of 0.41, as the cutoff point to divide patients into two groups. Although the comparisons according to the ECW/TBW showed significant differences, this ratio is not an absolute cutoff for defining volume overload because it can be affected by various factors including age, sex, and comorbidities [20,30]. Given this problem, further studies are needed to establish an optimal cutoff value for diagnosing volume overload in ICU patients.

owed significant differences, this ratio is not an absolute cutoff for defining volume overload because it can be affected by various factors including age, sex, and comorbidities [20,30]. Given this problem, further studies are needed to establish an optimal cutoff value for diagnosing volume overload in ICU patients. We investigated whether volume status assessed by BIA could predict mortality. We evaluated the time to 28-day mortality and found worse outcomes in more hydrated patients compared to less hydrated patients. That is, fluid balance before CRRT initiation might be better in patients with lower ECW/TBW than those with higher ECW/TBW, which could determine the results, although CRRT indications such as pulmonary edema did not differ. However, Cox regression analysis did not show a relationship between higher ECW/TBW and higher mortality due to small sample size, although it trended in the predicted direction. Additionally, we drew an ROC curve and obtained a significant AUC of ECW/TBW for predicting mortality. This result also shows that the ratio might be comparable to daily fluid balance, SOFA score, and serum lactate level which are known as predictors for mortality [6,7,18,31,32], although interactions could occur. Previous studies have investigated the usefulness of BIA in critically ill patients with or without AKI and have suggested that BIA can be a reliable method to determine fluid status and to predict outcomes [33-35]. However, these studies used whole-body BIA devices to estimate volume status. The present study investigated the usefulness of a segmental, multifrequency BIA device, which seems to be more accurate than whole-body BIA [16]. In addition to those studies, our study added the potentials of a segmental, multifrequency BIA for determining fluid status and predicting outcomes in CRRT patients. To confirm the utility of ECW/TBW, further large studies are required in ICU patients.

device, which seems to be more accurate than whole-body BIA [16]. In addition to those studies, our study added the potentials of a segmental, multifrequency BIA for determining fluid status and predicting outcomes in CRRT patients. To confirm the utility of ECW/TBW, further large studies are required in ICU patients. This study has several limitations. First, our small sample size limited our statistical power. Although we observed trends in the data in the direction expected, some results were not statistically significant. Furthermore, we could not perform multivariate analyses because of our small sample size. Some variables such as SOFA score and serum lactate could be confounders of mortality but could not be assessed here. Second, this study was observational; thus, we did not adjust fluid balance according to the BIA results. To determine if BIA-guided adjustment of volume status can improve outcomes in ICU patients, further controlled trials with large sample sizes are necessary. In conclusion, this prospective observational study investigated the utility of BIA-measured volume status in critically ill patients who received CRRT due to AKI in order to predict mortality. Our results suggest that hydration status can be determined using BIA, and that overhydration at the time of CRRT might be a predictor of mortality. Because of the small sample size and observational characteristics of this study, future trials are needed to confirm the usefulness of BIA in ICU patients.

ict mortality. Our results suggest that hydration status can be determined using BIA, and that overhydration at the time of CRRT might be a predictor of mortality. Because of the small sample size and observational characteristics of this study, future trials are needed to confirm the usefulness of BIA in ICU patients. No potential conflict of interest relevant to this article was reported. This work was supported by a grant from Baxter, Korea. The funders had no role in the design, collection, analysis or interpretation of this study. This work was supported by a grant from the National Research Foundation of Korea (NRF) (No. NRF- 2012R1A1A1011816) funded by the Korean government, and a grant from Baxter, Korea. Figure 1. Survival rate according to volume status in critically ill patients who received continuous renal replacement therapy. Survival rate was compared between the two groups. As shown, patients with ECW/TBW ≥0.41 had a lower survival rate than those with ECW/ TBW <0.41 (P = 0.044). Survival rate after 28 days was 73.3% in less hydrated patients and 36.0% in more hydrated patients. ECW/ TBW: the ratio of extracellular water to total body water. Figure 2. Receiver operating characteristics curve of volume status estimated by bioelectrical impedance analysis. ECW/TBW appeared to have the potential to predict mortality (P = 0.037). The area under the curve of ECW/TBW for 28-day mortality was 0.73 (95% confidence interval, 0.54 to 0.92) in this study. ECW/ TBW: the ratio of extracellular water to total body water.

f volume status estimated by bioelectrical impedance analysis. ECW/TBW appeared to have the potential to predict mortality (P = 0.037). The area under the curve of ECW/TBW for 28-day mortality was 0.73 (95% confidence interval, 0.54 to 0.92) in this study. ECW/ TBW: the ratio of extracellular water to total body water. Table 1. Baseline characteristics of AKI patients who received CRRT

f volume status estimated by bioelectrical impedance analysis. ECW/TBW appeared to have the potential to predict mortality (P = 0.037). The area under the curve of ECW/TBW for 28-day mortality was 0.73 (95% confidence interval, 0.54 to 0.92) in this study. ECW/ TBW: the ratio of extracellular water to total body water. Table 1. Baseline characteristics of AKI patients who received CRRT Variable Survivor (n = 17) Non-survivor (n = 14) P-value Age (yr) 67 (42–78) 69 (53–79) 0.653 Male 10 (58.8) 8 (57.1) 0.925 Comorbidity Hypertension 7 (41.2) 4 (28.6) 0.707 Diabetes 6 (35.3) 7 (50.0) 0.409 Liver disease 3 (17.6) 4 (28.6) 0.671 Heart failure 3 (17.6) 2 (14.3) 1.000 Reason for admission 0.560 Infection 9 (52.9) 9 (64.3) Cardiovascular 2 (11.8) 3 (21.4) Gastrointestinal 4 (23.5) 1 (7.1) Others 2 (11.8) 1 (7.1) Duration from ICU to CRRT (d) 1 (0–3) 1 (1–3) 0.246 Cause of AKIa 0.603 Septic 9 (52.9) 9 (64.3) Cardiorenal 1 (5.9) 2 (14.3) Ischemic 4 (23.5) 2 (14.3) Hepatorenal 2 (11.8) 0 Others 1 (5.9) 1 (7.1) CRRT indication Pulmonary edema 8 (47.1) 8 (57.1) 0.576 Hyperkalemia 3 (17.6) 4 (28.6) 0.671 Metabolic acidosis 7 (41.2) 8 (57.1) 0.376 Uremic symptom 6 (35.3) 5 (35.7) 0.981 SOFA score Respiratory 3 (2–3) 3 (3–3) 0.518 Cardiovascular 3 (0–4) 3 (1–4) 0.421 Nervous 2 (1–4) 4 (3–4) 0.026 Liver 0 (0–1) 0 (0–2) 0.493 Coagulation 1 (0–2) 1 (0–2) 0.518 Renal 3 (2–4) 2 (1–3) 0.186 Total 11 (8–13) 14 (11–15) 0.040 Oliguria (<400 ml/d) 9 (52.9) 10 (71.4) 0.461 Daily fluid balance (L/d) 1.3 (0.4–2.9) 2.3 (1.8–3.7) 0.048 Cumulative fluid balance (L) 2.2 (0.5–5.1) 5.9 (4.3–7.5) 0.008 Hematocrit (%) 31.1 (24.1–41.3) 28.8 (26.4–33.6) 0.769 Albumin (g/dl) 2.9 (2.6–3.5) 2.3 (2.1–2.8) 0.007 CRP (mg/L) 129.0 (64.6–151.4) 129.1 (67.0–216.0) 0.681 Lactate (mmol/L) 1.8 (1.0–5.2) 3.0 (2.1–8.4) 0.053 BNP (pg/ml) 122.5 (75.5–830.7) 570.4 (130.3–807.6) 0.201 Total CO2 (mmol/L) 15.5 (12.1–20.5) 13.9 (10.7–17.0) 0.316 ECW/TBW 0.40 (0.40–0.42) 0.42 (0.41–0.44) 0.044 Values are presented as median (interquartile range) or number (%).

(64.6–151.4) 129.1 (67.0–216.0) 0.681 Lactate (mmol/L) 1.8 (1.0–5.2) 3.0 (2.1–8.4) 0.053 BNP (pg/ml) 122.5 (75.5–830.7) 570.4 (130.3–807.6) 0.201 Total CO2 (mmol/L) 15.5 (12.1–20.5) 13.9 (10.7–17.0) 0.316 ECW/TBW 0.40 (0.40–0.42) 0.42 (0.41–0.44) 0.044 Values are presented as median (interquartile range) or number (%). AKI: acute kidney injury; CRRT: continuous renal replacement therapy; ICU: intensive care unit; SOFA: Sequential Organ Failure Assessment; CRP: C-reactive protein; BNP: brain natriuretic protein; CO2: carbon dioxide; ECW/TBW: the ratio of extracellular water to total body water. a AKI refers to an abrupt loss of kidney function requiring CRRT that develops within 7 days. Table 2. HRs for 28-day mortality in AKI patients who received CRRT Variable HR (95% confidence interval) P-value Age 1.0 (1.0–1.0) 0.577 Sex 0.9 (0.3–2.5) 0.810 SOFA 1.2 (1.0–1.6) 0.079 Daily fluid balance 1.3 (1.0–1.7) 0.072 Lactate 1.1 (1.0–1.3) 0.060 BNP (>500 pg/ml) 1.7 (0.6–4.9) 0.331 ECW/TBW (≥0.41) 3.0 (0.9–9.8) 0.061 HR: hazard ratio; AKI: acute kidney injury; CRRT: continuous renal replacement therapy; SOFA: Sequential Organ Failure Assessment; BNP: brain natriuretic protein; ECW/TBW: the ratio of extracellular water to total body water.

Introduction Patient safety and healthcare quality outcomes have been areas of increasing interest in Korea. In-hospital cardiopulmonary arrest is a known major patient safety and healthcare quality indicator. In most cases, clinical deterioration commonly precedes cardiopulmonary arrest [1-6]. Accordingly, early recognition of the signs and symptoms of deterioration could reduce the incidence of cardiopulmonary arrest, and this is the basis of rapid response systems (RRSs) [7,8]. In Korea, RRS implementation has been targeted at large academic medical centers. Although there have been some single-center studies, no nationwide survey has investigated the effect of RRSs on in-hospital cardiopulmonary arrest in Korea. Therefore, it is still necessary to estimate the effect of RRSs on cardiopulmonary arrest in domestic hospitals. Two general hospitals in Korea implemented RRSs in 2014. The present study aimed to compare in-hospital cardiopulmonary arrest rates of institutions with or without RRSs, as well as to determine the cardiopulmonary arrest rates both before and after RRS implementation. We also used a pilot study to ascertain the prevalence of in-hospital cardiopulmonary arrest at tertiary hospitals, which was followed with a nationwide survey.

pulmonary arrest rates of institutions with or without RRSs, as well as to determine the cardiopulmonary arrest rates both before and after RRS implementation. We also used a pilot study to ascertain the prevalence of in-hospital cardiopulmonary arrest at tertiary hospitals, which was followed with a nationwide survey. Materials and Methods 1) Study design and institution selection This was a retrospective, multi-center study that used in-hospital cardiopulmonary arrest data from 14 tertiary hospitals from January 2013 to December 2015. Among the existing 43 tertiary hospitals nationwide, the study included data from two institutions with RRSs and 12 facilities without RRSs. We excluded two hospitals with more than 1,500 beds and one hospital with less than 700 beds. We also excluded five hospitals that had introduced an RRS before January 2013 and two hospitals that implemented an RRS after January 2015 (Figure 1). From the remaining 33 hospitals, we selected 16 medical centers in consideration of their provincial arrangement, bed capacity, and ownership type using cluster sampling and judgment sampling. These methods are appropriate when the sample size is relatively small. The population of all tertiary hospitals was believed to have reached or exceeded a certain level of healthcare quality. Of the 16 selected medical centers, 14 directors or physicians in charge of intensive care units were willing to participate in this study.

s are appropriate when the sample size is relatively small. The population of all tertiary hospitals was believed to have reached or exceeded a certain level of healthcare quality. Of the 16 selected medical centers, 14 directors or physicians in charge of intensive care units were willing to participate in this study. To identify the current in-hospital cardiopulmonary arrest rate in tertiary hospitals in Korea, we estimated the cardiopulmonary arrest rate of hospitals with RRSs and also those without RRSs per bed capacity, type of ownership, and geographic distribution.

s are appropriate when the sample size is relatively small. The population of all tertiary hospitals was believed to have reached or exceeded a certain level of healthcare quality. Of the 16 selected medical centers, 14 directors or physicians in charge of intensive care units were willing to participate in this study. To identify the current in-hospital cardiopulmonary arrest rate in tertiary hospitals in Korea, we estimated the cardiopulmonary arrest rate of hospitals with RRSs and also those without RRSs per bed capacity, type of ownership, and geographic distribution. 2) Data collection and in-hospital cardiopulmonary arrest rates Each hospital included in the study used varying types of cardiopulmonary resuscitation (CPR) reports based on their own definition of cardiopulmonary arrest, which differed by location, cardiac rhythm, and patient population. Broadly, these definitions encompass circumstances such as the cessation of cardiac mechanical activity, absence of a detectable pulse, unresponsiveness, and apnea [9]. In this study, in-hospital cardiopulmonary arrest cases included those in the general ward; cases where the arrest took place in the emergency department, operating room, or intensive care unit were excluded. The inhospital cardiopulmonary arrest rate was defined as the CPR rate/1,000 adult admissions. The statistics on the CPR rate of each hospital were gathered from CPR reports and statistics about adult admissions and discharges. We collected detailed data from each hospital from 2013 to 2015 regarding the cardiopulmonary arrest rate of patients at least 18 years of age in the general ward, the number of annual adult admissions, and whether a dedicated intensivist was present. Additionally, rapid response team characteristics were analyzed at the two institutions that had implemented RRSs.

from 2013 to 2015 regarding the cardiopulmonary arrest rate of patients at least 18 years of age in the general ward, the number of annual adult admissions, and whether a dedicated intensivist was present. Additionally, rapid response team characteristics were analyzed at the two institutions that had implemented RRSs. To verify the clinical effect of RRSs, we selected a subset of hospitals among the 12 included institutions without RRSs. To minimize bias, we excluded the two hospitals with the highest and the lowest CPR rates and three hospitals that had not used electronic medical record (EMR) systems in 2013. Finally, seven hospitals were included in the group without an RRS. To identify the in-hospital cardiopulmonary arrest rate by hospital bed size, we categorized hospitals with more than 32,000 annual adult admissions (the median value among the 14 hospitals) as “high-volume hospitals” and those with less than 32,000 annual adult admissions as “medium-volume hospitals.”

without an RRS. To identify the in-hospital cardiopulmonary arrest rate by hospital bed size, we categorized hospitals with more than 32,000 annual adult admissions (the median value among the 14 hospitals) as “high-volume hospitals” and those with less than 32,000 annual adult admissions as “medium-volume hospitals.” 3) Statistical analysis and ethical considerations The chi-square test for categorical dummy variables was performed to compare cardiopulmonary arrest rates prior to and after implementation of an RRS. We also investigated cardiopulmonary arrest rates in accordance with hospital characteristics of location, type of organization, and bed capacity. Statistical analysis was performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL, USA). Microsoft Excel 2010 (Microsoft, Redmond, WA, USA) was used to collect and summarize the data and generate graphs. A P-value less than 0.05 was regarded as statistically significant. Because we only collected statistical data regarding mortality and did not use personal information, the Institutional Review Board waived deliberation of this study (IRB No. 2016-12-006).

was used to collect and summarize the data and generate graphs. A P-value less than 0.05 was regarded as statistically significant. Because we only collected statistical data regarding mortality and did not use personal information, the Institutional Review Board waived deliberation of this study (IRB No. 2016-12-006). Results 1) Hospital characteristics The characteristics of the 14 hospitals are presented Table 1. Twelve hospitals did not have an RRS, while two did have an RRS. In the group with an RRS, one hospital was a national university-affiliated institution, and the other was private and university-affiliated; both were located in provincial areas. Among the 12 hospitals without RRSs, four were in the capital area, and eight were situated in provincial areas. All hospitals were tertiary medical centers in urban settings. Of the locations without an RRS, four hospitals were national universityaffiliated and eight were private university-affiliated. In terms of the annual number of adult admissions, half of the hospitals without an RRS admitted over 30,000 patients per year, and both hospitals with an RRS admitted 30,000–39,999 patients annually. Three hospitals without an RRS did not have full-time intensivists on staff as of July 2015. In addition, all 14 hospitals offered CPR or basic life support training programs for healthcare providers and implemented departments of quality improvement. Detailed characteristics of the two hospitals with an RRS are presented in Supplementary Table 1.

S did not have full-time intensivists on staff as of July 2015. In addition, all 14 hospitals offered CPR or basic life support training programs for healthcare providers and implemented departments of quality improvement. Detailed characteristics of the two hospitals with an RRS are presented in Supplementary Table 1. 2) Comparison of in-hospital cardiopulmonary arrest rates We compared the cardiopulmonary arrest rates between the hospitals with an RRS and those without it (Figure 2). Hospitals with an RRS showed a statistically significant reduction in CPR rate between 2013 and 2015 (odds ratio [OR], 0.731; 95% confidence interval [CI], 0.577 to 0.927; P = 0.009). We analyzed the monthly CPR rate of the two hospitals with an RRS (Figure 3), which demonstrated a gradual reduction in CPR rate. The incidence of CPR in 2013 and 2015 was similar in hospitals without an RRS (OR, 0.988; 95% CI, 0.868 to 1.124; P = 0.854). Table 2 displays the annual CPR rates and yearly adult admissions statistics in each of the two groups; the number of CPR episodes declined from 161 cases to 119 cases after the implementation of an RRS. 3) Prevalence of in-hospital cardiopulmonary arrest We verified the in-hospital cardiopulmonary arrest rates of hospitals based on affiliation (national vs. private), location (capital vs. provincial), and volume (high-volume vs. medium-volume). We additionally excluded three hospitals that did not use an EMR in 2013 to avoid selection bias. Therefore, all included hospitals utilized an EMR during the period of interest (Table 3).

spitals based on affiliation (national vs. private), location (capital vs. provincial), and volume (high-volume vs. medium-volume). We additionally excluded three hospitals that did not use an EMR in 2013 to avoid selection bias. Therefore, all included hospitals utilized an EMR during the period of interest (Table 3). (1) National university-affiliated hospitals versus private university-affiliated hospitals From 2013 to 2015, the in-hospital cardiopulmonary arrest rate of the five national university-affiliated hospitals decreased from 2.29 to 1.92, and that of the nine private university-affiliated hospitals increased from 2.04 to 2.40. National university-affiliated hospitals had a significantly lower OR than private hospitals in 2015 (1.92 vs. 2.40, respectively; OR, 0.800; 95% CI, 0.702 to 0.912; P = 0.001) (Figure 4A). (2) Capital area versus provincial area Four hospitals represented the capital area and demonstrated a significant decrease in cardiopulmonary arrest rate (1.93 to 1.53) over 2 years. However, in the remaining 10 hospitals, the cardiopulmonary arrest rate increased from 2.19 to 2.51. In 2015, the capital area showed a significantly lower cardiopulmonary arrest rate compared to hospitals in provincial areas (1.53 vs. 2.51, respectively; OR, 0.609; 95% CI, 0.523 to 0.709; P < 0.001) (Figure 4B).

ears. However, in the remaining 10 hospitals, the cardiopulmonary arrest rate increased from 2.19 to 2.51. In 2015, the capital area showed a significantly lower cardiopulmonary arrest rate compared to hospitals in provincial areas (1.53 vs. 2.51, respectively; OR, 0.609; 95% CI, 0.523 to 0.709; P < 0.001) (Figure 4B). (3) High-volume hospitals versus medium-volume hospitals High-volume hospitals showed fewer in-hospital cardiopulmonary arrests compared with medium-volume hospitals in 2013 (1.76 vs. 2.63, respectively; OR, 0.667; 95% CI, 0.577 to 0.772; P < 0.001) and in 2015 (1.55 vs. 3.20, respectively; OR, 0.485; 95% CI, 0.428 to 0.550; P < 0.001) (Figure 4C). Discussion Prior studies have suggested that in-hospital mortality is not an appropriate indicator of hospital service or quality given that patient populations vary widely among hospitals, and a comparison of mortality without any correction for selection bias could skew such analyses [10-12]. Nonetheless, the number of in-hospital cardiopulmonary arrests can be reduced with proper interventions; this is a hot topic of interest for administrators and healthcare policy authorities. Various ways to prevent cardiopulmonary arrest have been introduced over the past few decades. They include the introduction of an EMR, use of Early Warning Score systems, creation of intermediate care units, and implementation of an RRS [1,13,14].

topic of interest for administrators and healthcare policy authorities. Various ways to prevent cardiopulmonary arrest have been introduced over the past few decades. They include the introduction of an EMR, use of Early Warning Score systems, creation of intermediate care units, and implementation of an RRS [1,13,14]. RRSs have been widely adopted around the world over the past two decades and effectively reduce in-hospital cardiopulmonary arrests [13-18]. RRSs can diminish inhospital cardiopulmonary arrests and improve patient safety by initiating earlier treatment attempts; impacting do-not-attempt-resuscitation decisions; monitoring atrisk patients; increasing levels of care; correcting wrong diagnoses; and requiring specialist consultations [19]. Despite these benefits, there have been barriers against the successful implementation of RRSs. Practically, there is a lack of specialists for RRS implementation and scarce financial support by hospitals. Moreover, the absence of government policy about RRSs is an important issue to be addressed. It is worth mentioning that some multi-center trials on RRSs have been conducted in other countries, and their results deserve attention despite being obtained in different healthcare systems and environments [7,16,17].

pitals. Moreover, the absence of government policy about RRSs is an important issue to be addressed. It is worth mentioning that some multi-center trials on RRSs have been conducted in other countries, and their results deserve attention despite being obtained in different healthcare systems and environments [7,16,17]. In this survey, the cardiopulmonary arrest rate of the 14 participating tertiary hospitals ranged from 1.02 to 7.1/1,000 adult admissions for the study period. Five hospitals without an RRS showed a low incidence of cardiopulmonary arrest (<2.0/1,000 admissions), whereas three hospitals without an RRS reported a higher incidence of cardiopulmonary arrest (>3.0/1,000 admissions). These variations in cardiopulmonary arrest rates might result from differences in hidden curricula and workforces, but this is beyond the scope of this study. Interestingly, this pilot analysis demonstrated that two hospitals with an operating RRS achieved a statistically significant reduction of in-hospital cardiopulmonary arrests between 2013 and 2015, although hospital A experienced an abrupt rise in the CPR rate in July 2015 due to an outbreak of Middle East respiratory syndrome.

ngly, this pilot analysis demonstrated that two hospitals with an operating RRS achieved a statistically significant reduction of in-hospital cardiopulmonary arrests between 2013 and 2015, although hospital A experienced an abrupt rise in the CPR rate in July 2015 due to an outbreak of Middle East respiratory syndrome. Regardless of the presence of RRSs, we found different cardiopulmonary arrest rates based on hospital characteristics. National university-affiliated hospitals showed relatively fewer cardiopulmonary arrests than private hospitals in 2015. High-volume hospitals also demonstrated a lower cardiopulmonary arrest rate than medium-volume hospitals. This tendency was presumed to be due to the strong medical and human resources available in national university-affiliated hospitals and high-volume hospitals. While all high-volume hospitals had dedicated intensivists, only 57% (four of seven hospitals) of medium-volume hospitals had them. None of the medium-volume hospitals had an RRS. Furthermore, three of five national university-affiliated hospitals (60%) were high-volume hospitals, whereas four of nine private hospitals (44%) were high-volume hospitals, which could explain the differences between cardiopulmonary arrest at national university-affiliated hospitals and private hospitals (Supplementary Table 2). However, these results should be interpreted with caution because this study was not a comprehensive investigation, but instead only compared the structural characteristics of a few hospitals. All included hospitals were tertiary hospitals and were accredited by Ministry of Health and Welfare. Accordingly, it can be assumed that the levels of care were above standard.

caution because this study was not a comprehensive investigation, but instead only compared the structural characteristics of a few hospitals. All included hospitals were tertiary hospitals and were accredited by Ministry of Health and Welfare. Accordingly, it can be assumed that the levels of care were above standard. In 2015, hospitals in the capital area demonstrated a lower cardiopulmonary arrest rate than those in provincial areas. Two of these were high-volume hospitals, but that difference does not fully establish an inference related with patient volume. Although we could not entirely rule out a selection bias in our sampling methods, this result might suggest that the discrepancy in cardiopulmonary arrest rate is due to a local gap in medical services as well as the severity of patient illnesses and comorbidities.

ly establish an inference related with patient volume. Although we could not entirely rule out a selection bias in our sampling methods, this result might suggest that the discrepancy in cardiopulmonary arrest rate is due to a local gap in medical services as well as the severity of patient illnesses and comorbidities. This is the first multicenter survey on the impacts of RRSs in Korea. To avoid selection bias, we made efforts to study hospitals with similar characteristics. First, we corrected combined probability sampling with nonprobability sampling. Second, we excluded data from hospitals that recorded the highest and lowest cardiopulmonary arrest rates and from the period when EMRs were not used. However, there are some limitations to this investigation. This study was a retrospective analysis that only included data from 14 of the 43 tertiary hospitals, which could have introduced a selection bias. Regarding the data, we collected admission statistics from each hospital and did not review individual medical records. Each hospital had a different system of reporting CPR rates, including standards to fulfill CPR reports and responsible parties. Hence, it is possible that some data are not apparent from the accessible CPR cases. Nevertheless, the aim of this research was to determine the incidence of cardiopulmonary arrest, not to analyze its causes. It was also not intended to be a comparative analysis that reflected various variables and characteristics of each hospital. Finally, this study was designed as a before and after study, and it is challenging to conduct randomized controlled trials in this manner.

e of cardiopulmonary arrest, not to analyze its causes. It was also not intended to be a comparative analysis that reflected various variables and characteristics of each hospital. Finally, this study was designed as a before and after study, and it is challenging to conduct randomized controlled trials in this manner. The implementation of an RRS in two tertiary hospitals reduced the incidence of cardiopulmonary arrest over 3 years. Although we cannot conclude that RRSs are the only method that can be used to reduce cardiopulmonary arrest, they might be a practical option to use to achieve this goal. The discrepancies in cardiopulmonary arrest rate submitted for this survey suggest that a nationwide survey on cardiopulmonary arrest and the effect of RRSs is necessary. Promoting patient safety and healthcare quality is within the public domain. Therefore, hospital administrators should take an interest in encouraging and supporting the implementation of RRSs. No potential conflict of interest relevant to this article was reported. We are grateful to the rapid response teams at Chungnam National University Hospital and Ulsan University Hospital for their key role in this work. Author contributions: substantial contributions to conception and design, acquisition of data or analysis and interpretation of data: all authors; drafting the article or revising it critically for important intellectual content: YHP, JJA, BJK, GHH, SWS, JHK, JYM; final approval of the version to be published: all authors.

Author contributions: substantial contributions to conception and design, acquisition of data or analysis and interpretation of data: all authors; drafting the article or revising it critically for important intellectual content: YHP, JJA, BJK, GHH, SWS, JHK, JYM; final approval of the version to be published: all authors. Supplementary Materials The online-only Supplement data are available with this article online: https://doi.org/10.4266/kjccm.2017.00024. Figure 1. Selection process of included tertiary hospitals. RRS: rapid response system. Figure 2. A comparison of cardiopulmonary arrest rates between hospitals with RRS and without RRS. CPR: cardiopulmonary resuscitation; RRS: rapid response system. Figure 3. The monthly cardiopulmonary arrest rate of each hospital with an RRS. RRS: rapid response system. Figure 4. A comparison of in-hospital cardiac arrest rate according to hospital characteristics. (A) National university-affiliated hospitals versus private university-affiliated hospitals, (B) capital area versus provincial area, (C) high-volume hospitals versus medium-volume hospitals. Table 1. Characteristics of included hospitals

Figure 4. A comparison of in-hospital cardiac arrest rate according to hospital characteristics. (A) National university-affiliated hospitals versus private university-affiliated hospitals, (B) capital area versus provincial area, (C) high-volume hospitals versus medium-volume hospitals. Table 1. Characteristics of included hospitals Variable Hospital without RRS (n = 12) Hospital with RRS (n = 2) Region Capital area 4 - Provincial area 8 2 Type Tertiary hospital 12 2 General hospital - - Operating National university-affiliated 4 1 Private university-affiliated 8 1 No. of admission (adult) 20,000–29,999/yr 6 - 30,000–39,999/yr 3 2 40,000–49,999/yr 3 - Capacity (bed) 1,000–1,300 2 1 700–999 10 1 Full-time intensivista Yes 9 2 No 3 - Presence of training program for CPR or BLS Yes 12 2 No 0 0 Presence of department of quality improvement Yes 12 2 No 0 0 RRS: rapid response system; CPR: cardiopulmonary resuscitation; BLS: basic life support. a On July 1, 2015. Table 2. Annual CPR number and annual adult admission number of two groups Variable 2 Hospitals with RRS 7 Hospitals without RRS No. of CPR In 2013 161 454 In 2015 119 471 No. of admissions (adult) In 2013 67,770 224,105 In 2015 68,438 235,319 CPR rate (per 1,000 admission) In 2013 2.38 2.03 In 2015 1.74 2.00 CPR: cardiopulmonary resuscitation; RRS: rapid response system. Table 3. Categorical data for in-hospital cardiopulmonary arrest rate at all 14 hospitals

Variable 2 Hospitals with RRS 7 Hospitals without RRS No. of CPR In 2013 161 454 In 2015 119 471 No. of admissions (adult) In 2013 67,770 224,105 In 2015 68,438 235,319 CPR rate (per 1,000 admission) In 2013 2.38 2.03 In 2015 1.74 2.00 CPR: cardiopulmonary resuscitation; RRS: rapid response system. Table 3. Categorical data for in-hospital cardiopulmonary arrest rate at all 14 hospitals 2013 Cardiopulmonary arrest (n = 11) OR (95% CI) 2015 Cardiopulmonary arrest (n = 14) OR (95% CI) Ownership 1.124 (0.957–1.318) 0.800 (0.702–0.912)* National 2.29 1.92 Private 2.04 2.40 Location 0.882 (0.757–1.028) 0.609 (0.523–0.709)* Capital area 1.93 1.53 Provincial area 2.19 2.51 Annual admission number 0.667 (0.577–0.772)* 0.485 (0.428–0.550)* >32,000 1.76 1.55 <32,000 2.63 3.20 OR: odds ratio; CI: confidence interval. * P < 0.05.

Dear Editor: Carotid dissection is a rare lesion after head injury and is often known to occur in the case of direct neck trauma. Although most of the carotid artery dissection occurs spontaneously, approximately 4% is associated with severe trauma [1]. For example, major blunt trauma resulting from high-speed motor vehicle accidents [2]. Clinical presentation encompasses a wide range of symptoms, often leading to a delay in diagnosis. Asymptomatic carotid artery injury may not be easily detected during clinical evaluation of head and neck trauma [3,4]. Case A 78-year-old woman was admitted after being hanging while experiencing a cart wearing a scarf 1 day ago. She had a history of hypercholesterolemia. The neck computed tomography (CT) scan showed bilateral common carotid artery dissection (Figure 1A and B), with patent blood flow and hematoma with swelling in the pretracheal area (Figure 1C). However, the true and false lumen could not be clearly distinguished and the true lumen was not compressed (Figure 1D). There was no problem with perfusion through the bilateral internal carotid artery. The patient was hemodynamically stable and without neurological deficits. We planned follow-up CT after 3 days and the patient will be transfer to her home region if there is no change in the lesion.

lumen was not compressed (Figure 1D). There was no problem with perfusion through the bilateral internal carotid artery. The patient was hemodynamically stable and without neurological deficits. We planned follow-up CT after 3 days and the patient will be transfer to her home region if there is no change in the lesion. Three days later, the patient complained of sudden severe dizziness. Patients were drowsy on the day of imaging follow-up with nausea and poor oral intake, and the Glasgow Coma Scale decreased from 4-5-6 to 3-4-6. Therefore, emergency transfemoral cerebral angiography and diffusion magnetic resonance imaging were performed. The magnetic resonance imaging showed multiple tiny high signal intensities with acute lacunar infarct in both cerebral white matter (Figure 2). On the transfemoral cerebral angiography, the dissection was observed in distal parts of both common carotid arteries and flow flap. As a result, spontaneous healing could not be expected, and endovascular stent insertion was performed in both common carotid arteries (Figure 3). We inserted 10 mm × 60 mm sized self-expanding Nitinol stents on each side without embolic protection device or ballooning.

ts of both common carotid arteries and flow flap. As a result, spontaneous healing could not be expected, and endovascular stent insertion was performed in both common carotid arteries (Figure 3). We inserted 10 mm × 60 mm sized self-expanding Nitinol stents on each side without embolic protection device or ballooning. After stent insertion, we started antithrombotic therapy (clopidogrel 75 mg and aspirin 100 mg per day as oral medications). The patient was discharged about 2 weeks later, complaining only of pain in the anterior neck without any specific neurologic deficiency. The study was performed according to the Helsinki Declaration and approved by the institutional review board of Jeju National University Hospital (No. 2016-12-009). The incidence of carotid artery dissection induced by blunt trauma ranges from less than 1% to 3% [2]. The actual incidence can be even higher and remain undiagnosed. Risk factors for traumatic carotid artery dissection include intense physical activity, blunt injury, and penetrating neck trauma. In general, about 10% of patients show immediate symptoms. In the first 24 hours after dissection, 55% had symptoms, and 35% had no symptoms for more than 24 hours after injury [5]. Patients with immediate symptoms usually have neurological deficits. Headache, including neck and facial pain, pulsatile tinnitus, decreased taste sensation, focal weakness, and migraine-like symptoms.

he first 24 hours after dissection, 55% had symptoms, and 35% had no symptoms for more than 24 hours after injury [5]. Patients with immediate symptoms usually have neurological deficits. Headache, including neck and facial pain, pulsatile tinnitus, decreased taste sensation, focal weakness, and migraine-like symptoms. Carotid artery dissection begins to tear in one of the carotid arteries, and the blood enters the arterial wall and splits the layer. As a result, intramural hematoma or aneurysmal dillatation is created and this process can either be the source of miroemboli. This injury involves an initial intimal tear and exposes the thrombogenic subendothelial collagen, initiating platelet aggregation with subsequent thrombus formation that can embolize the artery [6]. Therefore, some studies have recommended an initial antithrombotic regimen [7,8].

an either be the source of miroemboli. This injury involves an initial intimal tear and exposes the thrombogenic subendothelial collagen, initiating platelet aggregation with subsequent thrombus formation that can embolize the artery [6]. Therefore, some studies have recommended an initial antithrombotic regimen [7,8]. In the present case, the patient was unable to undergo antithrombotic therapy at the initial stage due to hematoma-induced neck swelling. On angiography, a thrombus was not present but a visible intimal flap was observed (Figure 3A and C), so the dissection was classified as grade II [9]. The management of low-grade (grade I and II) blunt carotid artery injuries remains controversial. However, antiplatelet agents or anticoagulants are used as first-line treatments, and endovascular stenting is generally reserved for symptomatic or higher-grade blunt injuries [6,10-13]. Low-grade blunt cerebrovascular injuries (BCVI) carry the low risk of cerebral infarction. Griessenauer et al. [14] searched the outcomes in 112 patients with BCVI. They concluded that most ischemic strokes occur before prescreening using CT angiography, and antiplatelet therapy. This indicates that subsequent imaging may not help prevent most ischemic strokes. The progression of injury was not changed by specific treatment or absence of treatment [13,15]. However, according to available literature, it is difficult to determine the actual rate of BCVI-related ischemic strokes. The incidence of BCVI-related ischemic stroke has been reported to be as low as 0.05% and as high as 50% [14,16-18]. At that time of admission, the risk of active bleeding or cerebral hemorrhage was low, but if the hematoma increased, dyspnea would occur and the patient’s management might be difficult. Thus, we did not proceed with the initial antithrombotic therapy and proceeded after the endovascular procedure.

h as 50% [14,16-18]. At that time of admission, the risk of active bleeding or cerebral hemorrhage was low, but if the hematoma increased, dyspnea would occur and the patient’s management might be difficult. Thus, we did not proceed with the initial antithrombotic therapy and proceeded after the endovascular procedure. Fortunately, this patient had a temporary neurological symptom that disappeared after stent insertion. There were no new symptoms during the 2 weeks of hospitalization. In addition, the hemorrhage in the neck did not progress after antithrombotic therapy. Therefore, we could manage the patient with low-grade carotid artery dissection with endovascular stent insertion without any complication, despite the delayed initial antithrombotic therapy. No potential conflict of interest relevant to this article was reported. We thank the intensive care unit members and primary care physicians for help and management for this patient. Figure 1. Neck computed tomography scan showing dissections of both common carotid arteries, mid portion. (A) The 2.6-cm segmental intimal flap in the left common carotid artery. (B) A short segmental intimal flap in the right common carotid artery. (C) Diffuse hematoma around the thyroid gland in the anterior neck. (D) Coronal view showing both common carotid artery dissections. Figure 2. Brain diffusion magnetic resonance image showing a tiny diffusion-restricted lesion in the left frontal white matter. Multiple tiny high signal intensities in both cerebral white matter.

Figure 1. Neck computed tomography scan showing dissections of both common carotid arteries, mid portion. (A) The 2.6-cm segmental intimal flap in the left common carotid artery. (B) A short segmental intimal flap in the right common carotid artery. (C) Diffuse hematoma around the thyroid gland in the anterior neck. (D) Coronal view showing both common carotid artery dissections. Figure 2. Brain diffusion magnetic resonance image showing a tiny diffusion-restricted lesion in the left frontal white matter. Multiple tiny high signal intensities in both cerebral white matter. Figure 3. Transfemoral cerebral angiography showing dissection of both distal common carotid arteries. (A) A movable flap was observed according to flow in the left common carotid artery without thrombus. (B) Flow patent after stent insertion in the left common carotid artery. (C) A movable flap was observed according to flow in the right common carotid artery without thrombus. (D) Flow patent after stent insertion in the right common carotid artery.

Introduction Sepsis is a life-threatening organ dysfunction resulting from a dysregulated host response to infection and is the leading cause of morbidity and mortality worldwide [1,2]. Sepsis was defined by a consensus conference in 1991 [3]; however, the definition was revised in 2016 (sepsis-3) [2]. In sepsis-3, the quick Sepsis-Related Organ Failure Assessment (qSOFA) score was introduced as a means of screening for sepsis at the bedside based on a patient’s respiratory rate, blood pressure, and level of consciousness [2]. In addition, the recommendations suggest that patients with a qSOFA score ≥2 suspected to have an infection should be monitored closely. Although several studies have suggested that compliance with the Surviving Sepsis Campaign bundles can benefit survival [4-7], compliance with resuscitation and management bundles is generally poor in many Asian intensive care units (ICUs) (including Korea) [8,9]. Moreover, there have been no multicenter studies regarding the current status of compliance with management recommendations at the national level because critical care resources and facilities at university hospitals in Korea are limited compared with Western countries [10]. Therefore, it is questionable whether qSOFA scores can be applied successfully in Korea. In addition, no largescale multicenter studies have reported the prognostic utility of this score. Also, the usefulness of the qSOFA score in Korea is unknown.

Although several studies have suggested that compliance with the Surviving Sepsis Campaign bundles can benefit survival [4-7], compliance with resuscitation and management bundles is generally poor in many Asian intensive care units (ICUs) (including Korea) [8,9]. Moreover, there have been no multicenter studies regarding the current status of compliance with management recommendations at the national level because critical care resources and facilities at university hospitals in Korea are limited compared with Western countries [10]. Therefore, it is questionable whether qSOFA scores can be applied successfully in Korea. In addition, no largescale multicenter studies have reported the prognostic utility of this score. Also, the usefulness of the qSOFA score in Korea is unknown. We hypothesized the qSOFA score would be useful in Korean patients. The present study investigated the clinical application and usefulness of the qSOFA score at ICU admission for predicting 28-day mortality in patients with a microbiologically diagnosed infection. In addition, we compared this score with other conventional early warning scores [3,11,12].

d be useful in Korean patients. The present study investigated the clinical application and usefulness of the qSOFA score at ICU admission for predicting 28-day mortality in patients with a microbiologically diagnosed infection. In addition, we compared this score with other conventional early warning scores [3,11,12]. Materials and Methods 1) Study design and subjects This retrospective study was conducted at a university-affiliated tertiary care hospital. This hospital has six functionally separate ICUs with 85 beds (medical, 12 beds; surgical, 10 beds; cardio-stroke, 14 beds; neurosurgical, 13 beds; emergency, 20 beds; and trauma, 16 beds) with full cardiovascular and close airway monitoring. All patients were managed according to therapeutic recommendations based on Surviving Sepsis guidelines and a lung-protective ventilator strategy [13,14].

ds; surgical, 10 beds; cardio-stroke, 14 beds; neurosurgical, 13 beds; emergency, 20 beds; and trauma, 16 beds) with full cardiovascular and close airway monitoring. All patients were managed according to therapeutic recommendations based on Surviving Sepsis guidelines and a lung-protective ventilator strategy [13,14]. We included patients who had various infectious causes with positive blood culture tests at ICU admission; all blood culture results were obtained within 3 days after ICU admission. The inclusion period was from March 2011 to February 2016. The exclusion criteria were patients younger than 18 years and those whose mental state could not be assessed. Also, patients who could not know 28-day mortality after ICU admission (example, transferred to other hospitals) were excluded. For all positive blood cultures, organism identification was performed by conventional and automated biochemical methods (VITEK 2; BioMérieux, Marcy l’Etoile, France) from March 2011 to February 2013, and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Daltonic, Bremen, Germany) from March 2013 to February 2016. The medical records and laboratory and radiological findings of all patients included in the study were reviewed. All investigators confirmed that the study objectives and procedures were complete, and they had full access to all data. The investigators completed a case report form for each patient; data were collected from September to December in 2016. This study was conducted with the approval of Institutional Review Board of Pusan National University Hospital (IRB No. 1612-003-049). This study had no impact on patient treatment.

cess to all data. The investigators completed a case report form for each patient; data were collected from September to December in 2016. This study was conducted with the approval of Institutional Review Board of Pusan National University Hospital (IRB No. 1612-003-049). This study had no impact on patient treatment. 2) Data collection The following data were gathered from the medical records of each patient: age, sex, comorbidities before ICU admission, ICU admission route, and length of stay (LOS; ICU and hospital). The severity of illness was measured by the Acute Physiology and Chronic Health Evaluation (APACHE) II score, and accompanying organ failure was measured by the Sequential Organ Failure Assessment (SOFA) score [15,16]. APACHE II and SOFA scores were calculated using data from the first 24 hours of ICU admission. The qSOFA was calculated at the time of ICU admission, which was defined as a systolic blood pressure ≤100 mmHg, respiratory rate ≥22 breaths per minute, and altered mental status (defined as a Glasgow Coma Scale score ≤13) [2]. To compare the prognostic utility with qSOFA, we also calculated the Modified Early Warning score (MEWS) at ICU admission and systemic inflammatory response syndrome (SIRS) criteria within the first 24 hours of ICU admission; these data were based on previously published definitions [3,11,12].

oma Scale score ≤13) [2]. To compare the prognostic utility with qSOFA, we also calculated the Modified Early Warning score (MEWS) at ICU admission and systemic inflammatory response syndrome (SIRS) criteria within the first 24 hours of ICU admission; these data were based on previously published definitions [3,11,12]. We also evaluated primary sources of infection at ICU admission, microbiological data (Gram staining, organism identification, and susceptibility testing), and the requirement for hemodialysis (defined as the use of any form of renal replacement therapy), neuromuscular blocking agents, vasopressors, and ventilator care within 3 days after ICU admission. In addition, the blood platelet count and arterial lactic acid level were determined during the first 3 days after ICU admission. Survivors were defined as patients that survived for 28 days after ICU admission.

herapy), neuromuscular blocking agents, vasopressors, and ventilator care within 3 days after ICU admission. In addition, the blood platelet count and arterial lactic acid level were determined during the first 3 days after ICU admission. Survivors were defined as patients that survived for 28 days after ICU admission. 3) Statistical analysis Continuous variables are expressed as medians (interquartile range [IQR]) and categorical variables are expressed as numbers (percentages). Student t-test and the Mann-Whitney U-test were applied to compare continuous variables. The chi-square and Fisher exact tests (for small numbers) were used to compare categorical variables. To estimate predictive capabilities of the qSOFA score and other scores for our cohort, the receiver operating characteristic curves were used to determine cutoff value. Pearson correlation coefficients between the qSOFA score and MEWS and SIRS were calculated. Logistic regression analyses were performed to evaluate the qSOFA score as an independent prognostic factor in 28-day mortality. All statistical analyses were performed using SPSS version 19.0 (IBM Corp., Armonk, NY, USA). A two-tailed P-value <0.05 was considered to indicate significant difference.

RS were calculated. Logistic regression analyses were performed to evaluate the qSOFA score as an independent prognostic factor in 28-day mortality. All statistical analyses were performed using SPSS version 19.0 (IBM Corp., Armonk, NY, USA). A two-tailed P-value <0.05 was considered to indicate significant difference. Results 1) Baseline characteristics During the study period, we identified 236 patients with infectious causes that had positive blood cultures within 3 days after ICU admission. In the total patient population, 151 (64.0%) were admitted to the ICU via the emergency department (ED) and 144 patients (61.0%) received ventilator care during their ICU stay. The median ICU LOS and hospital LOS were 10 days (IQR, 5 to 19 days) and 25 days (IQR, 14 to 53 days), respectively. Diabetes mellitus was the most common underlying disease, and pneumonia was the most common source of bacteremia (Table 1). Gram-positive bacteria were the most commonly identified organisms (Table 1). Of total enrolled patients, 49 patients (20.8%) received surgical drainage aside from antibiotics. The clinical characteristics of all patients enrolled in this study and comparisons between survivors and non-survivors are presented in Table 1.

am-positive bacteria were the most commonly identified organisms (Table 1). Of total enrolled patients, 49 patients (20.8%) received surgical drainage aside from antibiotics. The clinical characteristics of all patients enrolled in this study and comparisons between survivors and non-survivors are presented in Table 1. 2) qSOFA score and patient outcomes Figure 1 shows the number of patients for each qSOFA level and the corresponding mortality rates. Of the patients with a qSOFA score ≥2 (n = 127), 38 (29.9%) had all three criteria, followed by blood pressure and respiratory rate criteria (n = 35, 27.6%), mental status and blood pressure (n = 28, 22.0%), and mental status and respiratory rate criteria (n = 26, 20.5%). Patients with a qSOFA score ≥2 had significantly higher APACHE II and SOFA scores compared to those with a qSOFA score <2. In addition, these patients had significantly higher rates of septic shock, thrombocytopenia, and hyperlactatemia, and significantly greater requirements for ventilator care, neuromuscular blocking agents, vasopressors, and hemodialysis during the first 72 hours of ICU admission (Table 2). Further analysis indicated that patients with a qSOFA score ≥2 had significantly higher 28-day mortality rates than those with a qSOFA score <2 (Table 2). A univariate logistic regression analysis showed that a qSOFA score ≥2 was associated with 28-day mortality in our cohort (odds ratio, 2.722; 95% confidence interval, 1.582 to 4.683; P < 0.001).

at patients with a qSOFA score ≥2 had significantly higher 28-day mortality rates than those with a qSOFA score <2 (Table 2). A univariate logistic regression analysis showed that a qSOFA score ≥2 was associated with 28-day mortality in our cohort (odds ratio, 2.722; 95% confidence interval, 1.582 to 4.683; P < 0.001). 3) Comparison of qSOFA score with MEWS, SIRS, and SOFA When we compared two conventional early warning scores (MEWS and SIRS), we found correlations between qSOFA score and MEWS (γ = 0.401, P < 0.001) and between qSOFA score and SIRS criteria (γ = 0.271, P < 0.001). Also, positive correlation was found between qSOFA score and SOFA score (γ = 0.465, P < 0.001).

RS, and SOFA When we compared two conventional early warning scores (MEWS and SIRS), we found correlations between qSOFA score and MEWS (γ = 0.401, P < 0.001) and between qSOFA score and SIRS criteria (γ = 0.271, P < 0.001). Also, positive correlation was found between qSOFA score and SOFA score (γ = 0.465, P < 0.001). Further analysis using common thresholds for each conventional early warning score (MEWS ≥5, qSOFA ≥2, and SIRS criteria ≥2) according to published data was presented in Table 3 [3,11,17]. Patients with a MEWS ≥5 had significantly higher rates of septic shock, thrombocytopenia, and hyperlactatemia, and significantly greater requirements for ventilator care, neuromuscular blocking agents, vasopressors, and hemodialysis during the initial 72 hours after ICU admission than those of MEWS <5 (Table 3). In addition, they had higher 28-day mortality rates. However, patients with ≥2 SIRS criteria showed no significant differences compared to those with <2 SIRS criteria (Table 3). Also, we found the cutoff value of SOFA was 7, which were determined by receiver operating characteristic curves. When we compared between patients with a SOFA ≥7 and <7, patients with a SOFA ≥7 had same results as shown in patients with a MEWS ≥5 (Table 3).

ifferences compared to those with <2 SIRS criteria (Table 3). Also, we found the cutoff value of SOFA was 7, which were determined by receiver operating characteristic curves. When we compared between patients with a SOFA ≥7 and <7, patients with a SOFA ≥7 had same results as shown in patients with a MEWS ≥5 (Table 3). Discussion In the present study, we enrolled patients with bacteremia on ICU admission and evaluated the clinical utility of qSOFA scores at the time of ICU admission. In the present study, qSOFA score had positive correlation with SOFA score. Also, a qSOFA score ≥2 at ICU admission was associated with greater severity and higher medical resource use in the initial 72 hours after ICU admission. In addition, a qSOFA score ≥2 was a significant prognostic indicator for 28-day mortality. Although critical care resources are typically limited and there are distinct cultural differences compared to those in Western countries [9,10], our results suggested that a qSOFA score ≥2 at ICU admission would be a useful screening tool for predicting disease severity and mortality in patients with bacteremia.

lity. Although critical care resources are typically limited and there are distinct cultural differences compared to those in Western countries [9,10], our results suggested that a qSOFA score ≥2 at ICU admission would be a useful screening tool for predicting disease severity and mortality in patients with bacteremia. After introducing the qSOFA score in sepsis-3 as a screening tool for organ dysfunction [2], comparisons of qSOFA score with some conventional early warning scores were reported [17-20]. In our study, the cutoff levels of MEWS and SIRS were used according to previous reported data [3,11,17]. Our results showed a MEWS ≥5 was associated with greater severity and higher medical resource use within 72 hours after ICU admission, and 28-day mortality after ICU admission, consistent with the observations associated with a qSOFA score ≥2. However, we found no prognostic utility of ≥2 SIRS criteria because 96.6% of the total patient population had ≥2 SIRS criteria (Table 3). Our findings suggest that qSOFA scores may be more useful than SIRS criteria as a prognostic indicator, consistent with a previous report [21]. In comparison with previous studies [17-20], however, our patients had bacteremia with a documented infectious focus at ICU admission, and they had a higher mortality rate. Therefore, additional large-scale studies including patients with non-bacteremia are required to compare qSOFA with other early warning scores as early screening tools.

evious studies [17-20], however, our patients had bacteremia with a documented infectious focus at ICU admission, and they had a higher mortality rate. Therefore, additional large-scale studies including patients with non-bacteremia are required to compare qSOFA with other early warning scores as early screening tools. In the present study, we found the survival rate was different according to admission route. In patients admitted to ICU via ED, there was no significant difference in the 28-day mortality rate between patients with qSOFA score ≥2 and <2 (37.5% vs. 29.1%, respectively; P = 0.274). In patients admitted from general wards, however, patients with qSOFA score ≥2 had significantly higher 28-day mortality rate than those with qSOFA score <2 (70.9% vs. 26.7%, respectively; P < 0.001). To find out these differences, we further evaluated where the patients were before being transferred to the emergency room of our hospital (i.e., home, heath care facility, or another teaching hospital), however, we could not investigate accurately because of the shortage of medical records. Therefore, further investigation is needed to evaluate the prognostic utility of qSOFA score for patients presenting to the ED.

to the emergency room of our hospital (i.e., home, heath care facility, or another teaching hospital), however, we could not investigate accurately because of the shortage of medical records. Therefore, further investigation is needed to evaluate the prognostic utility of qSOFA score for patients presenting to the ED. Our study had several limitations. First, although qSOFA score was developed for patients with suspected infection presenting to the ED, in our study, we could not find the usefulness of qSOFA score for these patients due to the shortage of medical records. To assess the usefulness of this score, therefore, we enrolled patients who had documented infections with bacteremia admitted to ICU. Second, this study was conducted retrospectively; this may have resulted in information bias. Also, our enrolled patient populations were heterogeneous from six ICUs, which may be a bias. Third, our data represent the experience of a single center, so the results may not be representative of the general situation in Korea. Fourth, we expected that the qSOFA ≥2 score was associated with poor prognosis according to documented bacteria or sources of infection; however, we were unable to identify statistical significances in subgroup analysis due to the small sample size.

e results may not be representative of the general situation in Korea. Fourth, we expected that the qSOFA ≥2 score was associated with poor prognosis according to documented bacteria or sources of infection; however, we were unable to identify statistical significances in subgroup analysis due to the small sample size. In conclusion, we investigated the prognostic utility of the qSOFA score at ICU admission for patients with bacteremia. Our results show that a qSOFA score ≥2 at admission could be useful as a screening tool for predicting clinical severity and medical resource use within 72 hours after admission, and for predicting the 28-day mortality rate. In addition, a comparison of qSOFA score with MEWS and SIRS criteria suggested that qSOFA scores are more useful than SIRS criteria. Prospective and large-scale studies are required to determine the prognostic utility of qSOFA scores in Korean ICUs. No potential conflict of interest relevant to this article was reported. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science and ICT) (No. 2016R1C1B1008529). Figure 1. The number of patients for each qSOFA level (left Y axis) and the corresponding mortality (right Y axis). qSOFA: quick Sepsis-Related Organ Failure Assessment. Table 1. Comparison of baseline characteristics between survivors and non-survivors

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science and ICT) (No. 2016R1C1B1008529). Figure 1. The number of patients for each qSOFA level (left Y axis) and the corresponding mortality (right Y axis). qSOFA: quick Sepsis-Related Organ Failure Assessment. Table 1. Comparison of baseline characteristics between survivors and non-survivors Characteristic Total (n = 236) Survivor (n = 139) Non-survivor (n = 97) P-value Age (yr) 69 (57–76) 70 (58–77) 67 (55–75) 0.192 Male sex 144 (61.0) 82 (59.0) 62 (63.9) 0.498 ICU type Medical ICU 86 (36.4) 46 (33.1) 40 (41.2) 0.218 Surgical ICU 31 (31.1) 14 (10.1) 17 (17.5) 0.117 Cardio-stroke ICU 32 (13.6) 18 (12.9) 14 (14.4) 0.847 Emergency ICU 57 (24.2) 13 (28.1) 18 (18.6) 0.122 Neurosurgical ICU 27 (11.4) 19 (13.7) 8 (8.2) 0.219 Trauma ICU 3 (1.3) 3 (1.3) 0 0.271 APACHE II score on ICU admission day 23 (17–29) 20 (15–26) 26 (22–32) <0.001 SOFA score on ICU admission day 7 (4–9) 5 (3–8) 9 (6–11) <0.001 qSOFA score at ICU admission time 2 (0–3) 1 (0–3) 2 (0–3) <0.001 Comorbidities, overlapped Diabetes mellitus 69 (29.2) 46 (33.1) 23 (23.7) 0.146 Hemato-oncological disease 55 (23.3) 21 (15.1) 34 (35.1) 0.001 Cerebrovascular disease 34 (14.4) 20 (14.4) 14 (14.4) >0.999 Heart failure 34 (14.4) 21 (15.1) 13 (13.4) 0.851 Chronic kidney disease 25 (10.6) 14 (10.1) 11 (11.3) 0.831 Biliary disease 21 (8.9) 16 (11.5) 5 (5.2) 0.107 Chronic liver disease 19 (8.1) 9 (6.5) 10 (10.3) 0.335 Neuromuscular disease 13 (5.5) 11 (7.9) 2 (2.1) 0.079 Chronic lung diseasea 12 (8.9) 12 (8.6) 9 (9.3) >0.999 Source of infection Pneumonia 91 (38.6) 49 (35.3) 42 (43.3) 0.224 Intra-abdominal 61 (25.8) 37 (26.6) 24 (24.7) 0.765 Urinary tract 25 (10.6) 19 (13.7) 6 (6.2) 0.085 Musculoskeletal 41 (17.4) 28 (20.1) 13 (13.4) 0.222 Catheter-related 22 (9.3) 14 (10.1) 8 (8.2) 0.821 Neutropenia 13 (5.5) 2 (1.4) 11 (11.3) 0.002 Infectious endocarditis 11 (4.7) 5 (3.6) 6 (6.2) 0.366 Organism Gram-positive bacteremia 138 (58.5) 92 (66.2) 46 (47.4) 0.005 Gram-negative bacteremia 84 (35.6) 42 (30.2) 42 (43.3) 0.053 Multidrug-resistant bacteremiab 60 (25.4) 33 (23.7) 27 (27.8) 0.544 Fungemia 24 (10.2) 10 (7.2) 14 (14.4) 0.082 Polymicrobialc 27 (11.4) 14 (10.1) 13 (13.4) 0.534 Values are presented as median (interquartile range) or number (%).

5) 92 (66.2) 46 (47.4) 0.005 Gram-negative bacteremia 84 (35.6) 42 (30.2) 42 (43.3) 0.053 Multidrug-resistant bacteremiab 60 (25.4) 33 (23.7) 27 (27.8) 0.544 Fungemia 24 (10.2) 10 (7.2) 14 (14.4) 0.082 Polymicrobialc 27 (11.4) 14 (10.1) 13 (13.4) 0.534 Values are presented as median (interquartile range) or number (%). ICU: intensive care unit; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; qSOFA: quick Sepsis-Related Organ Failure Assessment. a Chronic obstructive pulmonary disease, asthma, and bronchiectasis; b Including methicillin-resistant Staphylococcus aureus, extended-spectrum ß-lactamase-producing Gram-negative bacteria (Escherichia coli and Klebsiella pneumoniae), carbapenem-resistant Gram-negative rods (Acinetobacter baumannii and Pseudomonas aeruginosa), and vancomycin-resistant Enterococcus faecium; c A blood culture test revealed more than two bacteria. Table 2. Comparison of clinical data from patients with a qSOFA ≥2 or <2

b Including methicillin-resistant Staphylococcus aureus, extended-spectrum ß-lactamase-producing Gram-negative bacteria (Escherichia coli and Klebsiella pneumoniae), carbapenem-resistant Gram-negative rods (Acinetobacter baumannii and Pseudomonas aeruginosa), and vancomycin-resistant Enterococcus faecium; c A blood culture test revealed more than two bacteria. Table 2. Comparison of clinical data from patients with a qSOFA ≥2 or <2 Variable qSOFA ≥2 (n = 127) qSOFA <2 (n = 109) P-value Age (yr) 69 (57–76) 68 (57–76) 0.505 Male sex 79 (62.2) 65 (59.6) 0.691 APACHE II score 25 (19–31) 21 (15–26) <0.001 SOFA score 9 (6–11) 5 (3–7) <0.001 Hospital LOS (d) 24 (15–42) 27 (13–57) 0.355 ICU LOS (d) 11 (5–19) 9 (4–18) 0.540 Source of infection Pneumonia 58 (45.7) 33 (30.3) 0.016 Intra-abdominal 36 (28.3) 25 (22.9) 0.373 Urinary tract 12 (9.4) 13 (11.9) 0.672 Musculoskeletal 17 (13.4) 24 (22.0) 0.088 Catheter-related 11 (8.7) 11 (10.1) 0.823 Neutropenia 11 (8.7) 2 (1.8) 0.024 Infective endocarditis 4 (3.1) 7 (6.4) 0.354 Organism Gram-positive bacteria 64 (50.4) 74 (67.9) 0.008 Gram-negative bacteria 56 (44.1) 28 (25.7) 0.004 Requirement for hemodialysisa 55 (43.3) 29 (26.6) 0.009 Requirement for NMBAsa 33 (26.0) 16 (14.7) 0.037 Requirement for vasopressorsa 93 (73.2) 53 (48.6) <0.001 Ventilator carea 76 (59.8) 51 (46.8) 0.050 Thrombocytopeniaa,b 102 (80.3) 62 (56.9) <0.001 Lactic acid >2.0 mmol/L (n = 161)a 80 (80.0) 39 (63.9) 0.028 Septic shock (n = 161)a,c 67 (67.0) 29 (47.5) 0.020 28-Day mortality 66 (52.0) 31 (28.4) <0.001 Values are presented as median (interquartile range) or number (%).

.001 Ventilator carea 76 (59.8) 51 (46.8) 0.050 Thrombocytopeniaa,b 102 (80.3) 62 (56.9) <0.001 Lactic acid >2.0 mmol/L (n = 161)a 80 (80.0) 39 (63.9) 0.028 Septic shock (n = 161)a,c 67 (67.0) 29 (47.5) 0.020 28-Day mortality 66 (52.0) 31 (28.4) <0.001 Values are presented as median (interquartile range) or number (%). qSOFA: quick Sepsis-Related Organ Failure Assessment; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; LOS: length of stay; ICU: intensive care unit; NMBA: neuromuscular blocking agent. a All clinical courses developed within 72 hours after ICU admission; b Defined as a platelet count ≤150×109/L; c Based on the sepsis-3 consensus statement. Table 3. Comparison of clinical courses among the cutoff levels of some scores (MEWS ≥5, SIRS criteria ≥2, and SOFA score ≥7) for 28-day mortality

qSOFA: quick Sepsis-Related Organ Failure Assessment; APACHE: Acute Physiology and Chronic Health Evaluation; SOFA: Sequential Organ Failure Assessment; LOS: length of stay; ICU: intensive care unit; NMBA: neuromuscular blocking agent. a All clinical courses developed within 72 hours after ICU admission; b Defined as a platelet count ≤150×109/L; c Based on the sepsis-3 consensus statement. Table 3. Comparison of clinical courses among the cutoff levels of some scores (MEWS ≥5, SIRS criteria ≥2, and SOFA score ≥7) for 28-day mortality Variable MEWS SIRS SOFA ≥5 (n = 175) <5 (n = 61) P-value ≥2 (n = 228) <2 P-value ≥7 (n = 125) <7 (n = 111) P-value Requirement for hemodialysisa 71 (40.6) 13 (21.3) 0.008 81 (35.5) 3 (37.5) >0.999 59 (47.2) 25 (22.5) <0.001 Requirement for NMBAsa 45 (25.7) 4 (6.6) 0.001 49 (21.5) 0 0.211 35 (28.0) 14 (12.6) 0.004 Requirement for vasopressorsa 129 (73.7) 17 (27.9) <0.001 141 (61.8) 5 (62.5) >0.999 100 (80.0) 46 (41.4) <0.001 Ventilator carea 113 (64.6) 14 (23.0) <0.001 123 (53.9) 4 (50.0) >0.999 81 (64.8) 46 (41.4) <0.001 Thrombocytopeniaa,b 131 (74.9) 33 (54.1) 0.004 160 (70.2) 4 (50.0) 0.252 113 (90.4) 51 (45.9) <0.001 Lactic acid >2.0 mmol/L (n = 161)a 105 (78.9) 14 (50.0) 0.004 115 (73.7) 4 (80.0) >0.999 81 (78.0) 38 (65.5) 0.002 Septic shock (n = 161)a,c 84 (65.4) 9 (32.1) 0.001 92 (59.0) 4 (80.0) 0.649 72 (64.9) 24 (25.0) <0.001 28-Day mortality 84 (48.0) 13 (21.3) <0.001 134 (58.8) 5 (62.5) >0.999 69 (55.2) 28 (25.2) <0.001 Values are presented as number (%). The cutoff levels of MEWS and SIRS were used according to published data, and the cutoff value of SOFA score was determined by receiver operating characteristic curves using our data.

01 28-Day mortality 84 (48.0) 13 (21.3) <0.001 134 (58.8) 5 (62.5) >0.999 69 (55.2) 28 (25.2) <0.001 Values are presented as number (%). The cutoff levels of MEWS and SIRS were used according to published data, and the cutoff value of SOFA score was determined by receiver operating characteristic curves using our data. MEWS: Modified Early Warning score; SIRS: systemic inflammatory response syndrome; SOFA: Sequential Organ Failure Assessment; NMBA: Neuromuscular blocking agent. a All clinical courses were developed within 72 hours after ICU admission; b Defined as a platelet count ≤150×109/L; c Based on the sepsis-3 consensus statement.

Introduction The ability to predict injury severity quickly and accurately should lead to improved patient outcomes. Patient treatment and disposition (intensive care unit or ward) is determined by these initial assessments [1]. Several trauma scores have been developed to predict injury severity and the risk of mortality; the injury severity score (ISS) is the most commonly used. The ISS correlates with mortality, and severe trauma is defined as an ISS >15 [2]. Despite being commonly used to predict mortality, there are limitations to using the ISS as a decision-making tool in the clinical setting. It is complex and time-consuming to calculate, and is therefore generally used for audit and research purposes rather than for clinical decision making [1]. To counter this disadvantage, several studies have attempted to identify biochemical markers that can be used to predict mortality [1,3].

in the clinical setting. It is complex and time-consuming to calculate, and is therefore generally used for audit and research purposes rather than for clinical decision making [1]. To counter this disadvantage, several studies have attempted to identify biochemical markers that can be used to predict mortality [1,3]. The trauma triad for death in patients with severe traumatic injuries comprises hypothermia, acidosis, and coagulopathy. Severe hemorrhage in trauma reduces oxygen delivery and may lead to hypothermia, acidosis, and coagulopathy [3]. In 1982, Kashuk et al. [4] showed that a “bloody vicious cycle” involving hemorrhage and tissue injury in patients with severe trauma causes this predictable triad of complicating factors. They suggested that medical providers have an accurate understanding of the triad and that this understanding should serve as the cornerstone of all interventions provided to patients who have sustained severe traumatic injury. Based on this triad, we hypothesized that hemoglobin level (HbL), potential of hydrogen level (pHL), and prothrombin time/international normalized ratio (PT/INR) on arrival would be early predictors of mortality. We attempted to analyze the usefulness of these biochemical values as predictors of in-hospital mortality in severe trauma patients with an ISS >15, and analyzed the correlations between these values and the ISS.

prothrombin time/international normalized ratio (PT/INR) on arrival would be early predictors of mortality. We attempted to analyze the usefulness of these biochemical values as predictors of in-hospital mortality in severe trauma patients with an ISS >15, and analyzed the correlations between these values and the ISS. Materials and Methods 1) Study population We retrospectively and consecutively evaluated all patients with an ISS >15 who were treated for severe trauma between January 2005 and December 2015 at Gyeongsang National University Hospital. We identified 454 eligible patients. We excluded patients who had penetrating injuries, those who were discharged from the emergency department, and those who received fluid in the field. Based on these criteria, 139 patients were excluded and 315 patients were finally included in our study (Figure 1). The medical records and electronic laboratory results were reviewed and data were extracted on demographics; mechanism of injury; ISS; HbL, pHL, and PT/INR on arrival at the emergency department; and in-hospital mortality.

teria, 139 patients were excluded and 315 patients were finally included in our study (Figure 1). The medical records and electronic laboratory results were reviewed and data were extracted on demographics; mechanism of injury; ISS; HbL, pHL, and PT/INR on arrival at the emergency department; and in-hospital mortality. 2) Treatment protocol Since 2008, the center has provided a 24-hour service. During each shift, an emergency medicine specialist—who acts as the trauma team’s leader—resides in the hospital for the prompt treatment of trauma patients. Patients are classified as being “severe trauma patients” if they are expected, on initial examination by an emergency medicine specialist, to have an ISS >15. Patients classified as having severe trauma are examined within 1 hour, in cooperation with related departments, and most are admitted to the intensive care unit.

tients are classified as being “severe trauma patients” if they are expected, on initial examination by an emergency medicine specialist, to have an ISS >15. Patients classified as having severe trauma are examined within 1 hour, in cooperation with related departments, and most are admitted to the intensive care unit. 3) Definitions The ISS is an anatomic scoring system that provides an overall score for patients with multiple injuries. Each injury is assigned an abbreviated injury scale score and is allocated to 1 of the following six body regions: head, face, chest, abdomen, extremities (including the pelvis), and external. Only the highest abbreviated injury scale score for each body region is used. The three most severely injured body regions have their scores squared and summed to produce the ISS score. Severe trauma is defined as an ISS >15 [2]. Cutoff values were estimated by using receiver operating characteristic (ROC) curves. The highest value of the sum of sensitivity and specificity was defined as the cutoff value. Mortality was defined as all-cause in-hospital mortality; we did not evaluate specific causes of death.

trauma is defined as an ISS >15 [2]. Cutoff values were estimated by using receiver operating characteristic (ROC) curves. The highest value of the sum of sensitivity and specificity was defined as the cutoff value. Mortality was defined as all-cause in-hospital mortality; we did not evaluate specific causes of death. 4) Data analysis Missing data were not replaced or imputed. We calculated P-values using Pearson chi-square test or Fisher exact for categorical variables, and continuous data were correlated using the Pearson correlation coefficient. A P-value <0.05 was considered statistically significant. The predictive values of these variables were determined using ROC curves, with Bonferroni corrections performed for multiple comparisons. All statistical analyses were performed using SPSS version 24.0 (IBM Corp., Armonk, NY, USA) and R version 3.3.4 for Windows (R Foundation for Statistical Computing, Vienna, Austria). 5) Ethics approval This retrospective medical record review of the trauma registry at Gyeongsang National University Hospital was approved by Institutional Review Board (No. GNUH 2017-06-024). Results Between January 2005 and December 2015, our hospital managed 315 patients with severe trauma. Of these, 255 patients (81%) were male, 60 (19%) were female, and 72 (22.9%) died. Motor vehicle collisions were main cause of trauma (77%): these included car (21%), motorcycle (20%), pedestrian (19%), and cultivator (17%) accidents (Figure 2). The age, HbL, pHL, PT/INR, and ISS distributions are shown in Table 1.

ma. Of these, 255 patients (81%) were male, 60 (19%) were female, and 72 (22.9%) died. Motor vehicle collisions were main cause of trauma (77%): these included car (21%), motorcycle (20%), pedestrian (19%), and cultivator (17%) accidents (Figure 2). The age, HbL, pHL, PT/INR, and ISS distributions are shown in Table 1. ROC curves were used to estimate the sensitivity, specificity, and cutoff values of HbL, pHL, and PT/INR to predict in-hospital mortality. The in-hospital mortality rate for severely injured trauma patients with an HbL <8.4 g/dl was 49.8%, compared with an in-hospital mortality rate of 9.9% for those with an HbL ≥8.4 g/dl (P < 0.001; odds ratio [OR], 13.56). At this level, the HbL had a sensitivity of 81.9% and a specificity of 86.4% for mortality. Hence, an HbL of 8.4 g/dl was determined to be the cutoff value for in-hospital mortality. The positive predictive value (PPV) was lower than expected at 59.8%, but the negative predictive value (NPV) was higher than expected at 90.1%. A pHL of 7.25 was evaluated as the cutoff value for in-hospital mortality, as 66.7% of patients with a pHL <7.25 died versus 22.2% with a pHL ≥7.25 (P < 0.001; OR, 7.0). At this cutoff value, the sensitivity and specificity values were 66.7% and 77.8%, respectively. The PPV was once again low at 47.1%, and the NPV was high at 88.7%. A PT/INR value of 1.4 was estimated as being the cutoff for in-hospital mortality, as 37.5% of patients with a PT/INR value ≥1.4 died versus 16% with a PT/INR value <1.4 (P < 0.001; OR, 3.14). At this cutoff value, the sensitivity was 37.5%, specificity 84%, PPV 40.9%, and NPV 81.9%.

.1%, and the NPV was high at 88.7%. A PT/INR value of 1.4 was estimated as being the cutoff for in-hospital mortality, as 37.5% of patients with a PT/INR value ≥1.4 died versus 16% with a PT/INR value <1.4 (P < 0.001; OR, 3.14). At this cutoff value, the sensitivity was 37.5%, specificity 84%, PPV 40.9%, and NPV 81.9%. A low HbL was considerably more specific than a low pHL or elevated PT/INR for predicting in-hospital mortality (Table 2). Combining these three values in an analysis yielded the following results: 87.5% of patients with an HbL <8.4, pHL <7.25, and PT/INR level ≥1.4 died. In contrast, 35.2% of patients who met only one or two of these three cutoff values died. This composite had a sensitivity of 36.8% and a specificity of 97.1% for mortality; the PPV was 87.5% and the NPV 73.9% (Table 3). The ROC was used to demonstrate the sensitivity and specificity of the HbL, pHL, PT/INR, and ISS for predicting in-hospital mortality (Figure 3). The HbL (area under the curve [AUC], 0.895) was a more significant biochemical predictor of in-hospital mortality than pHL (AUC, 0.736) and PT/INR (AUC, 0.593). Using Pearson correlation coefficients (Table 4), the ISS correlated significantly with HbL, pHL, and PT/INR (P = 0.01).

ting in-hospital mortality (Figure 3). The HbL (area under the curve [AUC], 0.895) was a more significant biochemical predictor of in-hospital mortality than pHL (AUC, 0.736) and PT/INR (AUC, 0.593). Using Pearson correlation coefficients (Table 4), the ISS correlated significantly with HbL, pHL, and PT/INR (P = 0.01). Discussion Several trauma scores have proven useful as indicators of mortality. These can be used to identify the severity of trauma injury and might influence therapeutic decisions. The ISS is one of these scores. Although numerous recent publications have questioned its accuracy, it remains the most commonly used trauma score internationally [5,6]. The ISS can be used to identify patients with severe traumatic injury, but it requires knowledge of all anatomical injuries [7]. Complete diagnosis of all injuries may take many hours after admission to the emergency department, thus, the usefulness of the score in clinical decision making is limited [1]. Because the ISS has such limitations, we hypothesized that HbL, pHL, and PT/INR value would be early predictors of mortality, based on the trauma triad of death [3,4]. We analyzed the correlations between in-hospital mortality and HbL, pHL, PT/INR, and ISS using ROC curves; all values were statistically significant (AUC: HbL, 0.895; pHL, 0.736; PT/INR, 0.593; and ISS, 0.629). Pearson correlation coefficients showed that the ISS correlated significantly with HbL, pHL, and PT/INR level (P = 0.01).

the correlations between in-hospital mortality and HbL, pHL, PT/INR, and ISS using ROC curves; all values were statistically significant (AUC: HbL, 0.895; pHL, 0.736; PT/INR, 0.593; and ISS, 0.629). Pearson correlation coefficients showed that the ISS correlated significantly with HbL, pHL, and PT/INR level (P = 0.01). Knottenbelt [8] suggested that the initial hemoglobin measurement may prove useful as a predictor of mortality in patients with severe traumatic injury. In that study, the initial HbLs of 1,000 patients were collected and analyzed; the mortality rate was higher in those with an HbL <8 g/dl than with an HbL ≥8 g/dl (P < 0.001). Knottenbelt [8] thus suggested that a low HbL observed soon after injury is usually an indicator of serious ongoing hemorrhage and has important implications for management and prognosis. Another study showed that hemorrhage in trauma patients is associated with an early decrease in HbL. An HbL ≤10 g/dl within the first 30 minutes of patient arrival will correctly identify the presence or absence of significant bleeding. Based on that finding, the authors suggested that a hemoglobin drop within minutes of sustaining injuries can predict mortality [9]. Similarly, we found that an HbL <8.4 g/dl predicted in-hospital mortality in patients with severe trauma (AUC, 0.895; P < 0.001).

identify the presence or absence of significant bleeding. Based on that finding, the authors suggested that a hemoglobin drop within minutes of sustaining injuries can predict mortality [9]. Similarly, we found that an HbL <8.4 g/dl predicted in-hospital mortality in patients with severe trauma (AUC, 0.895; P < 0.001). Acidosis can induce organ dysfunction, affecting the heart, kidneys, and liver and can increase pulmonary vascular resistance. This may result in hypoperfusion [10]. The utility of pH in the assessment of trauma patients has been debated in the literature [10-12]. Kaplan and Kellum [10] showed that the initial pH measured in the emergency department discriminated survivors from non-survivors of major vascular injury. Other studies showed that acidemia was correlated with serum lactate—lactate levels were higher in patients with acidemia—and that in patients with severe trauma, metabolic acidosis was correlated with survival [11,12]. Those research findings are similar to ours. We found that academia (pH <7.25) on arrival can be used to predict in-hospital mortality, as it was correlated with in-hospital mortality in patients with severe traumatic injury (AUC, 0.736; P < 0.001).

evere trauma, metabolic acidosis was correlated with survival [11,12]. Those research findings are similar to ours. We found that academia (pH <7.25) on arrival can be used to predict in-hospital mortality, as it was correlated with in-hospital mortality in patients with severe traumatic injury (AUC, 0.736; P < 0.001). Traumatic coagulopathy is a hypocoagulable state that occurs in most patients with severe trauma. Immediately after sustaining severe injury, hypoperfusion may induce coagulopathy. This coagulopathy is caused by increased anticoagulation and hyperfibrinolysis via increased production of protein C and tissue plasminogen activators and decreased concentrations of plasminogen activator inhibitors and thrombin activatable fibrinolysis inhibitors [13-15]. PT/INR is commonly measured in trauma patients; it is used as a measure of coagulopathy. Verma and Kole [16] published a study about the association between PT/INR and mortality in trauma patients. They analyzed 99 trauma patients and showed that the INR is a good predictor of mortality and has high diagnostic accuracy. Peltan et al. [17] also showed that an INR-based definition of acute traumatic coagulopathy is associated with mortality. In their multicenter prospective observational study, acute traumatic coagulopathy (defined as a PT/INR >1.5) was significantly associated with all-cause mortality (OR, 1.88; P < 0.001) [17]. In our study, the PT/INR cutoff value was ≥1.4, but this variable had relatively low statistical significance (AUC, 0.593; P < 0.001).

ir multicenter prospective observational study, acute traumatic coagulopathy (defined as a PT/INR >1.5) was significantly associated with all-cause mortality (OR, 1.88; P < 0.001) [17]. In our study, the PT/INR cutoff value was ≥1.4, but this variable had relatively low statistical significance (AUC, 0.593; P < 0.001). Our study has several limitations. First, we only evaluated trauma patients managed at a single hospital, which may have introduced selection bias and limits the extrapolation of our findings to the entire population. Second, the sample size was relatively small; larger studies are needed to validate our findings regarding the predictors of mortality in patients with severe trauma. Third, our study was a retrospective evaluation, and as with all trauma registries, the accuracy of the recorded data may vary [18]. However, it should be noted that while the ISS can be a subjective measure, the HbL, pHL, and PT/INR are objective measures. Moreover, the HbL, pHL, and PT/INR values were recorded more accurately than were the ISS scores. While this weakens the comparison between the variables, it strengthens the argument for the use of biochemical scores to predict mortality. Last, we excluded some patients from the study because they had received fluids in the field. We excluded these patients to prevent skewing the results based on presumed hemodilution or the hyperchloremic acidosis that may be caused by rapid saline infusion [19].

for the use of biochemical scores to predict mortality. Last, we excluded some patients from the study because they had received fluids in the field. We excluded these patients to prevent skewing the results based on presumed hemodilution or the hyperchloremic acidosis that may be caused by rapid saline infusion [19]. Despite these limitations, this study is an important investigation of predictors associated with in-hospital mortality in patients with severe traumatic injury. Our study showed that initial HbL, pHL, and PT/INR were significant predictors of in-hospital mortality in severely injured trauma patients. Comparisons of these findings with those of other reports will enhance the prediction of mortality. We anticipate that these biochemical predictors of in-hospital mortality can more easily be used in an emergency department setting. Our findings suggest that close monitoring should be considered for patients with severe trauma and an HbL <8.4, pHL <7.25, or PT/INR ≥1.4. No potential conflict of interest relevant to this article was reported. Figure 1. Outline of patient selection and exclusion. ISS: injury severity score. Figure 2. Mechanism of injury. Figure 3. Receiver operation characteristic curves for in-hospital mortality. AUC, area under the curve; Hb: hemoglobin; INR, international normalized ratio; ISS, injury severity score. Table 1. Distribution of age, hemoglobin, potential of hydrogen, INR, and ISS

Figure 1. Outline of patient selection and exclusion. ISS: injury severity score. Figure 2. Mechanism of injury. Figure 3. Receiver operation characteristic curves for in-hospital mortality. AUC, area under the curve; Hb: hemoglobin; INR, international normalized ratio; ISS, injury severity score. Table 1. Distribution of age, hemoglobin, potential of hydrogen, INR, and ISS Variable Median Minimum Maximum Interquartile range Age (yr) 66 14 87 51–74 Hb (g/dl) 11.9 5.4 19.7 9.0–14.0 pH 7.37 6.89 7.55 7.25–7.42 PT/INR 1.22 0.90 2.85 1.11–1.38 ISS 36 17 59 33–43 INR: international normalized ratio; ISS: injury severity score; Hb: hemoglobin; PT: prothrombin time. Table 2. Pairwise comparison of the AUC values Variable Difference between AUC values Standard error Z P-value Hb–pH 0.161 0.041 3.898 <0.001 HbL–PT/INR 0.302 0.043 7.084 <0.001 Hb–ISS 0.266 0.027 6.487 <0.001 pH–PT/INR 0.143 0.143 2.683 0.044 pH–ISS 0.107 0.107 2.059 0.237 PT/INR–ISS 0.036 0.036 0.672 1 AUC: area under the curve; Hb: hemoglobin; HbL: hemoglobin level; PT: prothrombin time; INR: international normalized ratio; ISS: injury severity score. Table 3. The sensitivity, specificity, PPV, and NPV for in-hospital mortality Variable Hb <8.4 g/dl pH <7.25 PT/INR ≥1.4 Composite groupa Mortality (%) 49.8 66.7 37.5 87.5 Sensitivity (%) 81.9 66.7 37.5 36.8 Specificity (%) 86.4 77.8 84.0 97.1 PPV (%) 59.8 47.1 40.9 87.5 NPV (%) 90.1 88.7 81.9 73.9 PPV: positive predictive value; NPV: negative predictive value; Hb: hemoglobin; PT: prothrombin time; INR: international normalized ratio.

.4 Composite groupa Mortality (%) 49.8 66.7 37.5 87.5 Sensitivity (%) 81.9 66.7 37.5 36.8 Specificity (%) 86.4 77.8 84.0 97.1 PPV (%) 59.8 47.1 40.9 87.5 NPV (%) 90.1 88.7 81.9 73.9 PPV: positive predictive value; NPV: negative predictive value; Hb: hemoglobin; PT: prothrombin time; INR: international normalized ratio. a Patients with an hemoglobin level <8.4 g/dl, potential of hydrogen level <7.25, and PT/INR level ≥1.4. Table 4. Pearson correlation coefficients Variable Hb pH PT/INR ISS Hb Pearson correlation 1 0.228a –0.154a –0.151a P-value (two-tailed) <0.001 0.006 0.007 pH Pearson correlation 0.228a 1 –0.202a –0.345a P-value (two-tailed) <0.001 <0.001 <0.001 PT/INR Pearson correlation –0.154a –0.202a 1 0.223a P-value (two-tailed) 0.006 <0.001 <0.001 ISS Pearson correlation –0.151a –0.345a 0.223a 1 P-value (two-tailed) 0.007 <0.001 <0.001 Hb: hemoglobin; PT: prothrombin time; INR: international normalized ratio; ISS: injury severity score. a Correlation is significant at the level P < 0.01 (two-tailed).

In Korea, patient’s safety is becoming an important issue. Patients who experience adverse events during their hospital stay, including cardiopulmonary arrest, unplanned intensive care unit admissions, and unexpected death, show clear signs of deterioration in the hours preceding the event [1,2]. About one-half of the serious adverse events are deemed to be preventable [3]. Patients often show some signs of physiological deterioration for several hours (median 6 hours) before cardiac arrest [4,5]. Early recognition and response to patient deterioration have reduced the potential impact of such adverse events [6,7]. Health professionals need to recognize and respond to early signs of patient deterioration and activate rapid response systems (RRSs) to provide rapid medical intervention. RRSs have been developed for timely identification and treatment of patients in general wards at risk for clinical deterioration [8]. RRSs have been implemented widely around the world over the past two decades and have been shown to effectively reduce in-hospital cardiopulmonary arrests. Recently, RRSs have been implemented in some large hospital in Korea; their effectiveness was uncertain. This is the first multicenter survey on the impacts of RRSs. Implementation of RRSs showed a statistically significant reduction of the cardiopulmonary arrest rates (odds ratio [OR], 0.731; 95% confidence interval [CI], 0.577 to 0.927; P = 0.009), whereas cardiopulmonary resuscitation rates of 2013 and 2015 did not change in hospitals without RRS (OR, 0.988; 95% CI, 0.868 to 1.124; P = 0.854).

of RRSs showed a statistically significant reduction of the cardiopulmonary arrest rates (odds ratio [OR], 0.731; 95% confidence interval [CI], 0.577 to 0.927; P = 0.009), whereas cardiopulmonary resuscitation rates of 2013 and 2015 did not change in hospitals without RRS (OR, 0.988; 95% CI, 0.868 to 1.124; P = 0.854). RRS can diminish in-hospital cardiopulmonary arrests and improve patient safety through earlier identification and treatment attempts. Despite these benefits, there have been barriers against successful implementation of RRS. First, there is a lack of specialists and physicians for RRS implementation. Also, the optimal composition of the RRS team is uncertain. Two previous single-center reports did not show the benefits of intensivist-led teams compared with registrar or resident-led teams [9,10]. The majority of RRS interventions did not require the presence of a physician (fluids, oxygen, and diuretics). Maharaj et al. [11] reported that RRSs were associated with a reduction in hospital mortality and cardiopulmonary arrest. However, meta-regression did not identify the presence of a physician in the RRS to be significantly associated with a mortality reduction. We need to develop proper RRS model that is applicable to our country. Second, there is a lack of financial support for RRS. Moreover, the absence of government policy about RRS is an important issue to be addressed. In particular, the government needs to make efforts not to increase regional medical gaps. In the future, we need time to share experiences with RSS systems between hospitals.

econd, there is a lack of financial support for RRS. Moreover, the absence of government policy about RRS is an important issue to be addressed. In particular, the government needs to make efforts not to increase regional medical gaps. In the future, we need time to share experiences with RSS systems between hospitals. No potential conflict of interest relevant to this article was reported.

Introduction Due to population aging and chronicization of diseases, the severity of diseases among patients admitted to intensive care units (ICUs) has been increasing. Treatment costs have also been increasing for these patients. Since ICUs have limited medical resources, an objective patient triage protocol and evaluation of the severity of a condition must be developed to minimize unnecessary use of ICUs and allow for efficient use of the limited healthcare resources. Prognostic scoring systems have been developed to estimate the in-hospital mortality of ICU patients [1-4]. For this reason, over the last three decades, severity scoring systems, such as the Acute Physiology and Chronic Health Evaluation (APACHE) score, the Simplified Acute Physiology Score (SAPS), and the Mortality Probability Model (MPM), have been attempted to be used as a critical care triage criterion beyond predicting hospital mortality in critically ill patients [5]. The APACHE scoring system developed by Knaus et al. [1] achieved higher calibration than Zimmerman et al. [6] after a series of improvements, suggesting that APACHE IV is more accurate [3,7]. According to recent studies, APACHE IV exhibits satisfactory discriminatory performance both in the United States, where it was first developed, and outside the United States [8-11]. The older APACHE II model has been validated in Korean populations; it exhibits poor calibration and modest discrimination for hospital mortality [12]. Whereas, the performance of the APACHE IV has not been sufficiently examined.

in the United States, where it was first developed, and outside the United States [8-11]. The older APACHE II model has been validated in Korean populations; it exhibits poor calibration and modest discrimination for hospital mortality [12]. Whereas, the performance of the APACHE IV has not been sufficiently examined. This study aimed to investigate the suitability of APACHE IV severity scores and MPMs in an ICU within a tertiary general hospital, by analyzing the relationships among APACHE IV scores at the time of the admission, the predicted mortality rate, and the actual mortality rate comparing with APACHE II model. This study also verified the usefulness of the APACHE IV score as a standard triage protocol for admission in the ICU.

ertiary general hospital, by analyzing the relationships among APACHE IV scores at the time of the admission, the predicted mortality rate, and the actual mortality rate comparing with APACHE II model. This study also verified the usefulness of the APACHE IV score as a standard triage protocol for admission in the ICU. Materials and Methods 1) Patients The study was conducted in a 1,200-bed capacity referral hospital with four adults’ ICU: medical, surgical, emergency, and cardiac ICU, respectively. Two dedicated intensivists supervised all of 56 ICU beds in all ICUs on a semi-closed system. All patients were admitted between August 1, 2013 and July 31, 2014. The same type of patients suggested by Zimmerman et al. [6] for the development of APACHE IV were involved in this study. We excluded patients if they were younger than 17 years old or if the primary outcome of hospital mortality was uncertain. We also excluded patients with an ICU stay <48 hours, patients with burns, patients missing an APACHE IV score on day 1 in the ICU, patients who were admitted for simple postcardiovascular intervention monitoring, and patients who gave do-not-resuscitate orders. Three hundred and sixty-four patients were screened during the study period and 318 eligible patients were enrolled for analysis (Figure 1). This study was approved by the institutional review board of Chungnam National University Hospital (No. 2015-06-053), in accordance with the Declaration of Helsinki.

itate orders. Three hundred and sixty-four patients were screened during the study period and 318 eligible patients were enrolled for analysis (Figure 1). This study was approved by the institutional review board of Chungnam National University Hospital (No. 2015-06-053), in accordance with the Declaration of Helsinki. 2) Differences between APACHE II and APACHE IV model APACHE II score is calculated based on 12 physiologic criteria and estimates risk based on data available within the first 24 hours of an ICU stay [13,14]. APACHE IV was designed to assess the severity of illness as well as the prognosis in the ICU and has 17 physiological criteria, adding new variables such as mechanical ventilation, thrombolysis, impact of sedation on Glasgow Coma Scale, rescaled Glasgow Coma Scale, PaO2/ FiO2 ratio and disease-specific subgroups, to the existing APACHE III variables [6]. Disease-specific scoring systems have been developed for several important subgroups treated in the ICU since an APACHE III model [15]. 3) Data collection Two critical care fellows and one trained nurse prospectively collected and reviewed electronic medical records. The electronic medical records provided all of the data required to predict the mortality rate using APACHE II and APACHE IV. Predicted hospital mortalities were calculated using the equations of APACHE II and APACHE IV as follows: logit for APACHE II = –3.517 + (APACHE II) × 0.146. The APACHE IV score and predicted mortality rate calculation on a website (http://www.mecriticalcare.net/icu_scores/apacheIV.php) were used in the present study.

Predicted hospital mortalities were calculated using the equations of APACHE II and APACHE IV as follows: logit for APACHE II = –3.517 + (APACHE II) × 0.146. The APACHE IV score and predicted mortality rate calculation on a website (http://www.mecriticalcare.net/icu_scores/apacheIV.php) were used in the present study. 4) Identification of risk factors independent of the APACHE IV score and triage model development Cox proportional hazards regression was conducted to examine associations with death after adjustment for the APACHE IV score. Hazard ratios (HRs) were used to quantify the relationship between risk factors and death. An ICU triage model, which predicts hospital mortality, was constructed by combining the APACHE IV score and the other risk factors identified above the Cox regression models. 5) Statistical analysis Descriptive statistics are presented as medians and interquartile ranges, or as numbers with percentages. A univariate logistic regression analysis was performed to evaluate any associations between various risk factors and hospital mortality. Among the variables used in the model, the risk factors with P-values less than 0.05 were selected for multivariate logistic regression analysis.

ges, or as numbers with percentages. A univariate logistic regression analysis was performed to evaluate any associations between various risk factors and hospital mortality. Among the variables used in the model, the risk factors with P-values less than 0.05 were selected for multivariate logistic regression analysis. Discrimination is defined as the power to distinguish between survivors and non-survivors, and this was evaluated by receiver operating characteristic (ROC) analysis [6]. Calibration was defined as agreement between individual probabilities and actual outcomes. It was assessed using the Hosmer-Lemeshow goodness-of-fit C statistic with P-values greater than 0.05 indicating good calibration [7]. The standardized mortality ratio (SMR) was the ratio between the observed and predicted number of deaths. To test for statistical significance, we calculated 95% confidence interval (CI) according to the method described by Hosmer and Lemeshow [6]. We estimated HRs in univariate Cox proportional hazards models with 95% CIs and level of statistical significance. To test the discrimination ability of different combinations of parameters like APACHE IV, Charlson Comorbidity index (CCI), and department, we used the area under the receiver operating characteristic curve (AUROC) by Delong method. Statistical analysis was performed using SPSS version 21.0 (IBM Corp., Armonk, NY, USA). STATA version 12.0 (StataCorp., College Station, TX, USA) was used for C-statistics and the R and MKmisc version 3.2.2 was used for H-statistics.

Discrimination is defined as the power to distinguish between survivors and non-survivors, and this was evaluated by receiver operating characteristic (ROC) analysis [6]. Calibration was defined as agreement between individual probabilities and actual outcomes. It was assessed using the Hosmer-Lemeshow goodness-of-fit C statistic with P-values greater than 0.05 indicating good calibration [7]. The standardized mortality ratio (SMR) was the ratio between the observed and predicted number of deaths. To test for statistical significance, we calculated 95% confidence interval (CI) according to the method described by Hosmer and Lemeshow [6]. We estimated HRs in univariate Cox proportional hazards models with 95% CIs and level of statistical significance. To test the discrimination ability of different combinations of parameters like APACHE IV, Charlson Comorbidity index (CCI), and department, we used the area under the receiver operating characteristic curve (AUROC) by Delong method. Statistical analysis was performed using SPSS version 21.0 (IBM Corp., Armonk, NY, USA). STATA version 12.0 (StataCorp., College Station, TX, USA) was used for C-statistics and the R and MKmisc version 3.2.2 was used for H-statistics. Results 1) Baseline characteristics of the population The baseline characteristics of the patients are shown in Table 1. The total number of patients was 318, 79 of whom were non-survivors, exhibiting a 24.8% mortality rate. Among the mortality factors, age, CCI scores, unplanned ICU admission, use of vasoactive agents, acute respiratory distress syndrome (ARDS), severe sepsis or septic shock, APACHE II scores, and APACHE IV scores exhibited statistically significant differences between the survivors and non-survivors (P < 0.001). The mean age of the non-survivors was 70.7 ± 12.6 years, which was higher than that of the survivors (63.1 ± 15.5 years). The mean CCI score of the non-survivors was 5.4 ± 3.2 points, exceeding that of the survivors (3.4 ± 2.5 points), suggesting that the non-survivors had more comorbidities than the survivors. No statistically significant differences were found for sex, body mass index, and ICU lengths of stay between the survivors and non-survivors.

score of the non-survivors was 5.4 ± 3.2 points, exceeding that of the survivors (3.4 ± 2.5 points), suggesting that the non-survivors had more comorbidities than the survivors. No statistically significant differences were found for sex, body mass index, and ICU lengths of stay between the survivors and non-survivors. 2) Performance of the APACHE IV and APACHE II models in the prediction of hospital mortality Both prognostic models showed reasonable discrimination and calibration (Table 2). The AUROC of the APACHE IV and APACHE II models were 0.759 and 0.752. ROC curves for the two scoring systems are shown in Figure 2. Hosmer-Lemeshow goodness-of-fit test generated P-values >0.05 for both APACHE II and APACHE IV, indicating that the models were comparable. APACHE IV exhibited the same mortality prediction rate as the observed mortality. The APACHE II model exhibited a lower SMR than the APACHE IV model (SMR APACHE IV, 1.000 [95% CI, 0.789 to 1.250]; APACHE II, 0.991 [95% CI, 0.788 to 1.248]).

5 for both APACHE II and APACHE IV, indicating that the models were comparable. APACHE IV exhibited the same mortality prediction rate as the observed mortality. The APACHE II model exhibited a lower SMR than the APACHE IV model (SMR APACHE IV, 1.000 [95% CI, 0.789 to 1.250]; APACHE II, 0.991 [95% CI, 0.788 to 1.248]). 3) Performance characteristics of different combination of parameters for predicting hospital mortality The multivariate logistic regression model was used to determine independent risk factors for hospital mortality by using all variables with a P-value <0.05 in the univariate model. Among the variables used in the analysis, the APACHE IV score (OR, 1.023; P < 0.001; 95% CI, 1.012 to 1.034), the CCI (OR, 1.200; P < 0.001; 95% CI, 1.053 to 1.368), ARDS (OR, 13.187; P < 0.001; 95% CI, 3.941 to 44.122) and unplanned ICU admission (OR, 2.239; P = 0.015; 95% CI, 1.169 to 4.287) were associated independently with hospital mortality (Table 3). With a cutoff score of 93, the APACHE IV score predicted hospital mortality with the highest sensitivity and specificity. The highest sensitivity and specificity were observed for CCI scores greater than 3. Table 4 shows that the HRs for hospital mortality. Medical admission, ARDS, CCI >3, and an APACHE IV >93 score were significant risk factors for hospital mortality.

icted hospital mortality with the highest sensitivity and specificity. The highest sensitivity and specificity were observed for CCI scores greater than 3. Table 4 shows that the HRs for hospital mortality. Medical admission, ARDS, CCI >3, and an APACHE IV >93 score were significant risk factors for hospital mortality. The explanatory power of APACHE IV scores >93 in predicting the hospital mortality rate was 44.1%; the model explained 44.1% of the total variance. With regards to models in which medical admission, which is a risk factor for hospital mortality, CCI scores >3, and APACHE IV scores >93 were added, the APACHE IV scores >93 and medical admission model had an explanatory power of 48.9%, and the APACHE IV scores >93, medical admission, and CCL >3 model had an explanatory power of 53.8%. Therefore, including risk factors in the models improved their explanatory power. However, the discriminative ability of the prediction models was not satisfactory to use them in a triage protocol for admission in the ICU, since all three models had a C index lower than 0.7 (Table 5). Discussion The present study aimed to investigate the suitability of APACHE IV severity scores and MPMs in the ICU of a tertiary general hospital. The study also evaluated the usefulness of APACHE IV scores as a single criterion of a triage protocol for admission in ICUs.

The explanatory power of APACHE IV scores >93 in predicting the hospital mortality rate was 44.1%; the model explained 44.1% of the total variance. With regards to models in which medical admission, which is a risk factor for hospital mortality, CCI scores >3, and APACHE IV scores >93 were added, the APACHE IV scores >93 and medical admission model had an explanatory power of 48.9%, and the APACHE IV scores >93, medical admission, and CCL >3 model had an explanatory power of 53.8%. Therefore, including risk factors in the models improved their explanatory power. However, the discriminative ability of the prediction models was not satisfactory to use them in a triage protocol for admission in the ICU, since all three models had a C index lower than 0.7 (Table 5). Discussion The present study aimed to investigate the suitability of APACHE IV severity scores and MPMs in the ICU of a tertiary general hospital. The study also evaluated the usefulness of APACHE IV scores as a single criterion of a triage protocol for admission in ICUs. There has been much debate on the need for objective directives to follow for an ICU admission triage, which aims to efficiently provide critically ill patients with resources within ICUs [16-18]. A substantial amount of research has also been conducted on the efficacy of existing physiological scores in predicting mortality in ICUs [19-22]. Previous studies have proposed various triage protocols for ICU admission [23].

which aims to efficiently provide critically ill patients with resources within ICUs [16-18]. A substantial amount of research has also been conducted on the efficacy of existing physiological scores in predicting mortality in ICUs [19-22]. Previous studies have proposed various triage protocols for ICU admission [23]. Since the APACHE scoring system is based on objective physiological factors, it eliminates the possibility of errors made by the user. It also allows for simultaneous comparison and prospective analyses of patients from different ICUs, and can be applied to a wide diversity of patients. Therefore, it was our hypothesis that the APACHE scoring system would be useful as the admission triage criterion.

iminates the possibility of errors made by the user. It also allows for simultaneous comparison and prospective analyses of patients from different ICUs, and can be applied to a wide diversity of patients. Therefore, it was our hypothesis that the APACHE scoring system would be useful as the admission triage criterion. This study showed that APACHE IV had good calibration and modest discrimination among the critically ill patients in a single center. Moreover, APACHE IV showed an SMR close to the actual mortality rate. Therefore, the suitability of APACHE IV for evaluating the severity of patients’ conditions and predicting their prognoses was verified in this urban referral hospital. It was also found that APACHE IV makes more accurate predictions of patients’ prognoses compared to APACHE II scores in even single center. This could be explained due to the advance of APACHE IV model using additional factors such as mechanical ventilation support, disease specific subgroup analysis, and the specific reason for ICU admission. Daley et al. [24] pointed out that APACHE II has been widely used for measuring ICU performance but this scoring system is not disease specific [25-29]. The same as that APACHE IV scoring system show more reliable prediction in Asia population were observed in other validation studies [30-33].

reason for ICU admission. Daley et al. [24] pointed out that APACHE II has been widely used for measuring ICU performance but this scoring system is not disease specific [25-29]. The same as that APACHE IV scoring system show more reliable prediction in Asia population were observed in other validation studies [30-33]. In aspect of the discriminative ability of new prediction models, the result exhibited unsatisfactory discrimination to use them in a triage protocol for admission in the ICU, as showing all three models had a C index lower than 0.7. There are some reasons why the new prediction model showed unsatisfactory discrimination as an ICU triage protocol. First, the semi-closed system of the ICUs in which the present study was conducted would have prevented lots of patients with a too-sick-to-benefit status from admitting to ICUs in advance by intensivists [34]. Second, the sample size was smaller than that of previous studies on triage models leading to insufficient statistical analysis [35]. Finally, despite establishing 93 as our cutoff score for APACHE IV, the higher score of a prognostic model has greater explanatory power. Though, the present study was limited to comparing the scores divided into only two ranges (<93 or ≥93).

s studies on triage models leading to insufficient statistical analysis [35]. Finally, despite establishing 93 as our cutoff score for APACHE IV, the higher score of a prognostic model has greater explanatory power. Though, the present study was limited to comparing the scores divided into only two ranges (<93 or ≥93). This study had some strengths. The validity of APACHE IV in surgical and medical ICUs, emergency ICUs, and cardiovascular ICUs was evaluated by using data of the same type of patients suggested by Zimmerman et al. [6] for the development of APACHE IV. Second, the present study is meaningful in that it attempted to verify the validity of a standard prognostic scoring system for few domestic studies. Moreover, the results suggest that the APACHE IV system makes more accurate predictions of patients’ prognoses compared to APACHE II scores even in single centers. Finally, within the scope what we know, there is rare research demonstrating that the APACHE IV model would not be advisable as a single criterion for admission in Korea ICU. However, the study results may contain selection bias since this study took place in only one institution. We could not obtain information on the survival status of patients who could not be monitored due to loss to follow up. And this study had a smaller sample size of 318 patients compared to the sample sizes used in previous studies [19,20,36].

y contain selection bias since this study took place in only one institution. We could not obtain information on the survival status of patients who could not be monitored due to loss to follow up. And this study had a smaller sample size of 318 patients compared to the sample sizes used in previous studies [19,20,36]. According to our results, in order to establish objective criteria for admission to ICUs, overall clinical judgment of internal medicine patients, severity of comorbidities, main diagnosis type (ARDS, septic shock), age, likelihood of recovery, opinions of medical professionals, patients’ conditions, and prognoses may be helpful. And multilateral research involving larger patient populations and disease groups is essential. In conclusion, in this work, the APACHE IV scoring system exhibits satisfactory discrimination and excellent calibration, but the result supposed that it was not appropriate to be used as a single criterion for ICU admission. Further research on determining the ICU admission priority of critically ill patients would be necessary. No potential conflict of interest relevant to this article was reported. Study design and interpretation of data: Sang IL Park Figure 1. Flow chart of the study population. Initially, 364 intensive care unit patients were enrolled from August 1, 2013 to July 31, 2014. The following patients were excluded: patients who were being readmitted (n = 2), patients who had missing data (n = 8), pediatric patients (n = 1), duplicated data (n = 7), and patients who lost to follow up (n = 28).

tion. Initially, 364 intensive care unit patients were enrolled from August 1, 2013 to July 31, 2014. The following patients were excluded: patients who were being readmitted (n = 2), patients who had missing data (n = 8), pediatric patients (n = 1), duplicated data (n = 7), and patients who lost to follow up (n = 28). Figure 2. Comparison of the area under the receiver operating characteristic curves of APACHE II and APACHE IV. The areas under the receiver operating characteristic curve were 0.759 and 0.752 in APACHE IV and APACHE II, respectively. APACHE: Acute Physiology and Chronic Health Evaluation. Table 1. Patient characteristics, scores, and predicted mortality in each prognostic model

Figure 2. Comparison of the area under the receiver operating characteristic curves of APACHE II and APACHE IV. The areas under the receiver operating characteristic curve were 0.759 and 0.752 in APACHE IV and APACHE II, respectively. APACHE: Acute Physiology and Chronic Health Evaluation. Table 1. Patient characteristics, scores, and predicted mortality in each prognostic model Variable Total (n = 318) Survivor (n = 239) Non-survivor (n = 79) P-value Age (yr) 63.1 ± 15.5 70.7 ± 12.6 <0.001 Male sex 131 (54.8) 49 (62.0) 0.262 Body mass index (kg/m2) 22.6 ± 3.9 22.6 ± 4.2 0.948 Charlson Comorbidity index score, age adjusted 3.4 ± 2.5 5.4 ± 3.2 <0.001 Route of admission 315 0.001 Ward 68 (29.2) 39 (50.6) Emergency room 168 (70.0) 38 (49.4) Other ICU/hospital 2 (0.9) 0 Admission type 305 0.092 Emergency surgery 30 (13.1) 9 (11.8) Elective surgery 47 (20.5) 5 (6.6) No surgery 152 (66.4) 62 (81.6) Unplanned ICU admission 307 79 (34.6) 47 (59.5) <0.001 Comorbidities Heart failure 22 (9.2) 3 (3.8) 0.122 Solid cancer 16 (6.7) 4 (5.1) 0.791 Chronic pulmonary disease 11 (4.6) 8 (10.1) 0.097 Infection-related admission 292 28 (13.0) 16 (20.8) 0.103 Mechanical ventilation (invasive) 74 (31.0) 35 (44.3) 0.030 CRRT 18 (7.5) 12 (15.2) 0.043 Vasoactive agent 52 (21.8) 42 (53.2) <0.001 Major diagnosis ARDS 5 (2.1) 19 (24.1) <0.001 Severe sepsis or septic shock 33 (13.8) 34 (43.0) <0.001 CPCR survivor 15 (6.3) 10 (12.7) 0.068 APACHE II score 21.0 ± 8.7 29.4 ± 8.8 <0.001 APACHE II predicted mortality 46.0 ± 11.1 56.6 ± 11.0 <0.001 APACHE IV score 71.4 ± 32.6 105.3 ± 37.8 <0.001 APACHE IV predicted mortality 24.5 ± 24.2 50.6 ± 30.1 <0.001 ICU length of stay (d) 11.0 ± 12.8 12.5 ± 12.2 0.367 Hospital length of stay (d) 41.7 ± 47.4 39.0 ± 89.2 0.736 Values are presented as mean ± standard deviation or number (%).

lity 46.0 ± 11.1 56.6 ± 11.0 <0.001 APACHE IV score 71.4 ± 32.6 105.3 ± 37.8 <0.001 APACHE IV predicted mortality 24.5 ± 24.2 50.6 ± 30.1 <0.001 ICU length of stay (d) 11.0 ± 12.8 12.5 ± 12.2 0.367 Hospital length of stay (d) 41.7 ± 47.4 39.0 ± 89.2 0.736 Values are presented as mean ± standard deviation or number (%). ICU: intensive care unit; CRRT: continuous renal replacement therapy; ARDS: acute respiratory distress syndrome; CPCR: cardiopulmonary cerebral resuscitation; APACHE: Acute Physiology and Chronic Health Evaluation. Table 2. Discrimination and calibration of APACHE IV and APACHE II Model No.a AUROC (95% CI) Hosmer-Lemeshow goodness-of-fit test SMR C-test P-value H-test P-value APACHE IV 304 0.759 (0.699–0.819) 3.42 0.905 7.679 0.465 1.000 (0.789–1.250) APACHE II 304 0.752 (0.692–0.811) 4.55 0.805 7.817 0.452 0.991 (0.788–1.248) APACHE: Acute Physiology and Chronic Health Evaluation; AUROC: area under the receiver operating characteristic curve; CI: confidence interval; SMR: standardized mortality ratio. a Among a total of 318 patients, 304 have both APACHE II and APACHE IV score data and the others could not be calculated. Table 3. Independent risk factors for hospital mortality

Model No.a AUROC (95% CI) Hosmer-Lemeshow goodness-of-fit test SMR C-test P-value H-test P-value APACHE IV 304 0.759 (0.699–0.819) 3.42 0.905 7.679 0.465 1.000 (0.789–1.250) APACHE II 304 0.752 (0.692–0.811) 4.55 0.805 7.817 0.452 0.991 (0.788–1.248) APACHE: Acute Physiology and Chronic Health Evaluation; AUROC: area under the receiver operating characteristic curve; CI: confidence interval; SMR: standardized mortality ratio. a Among a total of 318 patients, 304 have both APACHE II and APACHE IV score data and the others could not be calculated. Table 3. Independent risk factors for hospital mortality Variable Hospital mortality Odds ratio P-valuea APACHE IV score 1.023 (1.012–1.034) <0.001 Age 1.015 (0.989–1.045) 0.223 CCI score 1.200 (1.053–1.368) <0.001 Mechanical ventilation 0.534 (0.254–1.120) 0.097 Vasoactive agent 1.697 (0.798–3.610) 0.169 Sepsis 1.609 (0.729–3.550) 0.239 ARDS 13.187 (3.941–44.122) <0.001 Unplanned ICU admission 2.239 (1.169–4.287) 0.015 Among a total of 318 patients, only 307 were included in the regression analysis due to the variable of “unplanned ICU admission statuses” for 11 patients being unknown. APACHE: Acute Physiology and Chronic Health Evaluation; CCI: Charlson Comorbidity index; ARDS: acute respiratory distress syndrome; ICU: intensive care unit. a P < 0.05. Table 4. Predicting value for hospital mortality in patients assessed by Cox proportion hazards models

Variable Hospital mortality Odds ratio P-valuea APACHE IV score 1.023 (1.012–1.034) <0.001 Age 1.015 (0.989–1.045) 0.223 CCI score 1.200 (1.053–1.368) <0.001 Mechanical ventilation 0.534 (0.254–1.120) 0.097 Vasoactive agent 1.697 (0.798–3.610) 0.169 Sepsis 1.609 (0.729–3.550) 0.239 ARDS 13.187 (3.941–44.122) <0.001 Unplanned ICU admission 2.239 (1.169–4.287) 0.015 Among a total of 318 patients, only 307 were included in the regression analysis due to the variable of “unplanned ICU admission statuses” for 11 patients being unknown. APACHE: Acute Physiology and Chronic Health Evaluation; CCI: Charlson Comorbidity index; ARDS: acute respiratory distress syndrome; ICU: intensive care unit. a P < 0.05. Table 4. Predicting value for hospital mortality in patients assessed by Cox proportion hazards models Risk factor Hazard ratio (95% CI) P-valuea Multivariate analysis Surgical department admission Reference Medical department admission 3.534 (1.634–7.692) 0.001 Acute respiratory distress syndrome 2.661 (1.526–4.640) 0.001 CCI score >3 1.819 (1.061–3.120) 0.030 APACHE IV score >93 2.140 (1.304–3.510) 0.003 CI: confidence interval; CCI: Charlson Comorbidity index; APACHE: Acute Physiology and Chronic Health Evaluation. a P < 0.05. Table 5. Performance characteristics of different combination of parameters for predicting hospital mortality in patients

Risk factor Hazard ratio (95% CI) P-valuea Multivariate analysis Surgical department admission Reference Medical department admission 3.534 (1.634–7.692) 0.001 Acute respiratory distress syndrome 2.661 (1.526–4.640) 0.001 CCI score >3 1.819 (1.061–3.120) 0.030 APACHE IV score >93 2.140 (1.304–3.510) 0.003 CI: confidence interval; CCI: Charlson Comorbidity index; APACHE: Acute Physiology and Chronic Health Evaluation. a P < 0.05. Table 5. Performance characteristics of different combination of parameters for predicting hospital mortality in patients Model C-index (95% CI) P-value APACHE IV >93 0.680 (0.626–0.731) Reference APACHE IV >93 + medical department 0.686 (0.632–0.737) 0.679 APACHE IV >94 + medical department + CCI >3 0.659 (0.604–0.711) 0.372 The area under the receiver operating characteristic curve by Delong method was used. CI: confidence interval; APACHE: Acute Physiology and Chronic Health Evaluation; CCI: Charlson Comorbidity index.

Neuroleptic malignant syndrome (NMS) is an uncommon, but potentially life-threatening side effect of antipsycotic medications.[1] NMS is a known idiosyncratic and unpredictable adverse reaction that is related to the administration of dopamine antagonists.[2] We report a case of NMS in a schizophrenic patient who was treated with clozapine for years prior to the onset of NMS. Case Report A 47-year-old Korean woman was diagnosed with schizophrenia and treated with clozapine for years, from 2004 onward. It was determined through outpatient follow-up appointments that the patient had poor adherence to her prescribed medication regimen. Additionally, the patient and her spouse had divorced within the past year. The patients’ mother passed away 5 months preceding the event. The patient did not have a regular caregiver; however, 3 days prior to the adverse event she moved in with her son. While she was able to communicate with her son until the day preceding the event, she was drowsy and incoherent when found at home.