Browse the corpus

Thirteen U.S. and European studies have documented the benefits of blood glucose awareness training (BGAT) (1). These benefits include improvements in detecting and reducing the occurrences of extreme blood glucose levels and their sequelae, e.g., reducing occurrence of ketoacidosis, severe hypoglycemia, hypoglycemia-related driving mishaps, and fear of hypoglycemia. We hypothesized that an internet version of BGAT would be perceived as useful, be completed efficiently, and produce greater clinical benefits compared with a wait-list control group. RESEARCH DESIGN AND METHODS A notice in Diabetes Forecast inviting participants to evaluate BGAThome.com resulted in 210 individuals completing an online screening in 10 days. Participants were the first 40 individuals who, by telephone interviews, met the following inclusion criteria: type 1 diabetes, routinely measuring blood glucose levels more than twice a day, and willingness to devote 1–2 h/week for 8–10 weeks to completing BGAThome. Of 108 responders telephoned, 38 were unreachable, 14 were ineligible, and 10 declined participation (Table 1).

e interviews, met the following inclusion criteria: type 1 diabetes, routinely measuring blood glucose levels more than twice a day, and willingness to devote 1–2 h/week for 8–10 weeks to completing BGAThome. Of 108 responders telephoned, 38 were unreachable, 14 were ineligible, and 10 declined participation (Table 1). After signing institutional review board–approved informed consent, participants were mailed a handheld computer (HHC) and a LifeScan One-Touch meter with supplies for one month's use. Participants were instructed to 1) activate the HHC before performing routine self monitoring of blood glucose (SMBG); 2) enter an estimate of their current blood glucose level; 3) based on this estimate, indicate whether they should then eat fast-acting carbohydrates, engage in vigorous exercise, or drive; and 4) perform SMBG and record their actual blood glucose levels. After returning the HHC, participants completed online a demographic questionnaire, the Diabetes Knowledge Scale, and the Hypoglycemia Fear Survey (2). The HHC and questionnaire data were collected again 12 weeks later, along with Likert-scale items assessing BGAThome's benefits and usability.

ual blood glucose levels. After returning the HHC, participants completed online a demographic questionnaire, the Diabetes Knowledge Scale, and the Hypoglycemia Fear Survey (2). The HHC and questionnaire data were collected again 12 weeks later, along with Likert-scale items assessing BGAThome's benefits and usability. Users were given 12 weeks to complete BGAThome's units, detailed elsewhere (1). Central to BGAT is completing daily blood glucose diaries, in which participants 1) record relevant blood glucose information and symptoms, 2) estimate their current blood glucose, 3) receive feedback on their estimate accuracy by performing and recording SMBG, 4) interpret the clinical significance of their accuracy with the error grid (2), and 5) anticipate their blood glucose level 1 h later. To encourage use of blood glucose diaries, participants were only given access to the next unit 7 days following completion of the previous unit.

acy by performing and recording SMBG, 4) interpret the clinical significance of their accuracy with the error grid (2), and 5) anticipate their blood glucose level 1 h later. To encourage use of blood glucose diaries, participants were only given access to the next unit 7 days following completion of the previous unit. With internet delivery to a heterogeneous sample, individuals would be expected to pursue BGAT for various reasons. Thus, the primary outcome variable would need to incorporate a variety of possible desired outcomes. Consequently, our Improved Functioning Score (IFS) is a composite score where assessment-dependent variables are converted to Z scores. Assessment 2 performance was converted to Z scores based on assessment 1's mean and SD. Z scores for each outcome variable were totaled, where zero reflects average baseline functioning for all variables and +1 reflects performance across all variables one SD above the sample's baseline mean (3). It incorporated the following variables from questionnaires: Diabetes Knowledge Scale (percent correct) and Hypoglycemic Fear Survey (sum of Worry subscale). It also incorporated the following variables from the HHC: percent SMBG readings within target range (3.9 −10.0 mmol/l), number of undetected blood glucose readings <3.9 mmol/l, overall blood glucose estimation accuracy (Accuracy Index) (4), when blood glucose levels are <3.9 mmol/l, number of risky decisions to drive, not eat fast-acting carbohydrates, and exercise.

the HHC: percent SMBG readings within target range (3.9 −10.0 mmol/l), number of undetected blood glucose readings <3.9 mmol/l, overall blood glucose estimation accuracy (Accuracy Index) (4), when blood glucose levels are <3.9 mmol/l, number of risky decisions to drive, not eat fast-acting carbohydrates, and exercise. RESULTS Two wait-list control group participants and one BGAThome participant dropped out during the treatment period. Two BGAThome participants dropped out during assessment 1. ANOVA demonstrated that BGAThome resulted in greater improvement in IFS: interaction F(1,33) = 4.20; P = 0.048 (Table 1). On a scale of 1–5 in which 1 = Not at all and 5 = Very, treatment participants rated BGAThome as beneficial, easy to use, and enjoyable (3.8 ± 1.17, 3.9 ± 0.73, and 3.8 ± 1.04, respectively). On average, participants completed BGAThome in 11 weeks, logged onto BGAThome.com 30.4 ± 16.51 times, and spent 26.4 ± 16.3 min on each unit. These measures of use indicate trends toward a relationship between more website use and increased benefits. More time spent on units was associated with greater IFS improvement (r = −0.36, P = 0.10). More frequent log-ons were associated with greater improvement in knowledge (r = 0.49, P = 0.03) and lower blood glucose levels <50 mg/dl (r = −0.54, P = 0.02). Age was not correlated with IFS improvement; however, education tended to be associated with improved IFS (r = 0.45, P = 0.07).

rovement (r = −0.36, P = 0.10). More frequent log-ons were associated with greater improvement in knowledge (r = 0.49, P = 0.03) and lower blood glucose levels <50 mg/dl (r = −0.54, P = 0.02). Age was not correlated with IFS improvement; however, education tended to be associated with improved IFS (r = 0.45, P = 0.07). CONCLUSIONS BGAThome was found to be beneficial, easy to use, and enjoyable. This is the first time BGAT was made available to individuals with various goals, needs, diabetes regimens, and resources. Despite this heterogeneity, BGAThome improved performance, summed across all eight dependent variables an average of 2.37 SDs. Greater BGAThome use appeared to yield improved benefits. Engagement might be further enhanced by 1) incorporating a chat room where users share experiences and support; 2) employing a group context, led by a diabetes educator (5); 3) undergoing an initial motivational interview (6); 4) fiscally investing in training; and 5) having a pressing personal goal, such as achieving tight metabolic control because of pregnancy without increasing risk of severe hypoglycemia (7) or following a costly hypoglycemia-related driving mishap. While our final participant sample came from 35 different U.S. cities and 21 different states, allowing greater external validity, the sample size and its demographic composition (white, middle-aged, educated individuals) was a limitation of this study. A larger, more representative sample would also allow investigation into the role of socioeconomic status, race, and education.

. cities and 21 different states, allowing greater external validity, the sample size and its demographic composition (white, middle-aged, educated individuals) was a limitation of this study. A larger, more representative sample would also allow investigation into the role of socioeconomic status, race, and education. Nevertheless, this study indicates the possible benefits of disseminating BGAThome over the internet in a personalized and self-directed format, serving a large number of individuals in a cost-effective manner. This study was supported by the American Diabetes Association; the National Institutes of Health Grant DK28288; LifeScan, Inc.; and contributions by post-doctoral fellow Kushal Patel.

Recent studies found 1,5-anhydroglucitol (1,5-AG) to generally reflect postprandial hyperglycemia (1–4). However, these studies were cross-sectional, had a comparably short follow-up (1–4), or included patients in the pre-diabetic state (3). The present study assessed the correlation of 2-h postprandial glucose measurements, frequently recommended in clinical practice, with 1,5-AG and aimed at defining the time interval of glucose values yielding the closest correlation with 1,5-AG in diabetic patients.

, or included patients in the pre-diabetic state (3). The present study assessed the correlation of 2-h postprandial glucose measurements, frequently recommended in clinical practice, with 1,5-AG and aimed at defining the time interval of glucose values yielding the closest correlation with 1,5-AG in diabetic patients. RESEARCH DESIGN AND METHODS This was a prospective study at three large Swiss hospitals assessing the impact of strategies to improve postprandial glucose by optimizing the combination of oral antidiabetes drugs and/or insulin according to a prespecified scheme (5). Included patients had type 2 diabetes ≥6 months with A1C between 7.0 and 12.0%. Patients with renal insufficiency or proteinuria were excluded. Written consent was obtained, and the study was approved by the local ethics committees. The study duration was 12 months with clinical visits every 3 months. Patients regularly monitored their blood glucose levels before and 2 h after main meals (Glucotrend Premium; Roche Diagnostics, Mannheim, Germany), and 1,5-AG was measured at each time point in a central laboratory (Lana 1.5 Auto Liquid Reagent; InterBiotech, Tokyo, Japan; automated on a Hitachi 917 Analyzer, coefficient of variation 5.2%). A1C was measured on a DCA 2000 (Bayer, Leverkusen, Germany). Postprandial glucose values were included in the analysis if measured between 110 and 130 min after the corresponding preprandial measurement. Correlation coefficients were calculated using standardized linear regression. In sensitivity analyses, the models were adjusted for age, sex, treatment modalities, time in study, and time of day. 1,5-AG measurements were examined for correlations with postprandial glucose values taken over the following prespecified preceding time periods: 3 days, 1 week, then weekly up to 12 weeks. All analyses were performed using Stata 10.0 (Stata, College Station, TX).

, treatment modalities, time in study, and time of day. 1,5-AG measurements were examined for correlations with postprandial glucose values taken over the following prespecified preceding time periods: 3 days, 1 week, then weekly up to 12 weeks. All analyses were performed using Stata 10.0 (Stata, College Station, TX). RESULTS All 55 patients (19 women and 36 men) contributed to the present analysis. The mean age was 61.3 ± 9.6 years (mean ± SD). The average number of self-measurements of blood glucose per patient and year were 405 ± 224 (fasting/preprandial) and 230 ± 122 (postprandial). Mean fasting/preprandial glucose was 155 ± 48 mg/dl at the beginning and 133 ± 46 mg/dl at the end of the study, and corresponding values for postprandial glucose were 172 ± 55 and 162 ± 53 mg/dl. Mean A1C was 8.7 ± 1.3% at baseline and 7.7 ± 1.0% at study end. Mean 1,5-AG was 4.2 ± 3.5 μg/ml at baseline and 6.4 ± 3.5 μg/ml at study end. A1C and 1,5-AG were negatively correlated (r = −0.42, P < 0.001). The correlation coefficient of postprandial glucose values with 1,5-AG varied across the prespecified time periods preceding the measurement of 1,5-AG (Fig. 1): −0.34 (P < 0.05) for 3 days, −0.38 (P < 0.001) for 1 week, and −0.40 (P < 0.001) for 2 weeks. Afterward, the strength of the correlation decreased (P < 0.001 for all remaining correlations). The association of fasting/preprandial glucose and 1,5-AG was lower (−0.19 to −0.23, P < 0.001 for all correlations) and did not reveal a time dependency (Fig. 1). Adjusting analyses for age, sex, treatment modalities, time in study, and time of day did not change the observed time dependency (data not shown). Time-specific correlation coefficients were 0.26–0.28 for A1C with fasting/preprandial and 0.22–0.30 for A1C with postprandial glucose, without evidence of a time dependency.

justing analyses for age, sex, treatment modalities, time in study, and time of day did not change the observed time dependency (data not shown). Time-specific correlation coefficients were 0.26–0.28 for A1C with fasting/preprandial and 0.22–0.30 for A1C with postprandial glucose, without evidence of a time dependency. CONCLUSIONS The present study is the first to longitudinally assess the association of 1,5-AG with 2-h postprandial glucose values in diabetic patients followed over 12 months in an outpatient setting. Although correlations were moderate (<0.5 in magnitude), 1,5-AG best reflected 2-h postprandial glucose values in the 2 preceding weeks. No time dependency resulted for the association of fasting glucose values and 1,5-AG. Correlations were weaker for A1C with fasting/preprandial as well as with postprandial glucose. A comparable albeit slightly stronger correlation of 1,5-AG with postprandial glucose in a population of type 1 and 2 diabetic patients was recently reported. Of note, Dungan et al. (1) used a continuous glucose monitoring system to measure the area under the curve for glucose levels exceeding 180 mg/dl, while the present study was based on self-measured glucose values. The lower number of postprandial compared with preprandial values reflects the difficulty to motivate patients for additional measurements and substantiates the role of 1,5-AG as a substitute for postprandial glucose measurements, complementing the widely used A1C and fructosamine measurements.

measured glucose values. The lower number of postprandial compared with preprandial values reflects the difficulty to motivate patients for additional measurements and substantiates the role of 1,5-AG as a substitute for postprandial glucose measurements, complementing the widely used A1C and fructosamine measurements. A recent cross-sectional study in prediabetic Japanese patients found a higher correlation with 2-h postprandial glucose (3). It is noteworthy that 1,5-AG is subject to urinary excretion followed by almost complete reabsorption, which is competitively inhibited by glucose if the renal threshold for glucosuria is exceeded (6). While glucosuria in prediabetic individuals mainly occurs in the context of carbohydrate loading, diabetic patients may reveal increased glucose values independently of meals. Moreover, differences in absolute levels of 1,5-AG between Asian and Caucasian individuals have been previously reported (7). In conclusion, this longitudinal study in an outpatient setting shows that 1,5-AG best reflects 2-h postprandial glucose values in the 2 preceding weeks. This study was supported by the Swiss National Science Foundation (grant 3233B0–115212 to C.S.) and unrestricted grants from Roche Switzerland, NovoNordisk Switzerland, and Lilly Switzerland. We thank Vreni Wyss, Gabi Pfenninger, Christiane Schwarzenbach, and Heinz Karrer for technical and administrative support.

Paraoxonase-1 (PON1) is a calcium-dependent HDL-associated enzyme that protects LDLs from oxidation. In type 1 diabetic patients, the serum paraoxonase concentration is lower and HDL is a less efficient antioxidant than in healthy individuals (1). Oxidized LDL is implicated in the pathogenesis of atherosclerosis, diabetic retinopathy, and nephropathy (2). Variations in lipoprotein-related enzymes and genotypes may also promote diabetic microvascular damage (3). Soft-tissue thickening is associated with chronic hyperglycemia and is hypothesized to be due to collagen glycation (4). With use of ultrasound techniques to measure plantar aponeurosis, a collagen-rich tissue, researchers demonstrated previously that people with diabetes have increased plantar fascia thickness (PFT) (5). Recently, this group reported that increased PFT predicted the development of microvascular complications in adolescents with type 1 diabetes and proposed abnormal PFT as a putative marker of soft-tissue glycation (6). The PON gene cluster maps to chromosome 7q21-22 and influences gene expression and serum activity. There is an established link between PON1 and macrovascular disease (7) and emerging evidence linking PON1 to microvascular complications (8,9). In this study we investigated whether the variants c.-107C>T at the promoter region and p.Leu54Met and p.Gln192Arg at the coding regions of PON1 are associated with PFT in type 1 diabetes.

stablished link between PON1 and macrovascular disease (7) and emerging evidence linking PON1 to microvascular complications (8,9). In this study we investigated whether the variants c.-107C>T at the promoter region and p.Leu54Met and p.Gln192Arg at the coding regions of PON1 are associated with PFT in type 1 diabetes. RESEARCH AND METHODS The cohort consisted of 331 Caucasian adolescents and young adults (162 male and 169 female) with a median (interquartile range) age of 15.4 (13.5–17.3) years and type 1 diabetes duration of 7.6 (4.9–10.6) years, who presented for routine complications assessment at the Children's Hospital at Westmead (Sydney, Australia) between 1998 and 2004. The study was approved by the hospital's ethics committee, and written informed consent was obtained. PFT measurement A single investigator (A.D.) measured PFT using ultrasound (Acuson 128 gray scale imager; Acuson, Mountain View, CA). For aponeurosis measurements, a linear-array, 7-MHz high-resolution transducer was placed longitudinally over the center of the arch at least 3 cm from the calcaneal insertion. To assess test-retest variability, four individuals were assessed 10 times in the same session. The assessor was masked to the measurements until they were completed and stored. The intraclass coefficient was 0.89. Abnormal PFT (>1.7 mm) was defined as 2 SDs above the mean measurement of 57 age-matched unrelated nondiabetic control subjects (27 male, median age 15.6 years) (5).

essed 10 times in the same session. The assessor was masked to the measurements until they were completed and stored. The intraclass coefficient was 0.89. Abnormal PFT (>1.7 mm) was defined as 2 SDs above the mean measurement of 57 age-matched unrelated nondiabetic control subjects (27 male, median age 15.6 years) (5). PON1 genotyping DNA was extracted from peripheral white blood cells collected in lithium-heparin tubes using a modified salting out protocol. Polymorphisms were analyzed by PCR-restriction fragment–length polymorphism using a slightly modified version of the procedure described by Humbert et al. (10). DNA (100 ng) was denatured (94°C, 12 min) and then amplified for 35 cycles using PCR primers. Each cycle consisted of denaturation (94°C, 30 s), annealing (55°C, 30 s), and extension (72°C, 30 s) with a final extension of 5 min. For p.Leu54Met genotyping, the 171-bp PCR product was digested (37°C, 2 h) with Hsp92II (Promega, Madison, WI), and products were separated by PAGE (10%) and stained by ethidium bromide. The leucine allele corresponded to the presence of a nondigested fragment of 171 bp; the methionine allele corresponded to two digestion fragments of 127 and 44 bp. For p.Gln192Arg genotyping, the 99-bp PCR product was digested with DpnII, and products were separated by PAGE (10%). The arginine allele corresponded to the presence of three digestion fragments (40, 31, and 28 bp); the glutamine allele corresponded to two digestion fragments of 59 and 40 bp.

PON1 genotyping DNA was extracted from peripheral white blood cells collected in lithium-heparin tubes using a modified salting out protocol. Polymorphisms were analyzed by PCR-restriction fragment–length polymorphism using a slightly modified version of the procedure described by Humbert et al. (10). DNA (100 ng) was denatured (94°C, 12 min) and then amplified for 35 cycles using PCR primers. Each cycle consisted of denaturation (94°C, 30 s), annealing (55°C, 30 s), and extension (72°C, 30 s) with a final extension of 5 min. For p.Leu54Met genotyping, the 171-bp PCR product was digested (37°C, 2 h) with Hsp92II (Promega, Madison, WI), and products were separated by PAGE (10%) and stained by ethidium bromide. The leucine allele corresponded to the presence of a nondigested fragment of 171 bp; the methionine allele corresponded to two digestion fragments of 127 and 44 bp. For p.Gln192Arg genotyping, the 99-bp PCR product was digested with DpnII, and products were separated by PAGE (10%). The arginine allele corresponded to the presence of three digestion fragments (40, 31, and 28 bp); the glutamine allele corresponded to two digestion fragments of 59 and 40 bp. For c.-107C>T genotyping, the 240-bp PCR product was digested with BsrBI, and products were separated by PAGE (10%). The TT genotype corresponded to a nondigested 240-bp fragment, the CC genotype corresponded to two digestion fragments of 212 and 28 bp, and the CT genotype corresponded to three digestion fragments of 240, 212, and 28 bp.

yping, the 240-bp PCR product was digested with BsrBI, and products were separated by PAGE (10%). The TT genotype corresponded to a nondigested 240-bp fragment, the CC genotype corresponded to two digestion fragments of 212 and 28 bp, and the CT genotype corresponded to three digestion fragments of 240, 212, and 28 bp. PON1 activity In a representative subgroup of subjects (n = 144), PON1 activity of lithium-heparin plasma was measured by the rates of hydrolysis of paraoxon and phenylacetate as described previously (9). Other variables A1C was measured by a Bio-Rad Diamat analyzer (nondiabetic range of 4–6%; Bio-Rad Laboratories, Hercules, CA). Nonfasting plasma cholesterol was measured by Cobas Mira. Systolic blood pressure (SBP), diastolic blood pressure (DBP), and BMI percentiles were determined using age- and sex-related reference standards. Statistical analysis Descriptive statistics are reported as means ± SD or median (interquartile range) for skewed distributions. Groups were compared by χ2 test for categorical variables. Differences between independent samples were evaluated using Student's t test and one-way ANOVA for normally distributed data or the Mann-Whitney U test for skewed data.

istics are reported as means ± SD or median (interquartile range) for skewed distributions. Groups were compared by χ2 test for categorical variables. Differences between independent samples were evaluated using Student's t test and one-way ANOVA for normally distributed data or the Mann-Whitney U test for skewed data. Multiple logistic regression was used to evaluate the association between abnormal PFT and biological and genetic variables. Explanatory variables were sex, age, diabetes duration, A1C, BMI, SBP, cholesterol, and PON1 genotypes (MM vs. ML/LL for p.Leu54Met; QQ vs. RQ/RR for p.Gln192Arg, and TT vs. CT/CC for c.-107C>T). Results are expressed as odds ratios and 95% CI. P ≤ 0.05 indicated statistical significance. RESULTS Clinical characteristics are summarized in Table 1. Patients with abnormal PFT had higher SBP and BMI percentiles and were more likely to be male. PON1 activity was not significantly different between those with or without abnormal PFT for any substrate. Table 2 shows genotype distributions and allele frequencies (all mutations were in Hardy-Weinberg equilibrium). Activities for phenylacetate and paraoxon substrates were higher for patients carrying LL alleles at PON1 p.Leu54Met versus those with the MM genotype. Similarly, those with CC alleles at c.-107C>T had higher enzyme activity for both substrates versus those with the TT genotype. At p.Gln192Arg, enzyme activity for hydrolysis of paraoxon was higher in those with the RR versus the QQ genotype, but phenylacetate activities did not differ.

rsus those with the MM genotype. Similarly, those with CC alleles at c.-107C>T had higher enzyme activity for both substrates versus those with the TT genotype. At p.Gln192Arg, enzyme activity for hydrolysis of paraoxon was higher in those with the RR versus the QQ genotype, but phenylacetate activities did not differ. The p.Leu54Met genotype distribution differed between the two PFT groups. The MM genotype was less frequent among those with thickened aponeurosis (normal PFT 20% vs. abnormal PFT 8%, χ2, P = 0.01). There was no difference in frequency of p.Gln192Arg or c.-107C>T genotype distribution by PFT (Table 3). In the multiple regression, LL/ML at p.Leu54Met, male sex, and BMI and SBP percentiles were associated with increased PFT (Table 4). There was no interaction between sex and BMI, sex and PON1 variants, or BMI and PON1 variants. There was strong linkage disequilibrium between p.Leu54Met and p.Gln192Arg and between p.Leu54Met and c.-107C>T genotypes (not shown). CONCLUSIONS This is the first report demonstrating an association between PFT, a marker of tissue glycation and/or oxidation, and PON1 gene polymorphisms in type 1 diabetes. In this cross-sectional study, the L allele at the p.Leu54Met gene was associated with abnormal PFT, or, conversely, the M allele had a protective effect. These results support the association between PON1 gene polymorphisms and plantar fascia changes. Previously, our group showed that the LL genotype was closely associated with microvascular complications (9,11).

the p.Leu54Met gene was associated with abnormal PFT, or, conversely, the M allele had a protective effect. These results support the association between PON1 gene polymorphisms and plantar fascia changes. Previously, our group showed that the LL genotype was closely associated with microvascular complications (9,11). After a relatively short median diabetes duration of 7 years, approximately half of the cohort had abnormal plantar fascia measurements. The early appearance of abnormal PFT, although unrelated to A1C measured at the time of the assessment, does not exclude the possibility that early glycemic variability could leave an imprinting in target organs, including changes in collagen, predisposing these organs to the future development of complications. In support of this possibility, we have recently demonstrated that PFT predicts retinopathy, early elevation of albumin excretion rate, and nerve abnormalities in young people with type 1 diabetes (6).

in target organs, including changes in collagen, predisposing these organs to the future development of complications. In support of this possibility, we have recently demonstrated that PFT predicts retinopathy, early elevation of albumin excretion rate, and nerve abnormalities in young people with type 1 diabetes (6). Oxidative stress and abnormalities of lipoprotein quantity and quality are implicated in the pathogenesis of microvascular complications (3). We hypothesize that genetic variants at the PON1 gene, for an antioxidant enzyme, could also predispose to collagen abnormalities via advanced glycation end product (AGE) formation, which involves glycation and oxidation. Tissue glycation and/or oxidation via intra- and extracellular generation of AGEs may promote diabetes complications. The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) demonstrated in skin biopsies that collagen AGEs predict microvascular disease (12). Skin collagen methionine sulfoxide, a marker of oxidative damage independent of glycemia, has also been associated with type 1 diabetes complications (13).

ions Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) demonstrated in skin biopsies that collagen AGEs predict microvascular disease (12). Skin collagen methionine sulfoxide, a marker of oxidative damage independent of glycemia, has also been associated with type 1 diabetes complications (13). There are divergent reports of PON1 activity and genotype and their relationship to microvascular complications. Our group described a higher allelic frequency of leucine 54 among type 1 diabetic subjects with versus without retinopathy (14). Similar findings were subsequently confirmed in a larger cohort of 372 type 1 diabetic adolescents in whom the LL genotype increased risk for early retinopathy by almost threefold (8). The presence of the LL genotype and two other PON1 polymorphisms was found to influence urinary albumin excretion in 156 Caucasian type 1 diabetic adolescents (9). In contrast, others found no correlation between microvascular complications and PON1 polymorphisms in adults with type 1 diabetes (15). Divergent results may be explained by differences in ethnicity, age, sample size, smoking, diet, concomitant medications (16), PON1 assays, and differing relationships between PON activity or genotype for initiation and progression of complications. Conversely, there is strong evidence showing that “high-expressor” PON1 alleles influence macrovascular disease. Homozygosity for the L allele doubled cardiovascular disease risk after adjustment for other risk factors in diabetes (17).

There are divergent reports of PON1 activity and genotype and their relationship to microvascular complications. Our group described a higher allelic frequency of leucine 54 among type 1 diabetic subjects with versus without retinopathy (14). Similar findings were subsequently confirmed in a larger cohort of 372 type 1 diabetic adolescents in whom the LL genotype increased risk for early retinopathy by almost threefold (8). The presence of the LL genotype and two other PON1 polymorphisms was found to influence urinary albumin excretion in 156 Caucasian type 1 diabetic adolescents (9). In contrast, others found no correlation between microvascular complications and PON1 polymorphisms in adults with type 1 diabetes (15). Divergent results may be explained by differences in ethnicity, age, sample size, smoking, diet, concomitant medications (16), PON1 assays, and differing relationships between PON activity or genotype for initiation and progression of complications. Conversely, there is strong evidence showing that “high-expressor” PON1 alleles influence macrovascular disease. Homozygosity for the L allele doubled cardiovascular disease risk after adjustment for other risk factors in diabetes (17). If PON1 protects lipids from oxidation, one might expect a higher expressor allele and a higher enzymatic activity to be protective against oxidative stress, AGE formation, and diabetes complications. However, the artificial nature of the substrates in this study is recognized and may not represent physiological substrates. A recent report demonstrated that the primary activity of the paraoxonases is that of a lactonase (18). As yet, there are no studies relating lactonase activity to diabetes complications.

ations. However, the artificial nature of the substrates in this study is recognized and may not represent physiological substrates. A recent report demonstrated that the primary activity of the paraoxonases is that of a lactonase (18). As yet, there are no studies relating lactonase activity to diabetes complications. Although PON1 activity was measured in a subgroup of subjects, PON1 polymorphisms influenced at least one of the two measured activities, in agreement with published data (19). PON1 activity was higher in the abnormal PFT group but did not achieve statistical significance. There is a large interindividual variation in PON activity that mostly depends on functional and promoter variations at the PON gene. Lower PON activity has been reported in type 1 diabetic subjects versus control subjects regardless of differences in PON1 phenotype (20). Hyperglycemia may influence the in vivo concentration and in vitro activity of PON1. Kordonouri et al. (11) showed that higher PON1 activity was associated with high glucose levels but not with A1C. With adjustment for blood glucose and diabetes duration, PON1 activity was higher in subjects with different stages of retinopathy versus those without retinopathy.

centration and in vitro activity of PON1. Kordonouri et al. (11) showed that higher PON1 activity was associated with high glucose levels but not with A1C. With adjustment for blood glucose and diabetes duration, PON1 activity was higher in subjects with different stages of retinopathy versus those without retinopathy. Our findings suggest that high-expressor alleles and a higher enzymatic activity are associated with increased susceptibility to microvascular complications (8,9,11). One explanation could be that although paraoxonase hydrolyzes harmful lysolipids produced by peroxidation, it might also increase their production from phospholipid peroxidation products and create a more harmful lipoprotein profile. The LL genotype has been shown to be least protective against oxidation (21). Another possibility is that relatively reduced enzyme activity rather than increased absolute PON1 activity promotes vascular complications. Some have reported that a higher LDL cholesterol–to-PON concentration ratio may indicate a reduced capacity of the enzyme to limit LDL oxidation (22). The low-activity allele at the PON1 gene has also been associated with a less harmful lipoprotein profile (23). Data on the LDL cholesterol–to-PON concentration ratio or levels of LDL cholesterol, triglycerides, and apolipoprotein B were not available for this study.

capacity of the enzyme to limit LDL oxidation (22). The low-activity allele at the PON1 gene has also been associated with a less harmful lipoprotein profile (23). Data on the LDL cholesterol–to-PON concentration ratio or levels of LDL cholesterol, triglycerides, and apolipoprotein B were not available for this study. Male sex presented a threefold increased risk for thickened aponeurosis independent of other risk factors. The reasons for this are unclear, but men have a greater risk of lower limb diabetes complications (24). Hormonal variation and different recreational activities may be contributors. Although level of physical activity was not measured, it is unlikely that plantar fasciitis, a common condition among athletes, is the cause of fascia thickness. Fasciitis is associated with heel pain, and its ultrasonographic changes are limited to the proximal insertion of the aponeurosis (25). Higher BMI and SBP were associated with abnormal PFT. Although BMI may increase mechanical load on the plantar aponeurosis, overweight is associated with microvascular complications in diabetes, insulin resistance, hypertension, and an atherogenic profile. In summary, this study underlines the association between p.Leu54Met variants and PFT, implicating PON1 in the pathogenesis of diabetes-related collagen changes. Although relationships between PON1 genes and collagen abnormalities merit further investigation, these initial findings support the concept of PON1 genetic variants as a link predisposing to development of complications in type 1 diabetes.

PFT, implicating PON1 in the pathogenesis of diabetes-related collagen changes. Although relationships between PON1 genes and collagen abnormalities merit further investigation, these initial findings support the concept of PON1 genetic variants as a link predisposing to development of complications in type 1 diabetes. This study was supported by a Diabetes Australia Research Trust New Millennium Grant. We thank Connie Karschimkus and Chee Lee for PON1 activity measurements.

Acute hyperglycemia occurs in up to 50% of all ST-segment elevation myocardial infarctions, whereas patients with diabetes represent ∼25% of patients with ST-segment elevation myocardial infarctions (1). When glucose tolerance testing is performed, 65% of patients with myocardial infarction and a negative history of diabetes can be diagnosed with diabetes or impaired glucose tolerance (2). Acute hyperglycemia on admission has been reported to worsen the prognosis in myocardial infarction patients with and without known diabetes (3), including increased risk of in-hospital mortality in both groups (4). Cardiovascular stress induces release of catecholamines, cortisol, and glucagons, leading to increases in glucose and free fatty acids that enhance hepatic gluconeogenesis and diminish peripheral glucose uptake. Unfavorable effects of high blood glucose levels in myocardial infarction involve impaired left ventricular function, increased incidence of the no-reflow phenomenon, and a tendency for arrhythmias (5). Several mechanisms implicated in the detrimental impact of hyperglycemia during acute myocardial ischemia have been postulated, i.e., enhanced oxidative stress, the activation of blood coagulation and platelets, stimulation of inflammation, and endothelial cell dysfunction (5). All of these have also been reported in type 2 diabetes (6,7).

plicated in the detrimental impact of hyperglycemia during acute myocardial ischemia have been postulated, i.e., enhanced oxidative stress, the activation of blood coagulation and platelets, stimulation of inflammation, and endothelial cell dysfunction (5). All of these have also been reported in type 2 diabetes (6,7). Evidence for the prothrombotic effects of acute hyperglycemia in vivo is scanty. Exposure to 24-h selective hyperglycemia in healthy volunteers results in increased tissue factor procoagulant activity (8). Acute hyperglycemia activates platelet aggregation, enhances thrombin generation, and activates coagulation factor VII (9). It is not known whether acute hyperglycemia during myocardial infarction is potent enough to influence hemostasis. Moreover, hyperglycemia, both in diabetic patients and under in vitro conditions, is linked to unfavorably altered fibrin clot properties and reduced fibrinolysis compared with the results at normoglycemia (10,11). Recently, we have showed that in patients with acute myocardial infarction, a history of type 2 diabetes is associated with impaired plasma clot permeability and fibrinolysis (12). The effect of hyperglycemia on clot properties in acute myocardial infarction patients with no history of diabetes has not been investigated yet. The aim of the study was to evaluate potential prothrombotic alterations in acute myocardial infarction patients in relation to hyperglycemia, including thrombin formation, platelet activation, and fibrin network structure/function.

ocardial infarction patients with no history of diabetes has not been investigated yet. The aim of the study was to evaluate potential prothrombotic alterations in acute myocardial infarction patients in relation to hyperglycemia, including thrombin formation, platelet activation, and fibrin network structure/function. RESEARCH DESIGN AND METHODS Patients with acute myocardial infarction admitted to the coronary care unit within the first 12 h after the onset of chest pain were enrolled in the study. We recruited 20 consecutive acute myocardial infarction patients with a history of type 2 diabetes, who self-reported taking insulin or oral hypoglycemic drugs on a regular basis (the DM group) and 20 patients with a negative history of diabetes, who had a serum glucose level of ≥7 mmol/l on admission (the HG group). Twenty patients with glucose levels <7 mmol/l (the NG group) served as a reference group. Inclusion criteria were typical chest pain and elevated cardiac troponin levels. Changes in electrocardiogram (ECG) recordings such as either ST-segment elevation ≥0.1 mV or ST-segment depression ≥0.1 mV in at least two contiguous ECG leads or normal ECG results were allowed. Exclusion criteria were as follows: cardiogenic shock, any acute illness, cancer, hepatic or renal dysfunction, a history of venous thromboembolism or stroke, anticoagulant therapy, and recent myocardial infarction (within the previous 3 months). All subjects received 300 mg aspirin 2–8 h before the study. Major adverse coronary events were recorded within the first 30 days after enrollment.

cancer, hepatic or renal dysfunction, a history of venous thromboembolism or stroke, anticoagulant therapy, and recent myocardial infarction (within the previous 3 months). All subjects received 300 mg aspirin 2–8 h before the study. Major adverse coronary events were recorded within the first 30 days after enrollment. All subjects enrolled in the study provided written, informed consent. The University ethics committee approved the study.

cancer, hepatic or renal dysfunction, a history of venous thromboembolism or stroke, anticoagulant therapy, and recent myocardial infarction (within the previous 3 months). All subjects received 300 mg aspirin 2–8 h before the study. Major adverse coronary events were recorded within the first 30 days after enrollment. All subjects enrolled in the study provided written, informed consent. The University ethics committee approved the study. Laboratory investigations Blood samples were obtained from an antecubital vein using a 21-gauge butterfly needle within 15 min upon admission. The lipid profile, C-reactive protein (CRP), glucose, creatinine, platelet count, and cardiac troponin T were determined using routine laboratory methods. A1C was analyzed by high-performance liquid chromatography using a Variant II analyzer (Bio-Rad, Hercules, CA). A human-specific radioimmunoassay kit (Linco Research, St. Charles, MO) was used to measure plasma insulin levels. Fibrinogen was determined using the Clauss method. High-sensitivity CRP was measured by latex nephelometry (Dade Behring, Marburg, Germany). Blood samples for thrombin and platelet markers were centrifuged at 2,500g for 15 min, and plasma was stored at −80°C. Using commercially available enzyme-linked immunosorbent assays, we determined the following in plasma: interleukin-6 (IL-6) (R&D Systems, Abingdon, U.K.); thrombin-antithrombin complexes (TATs) and prothrombin 1.2 fragments (F1.2), markers of thrombin formation (Enzygnost, Dade Behring); and soluble CD40 ligand (sCD40L), a marker of platelet activation (R&D Systems). Routine laboratory data and hemostatic variables were also obtained after 30 days from the event.

U.K.); thrombin-antithrombin complexes (TATs) and prothrombin 1.2 fragments (F1.2), markers of thrombin formation (Enzygnost, Dade Behring); and soluble CD40 ligand (sCD40L), a marker of platelet activation (R&D Systems). Routine laboratory data and hemostatic variables were also obtained after 30 days from the event. Model of vascular injury Measurements were performed in blood collected at 60-s intervals from a standardized skin incision, made using a Simplate IR device (Organon Teknika, Durham, NC) at the inflation of the sphygmomanometer cuff at 40 mmHg, as described previously (13–15). Blood was collected by means of heparinized tubes (Kabe Labortechnik, Numbrecht-Elsenroth, Germany) into Eppendorf tubes containing anticoagulants as described previously (14,15). After centrifugation at 3,000g at 4°C for 20 min, supernatants were frozen at −80°C. Both TAT (Dade Behring) and sCD40L (R&D Systems) were measured in the samples. Interassay and intra-assay coefficients of variation (CVs) were 5–7%. Thrombin formation and platelet activation were described as maximum velocity of both processes and total amounts of each marker produced within the first 6 min of bleeding (using the trapezoid rule) (14,15).

&D Systems) were measured in the samples. Interassay and intra-assay coefficients of variation (CVs) were 5–7%. Thrombin formation and platelet activation were described as maximum velocity of both processes and total amounts of each marker produced within the first 6 min of bleeding (using the trapezoid rule) (14,15). Clot permeability Permeation properties of fibrin clots were assessed according to the method of Mills et al. (16). Briefly, tubes containing plasma clots formed upon addition of calcium chloride and human thrombin (Sigma) were connected via plastic tubing to a reservoir of 0.05 mol/l Tris-HCl, and its volume flowing through the gels was measured within 60 min. A permeation coefficient (Ks), which indicates the pore size, was calculated from the equation, as described (16). The interassay CV was 9.2%. Plasma clot lysis assay To determine lysis time, we used an assay by Lisman et al. (17) with some modifications. Briefly, citrated plasma was mixed (1:1) with HEPES buffer containing calcium chloride, diluted recombinant tissue factor (Innovin, Dade Behring), phospholipid vesicles, and recombinant tissue plasminogen activator (Boehringer Ingelheim). The turbidity of this mixture (100 μl) was measured at 405 nm at 37°C in a SpectraMax 340 kinetic microplate reader (Molecular Devices). Clot lysis time was defined as the time from the midpoint of the baseline to maximum turbid transition, to the final plateau phase. The interassay and intra-assay CVs were 8.1 and 6.2%, respectively.

this mixture (100 μl) was measured at 405 nm at 37°C in a SpectraMax 340 kinetic microplate reader (Molecular Devices). Clot lysis time was defined as the time from the midpoint of the baseline to maximum turbid transition, to the final plateau phase. The interassay and intra-assay CVs were 8.1 and 6.2%, respectively. Statistical analysis The study was powered to have an 80% chance of detecting a 10% intergroup difference in maximum rate of TAT generation at the site of microvascular injury using a P value of 0.05, based on mean values in published articles (13–15). To demonstrate such a difference or greater, 12 patients were required in each group. The corresponding number of patients for local sCD40L release was calculated to be 12. Continuous data are presented as means ± SD or as median (interquartile range). The Kolmogorov-Smirnov test was used to determine normal distribution. The significance of between-group differences was tested by ANOVA with Scheffe's adjustment. Post hoc comparisons were made using a Tukey test. The χ2 test or Fisher's exact test was used to compare categorical variables. Pearson's correlations were used to identify associations between variables. A two-sided P value < 0.05 was considered statistically significant.

as tested by ANOVA with Scheffe's adjustment. Post hoc comparisons were made using a Tukey test. The χ2 test or Fisher's exact test was used to compare categorical variables. Pearson's correlations were used to identify associations between variables. A two-sided P value < 0.05 was considered statistically significant. RESULTS The three myocardial infarction groups did not differ with regard to demographic and clinical variables (Table 1). All three patient groups were enrolled after 5.2 ± 0.3 h of chest pain onset (P = 0.9). Patients with diabetes were treated either with insulin (n = 8; 40%) or with oral hypoglycemic agents (n = 12; 60%). Duration of the disease ranged from 0.5 to 11 (median 5) years. As expected, glucose levels were higher in both hyperglycemic groups and in patients with normoglycemia, whereas serum insulin and A1C were elevated in the DM group, with no difference between the HG and NG groups (Table 1). Higher cardiac troponin T was observed in the DM group than in the HG group (Table 1). In contrast to CRP, IL-6 levels were elevated by 86% both in the DM and HG groups compared with the NG group. Fibrinogen levels were 29% higher in the DM group than in the NG group, with similar values in both hyperglycemic groups (Table 1). Bleeding time did not differ among the three groups (Table 1). The total volume of blood collected from wounds was similar in all groups (data not shown).

RESULTS The three myocardial infarction groups did not differ with regard to demographic and clinical variables (Table 1). All three patient groups were enrolled after 5.2 ± 0.3 h of chest pain onset (P = 0.9). Patients with diabetes were treated either with insulin (n = 8; 40%) or with oral hypoglycemic agents (n = 12; 60%). Duration of the disease ranged from 0.5 to 11 (median 5) years. As expected, glucose levels were higher in both hyperglycemic groups and in patients with normoglycemia, whereas serum insulin and A1C were elevated in the DM group, with no difference between the HG and NG groups (Table 1). Higher cardiac troponin T was observed in the DM group than in the HG group (Table 1). In contrast to CRP, IL-6 levels were elevated by 86% both in the DM and HG groups compared with the NG group. Fibrinogen levels were 29% higher in the DM group than in the NG group, with similar values in both hyperglycemic groups (Table 1). Bleeding time did not differ among the three groups (Table 1). The total volume of blood collected from wounds was similar in all groups (data not shown). Thrombin formation Plasma TAT and F1.2 concentrations did not differ between the DM and HG groups. However, diabetic patients with acute myocardial infarction, but not those from the HG group, had higher plasma levels of F1.2 (by 27.5%) and TATs (by 30%) than those observed in the NG group (Table 1).

Bleeding time did not differ among the three groups (Table 1). The total volume of blood collected from wounds was similar in all groups (data not shown). Thrombin formation Plasma TAT and F1.2 concentrations did not differ between the DM and HG groups. However, diabetic patients with acute myocardial infarction, but not those from the HG group, had higher plasma levels of F1.2 (by 27.5%) and TATs (by 30%) than those observed in the NG group (Table 1). Time courses of TAT generation at the site of injury were similar regardless of the presence or absence of hyperglycemia (Fig. 1A). Maximum TAT levels were found at 6 min, with the highest values in the DM group (112.6 ± 10.4 nmol/l) and the lowest in the NG group (89.7 ± 9.1 nmol/l; P = 0.006). There was no difference between maximum TAT levels in bleeding time blood in the HG (96.1 ± 5.9 nmol/l) and NG groups (P = 0.3). A peak rate of TAT formation after vascular injury was higher in hyperglycemia (0.36 ± 0.03 for the DM group and 0.3 ± 0.03 nmol/l/s for the HG group, respectively) compared with patients with normoglycemia (0.21 ± 0.03 nmol/l/s; P < 0.0001 for both comparisons). However, TAT was also generated faster in the DM group than in the HG group (P < 0.0001). Total amounts generated after injury within 6 min were increased by 24.3% in diabetic patients with acute myocardial infarction compared with amounts in those with elevated glucose levels without a history of diabetes (P < 0.0001) as well as by 55.4% compared with amounts in those with normoglycemia during acute myocardial infarction (P < 0.0001) (Fig. 2A of the online appendix [available at http://dx.doi.org/10.2337/dc08-0282]).

ial infarction compared with amounts in those with elevated glucose levels without a history of diabetes (P < 0.0001) as well as by 55.4% compared with amounts in those with normoglycemia during acute myocardial infarction (P < 0.0001) (Fig. 2A of the online appendix [available at http://dx.doi.org/10.2337/dc08-0282]). None of the variables describing TAT formation at the site of vascular injury showed associations with plasma TAT levels, glycemia, insulinemia, age, or other clinical or laboratory variables in the three groups studied. Total formation of TATs within the first 6 min was associated with triglycerides in the HG group, but not in the other two groups (r = 0.48; P = 0.03). The maximum rate of TAT generation and TAT levels tended to be higher in patients whose blood was drawn after a longer time from pain onset only in the DM group (r = 0.38; P = 0.1 for both). Other variables showed no correlation with time from pain onset (data not shown).

t in the other two groups (r = 0.48; P = 0.03). The maximum rate of TAT generation and TAT levels tended to be higher in patients whose blood was drawn after a longer time from pain onset only in the DM group (r = 0.38; P = 0.1 for both). Other variables showed no correlation with time from pain onset (data not shown). Platelet activation Plasma sCD40L levels were similar in the DM and HG groups. Compared with the normoglycemic patients, patients in both the DM and HG groups displayed higher plasma sCD40L levels by 120 and 82.5%, respectively (Table 1). Profiles of sCD40L release, reflected in its levels in blood obtained from bleeding time wounds, shared common kinetics in acute myocardial infarction patients, with the steepest increase in diabetic subjects (Fig. 1B). The highest local sCD40L value of 23.4 ± 2.6 ng/ml was observed in the DM group. A lower maximum sCD40L level of 20.3 ± 1.9 ng/ml (P < 0.001) was found in the HG group. Maximum rates of sCD40L release were higher in patients in the DM (0.087 ± 0.009 ng · ml−1 · s−1) and HG (0.086 ± 0.01 ng · ml−1 · s−1) groups than in individuals with normoglycemia (0.074 ± 0.012 ng · ml−1 · s−1; P < 0.001 for both comparisons). There was no difference in this variable between the DM and HG groups (P = 0.2). The velocity of the sCD40L increase in shed blood was increased in the HG group compared with that in the NG group (P = 0.011).

groups than in individuals with normoglycemia (0.074 ± 0.012 ng · ml−1 · s−1; P < 0.001 for both comparisons). There was no difference in this variable between the DM and HG groups (P = 0.2). The velocity of the sCD40L increase in shed blood was increased in the HG group compared with that in the NG group (P = 0.011). Total release of sCD40L within the first 6 min was similar in the DM and HG groups. Both of these groups were characterized by increased amounts of sCD40L measured after injury (by 28 and 16.3%, P < 0.001, respectively) compared with the NG group (Fig. 2B of the online appendix). In the DM group, the maximum rate of the sCD40L release showed no association with the duration of diabetes, insulin administration, age, or other clinical or laboratory variables with two exceptions. It was correlated with glucose (r = 0.56; P = 0.01) and with plasma TAT levels (r = 0.53; P = 0.02). No similar associations were observed in the two other groups. Total release of sCD40L within the first 6 min was associated with total cholesterol (r = 0.47; P = 0.036) and plasma sCD40L levels (r = 0.48; P = 0.03) but only in the HG group. Variables describing local sCD40L release showed no significant correlations with time from pain onset (data not shown).

he two other groups. Total release of sCD40L within the first 6 min was associated with total cholesterol (r = 0.47; P = 0.036) and plasma sCD40L levels (r = 0.48; P = 0.03) but only in the HG group. Variables describing local sCD40L release showed no significant correlations with time from pain onset (data not shown). Clot permeability Lower clot permeability was found in patients with a prior history of diabetes compared with subjects from both the HG and NG groups (Table 2). However, Ks was similar in the HG and NG groups. Ks was correlated with fibrinogen in all groups (r from −0.36 to −0.51; P < 0.05). Ks was inversely associated with CRP only in the DM group (r = −0.42, P = 0.03), but showed no associations with lipids or thrombin or platelet parameters in venous or bleeding time blood in either group. Fibrinolysis Clot lysis time was the longest in the diabetic patients admitted for acute myocardial infarction and was significantly shorter in the HG group than in subjects with normoglycemia (Table 2). Lysis time showed correlations only with CRP in all three groups (r from 0.35 to 0.49; P < 0.05). No associations between lysis time and glucose or insulin levels were observed in any of the groups. There were no correlations of lysis time with thrombin generation or platelet activation in any of the patients and in the three groups or with time from the onset of myocardial infarction symptoms or troponin levels (data not shown).

between lysis time and glucose or insulin levels were observed in any of the groups. There were no correlations of lysis time with thrombin generation or platelet activation in any of the patients and in the three groups or with time from the onset of myocardial infarction symptoms or troponin levels (data not shown). Short-term outcomes During a 30-day follow-up, there were three cardiovascular deaths (two in the DM group and one in the NG group). Recurrent myocardial ischemia was observed in six patients, two in each group. No intergroup differences in major adverse cardiovascular events were observed. Glucose levels determined 1 month after enrollment revealed that all normoglycemic subjects had still normoglycemia, whereas three subjects from the HG group had glycemia >7 mmol/l; exclusion of these patients did not alter the results for hemostatic variables (data not shown).

erse cardiovascular events were observed. Glucose levels determined 1 month after enrollment revealed that all normoglycemic subjects had still normoglycemia, whereas three subjects from the HG group had glycemia >7 mmol/l; exclusion of these patients did not alter the results for hemostatic variables (data not shown). CONCLUSIONS The current study shows that elevated glucose levels are associated with significantly augmented thrombin formation and platelet protein secretion in response to vascular injury not only in patients with type 2 diabetes but also in those with no prior history of diabetes and hyperglycemia during acute myocardial infarction. Moreover, we demonstrated that hyperglycemia observed in acute myocardial infarction results in hypofibrinolysis, regardless of a history of type 2 diabetes, whereas reduced clot permeability was found only in patients with previously diagnosed diabetes compared with normoglycemic individuals. Our findings indicate that not only diabetes but also hyperglycemia occurring in acute myocardial infarction patients with no prior diagnosis of diabetes produces several prothrombotic effects that may contribute to an increased risk for thrombotic complications after an acute coronary event. The impact of hyperglycemia in myocardial infarction patients appeared potent enough to be detected despite strong prothrombotic effects of coronary plaque injury during myocardial infarction. Our findings may also help explain a recent observation that glucose-insulin-potassium therapy, resulting in increased glucose levels, could be harmful within the first days of acute myocardial infarction (18).

h to be detected despite strong prothrombotic effects of coronary plaque injury during myocardial infarction. Our findings may also help explain a recent observation that glucose-insulin-potassium therapy, resulting in increased glucose levels, could be harmful within the first days of acute myocardial infarction (18). Because efficient hemostasis occurs only at vascular lesions where tissue factor is exposed and platelets rapidly aggregate, measurements of hemostatic markers at the site of vascular injury are more sensitive than those in venous blood in the assessment of local thrombotic reactions (13,14,19). We did not observe elevated levels of thrombin or platelet markers in venous blood in diabetic patients compared with those from the HG group; the differences were detectable at the site of injury. Probable mechanisms for this effect of hyperglycemia involve enhanced activation of proinflammatory transcription factors that can increase tissue factor expression (20). Augmented local thrombin production in myocardial infarction patients with glucose >7.0 mmol/l was accompanied by increased platelet activation, reflected by elevated sCD40L levels in venous plasma and bleeding time blood. Of several soluble platelet activation markers, including β-thromboglobulin or P-selectin, sCD40L has been extensively studied in hyperglycemic subjects (8,9,21) and measured at the site of injury (19,22); ∼95% of circulating sCD40L is platelet-derived (11,23). For these reasons, sCD40L was chosen as the platelet activation marker in the current study. Importantly, a similar increase in sCD40L release correlated with thrombin formation has been reported in patients with the metabolic syndrome (24).

ite of injury (19,22); ∼95% of circulating sCD40L is platelet-derived (11,23). For these reasons, sCD40L was chosen as the platelet activation marker in the current study. Importantly, a similar increase in sCD40L release correlated with thrombin formation has been reported in patients with the metabolic syndrome (24). Fibrin clot analysis revealed reduced lysis time in the DM and HG groups compared with that in subjects with glycemia <7 mmol/l, without any intergroup differences in clot permeability except for significantly higher permeability in diabetic subjects. Glycation of the fibrinogen molecules is largely responsible for altered fibrin clot features found at elevated glucose levels (10,11). We extended previous observations by showing a potent impact of diabetes on fibrin properties, easily detectable also in myocardial infarction patients despite the fact that acute myocardial ischemia itself is associated with deleterious clot alterations similar to those described in diabetic patients (12). A short-term increase in glucose levels does not modify fibrin structure, which explains the similar permeability observed in the HG and NG groups. Reduced lysis efficiency in the HG and DM groups indicates the presence of some glucose-mediated rapid mechanisms impairing fibrinolysis even if the extent of glycation is negligible. This effect could be explained by elevated plasminogen activator inhibitor 1 observed in hyperglycemia (5,6). It might be speculated that altered fibrin in hyperglycemia leads to lower binding affinity of both tissue plasminogen activator and plasminogen toward fibrin (11) and, as a consequence, impaired clot lysis in our assay.

t could be explained by elevated plasminogen activator inhibitor 1 observed in hyperglycemia (5,6). It might be speculated that altered fibrin in hyperglycemia leads to lower binding affinity of both tissue plasminogen activator and plasminogen toward fibrin (11) and, as a consequence, impaired clot lysis in our assay. One might suspect that insulin or oral hypoglycemic agents taken only by diabetic patients confounded the data interpretation. However, there is no evidence that in myocardial infarction patients such therapy alters thrombin formation or platelet activation. In terms of fibrin-modifying properties, insulin, gliclazide, and metformin have been shown to enhance clot lysis (25). We might speculate that susceptibility to lysis is probably even weaker in untreated diabetic patients with myocardial infarction. Another potential effect could be mediated by statins that were taken by a significantly lower percentage of the HG group before myocardial infarction. Because statins can reduce thrombin generation (13) and platelet activation (20) after injury in stable patients, both processes may have been relatively more vigorous in the HG group than in the DM and NG groups. However, no data support the view that statins are potent enough to suppress the massive activation of hemostasis observed in patients with acute myocardial infarction (26).

tivation (20) after injury in stable patients, both processes may have been relatively more vigorous in the HG group than in the DM and NG groups. However, no data support the view that statins are potent enough to suppress the massive activation of hemostasis observed in patients with acute myocardial infarction (26). This study has limitations. First, the number of patients studied is limited. However, we matched the myocardial infarction patients with and without elevated glucose levels as well as those with normoglycemia well. Second, our analysis was based on a determination of each variable at a single time point. Third, results of oral glucose tests after myocardial infarction were not analyzed. However, lack of significant differences in A1C between the HG and NG groups speaks against the possibility that patients with undiagnosed diabetes before the acute event were enrolled in the HG group. Finally, statistical associations reported here do not necessarily mean cause-effect relationships. Further studies are needed to elucidate this issue. In summary, our findings demonstrate that acute hyperglycemia in acute myocardial infarction patients without a previous history of diabetes is associated with increased thrombin generation and platelet activation at the site of vascular injury as well as greater resistance to fibrinolysis. This study provides further insights into the relationship between hyperglycemia and thrombosis in myocardial infarction patients.

without a previous history of diabetes is associated with increased thrombin generation and platelet activation at the site of vascular injury as well as greater resistance to fibrinolysis. This study provides further insights into the relationship between hyperglycemia and thrombosis in myocardial infarction patients. Supplementary Material Online-Only Appendix This work was supported by a grant from the Polish Ministry of Science and Education (to A.U.).

Diabetic microangiopathy can involve alveolar tissue and capillaries, the largest microvascular bed in the body, leading to restriction of lung volume and alveolar gas transport, as manifested by reduced diffusing capacity of the lung for carbon monoxide (DLCO), as well as its components: membrane diffusing capacity and pulmonary capillary blood volume (VC). Lung diffusing capacity is the gas conductance across the lung, modeled as diffusion across alveolar-capillary membrane barrier followed by chemical binding to capillary hemoglobin. In young nonsmokers with poorly controlled type 1 diabetes, DLCO and its components were reduced 15–30% at rest and exercise compared with age-matched nondiabetic subjects (1,2). In type 1 diabetic patients who maintained near-normoglycemia, these parameters are near normal, suggesting a relationship between alveolar function and systemic microangiopathy. Impaired alveolar gas transfer in type 1 diabetes signifies erosion of microvascular reserves that could accelerate clinical decline in conjunction with primary lung disease, aging, or cardiorenal complications and affect long-term tolerance to the use of inhaled insulin.

alveolar function and systemic microangiopathy. Impaired alveolar gas transfer in type 1 diabetes signifies erosion of microvascular reserves that could accelerate clinical decline in conjunction with primary lung disease, aging, or cardiorenal complications and affect long-term tolerance to the use of inhaled insulin. Type 2 diabetes has also been linked to lower spirometric indexes (3,4) and resting DLCO (5,6). However, previous studies had not taken into account the dependence of DLCO on pulmonary blood flow (Q̇). Normally, DLCO and its components increase 40–60% in a linear relationship as Q̇ increases up to peak exercise. The ability to augment DLCO and its components indexes the recruitment of alveolar microvascular reserves via enlarged membrane surfaces, as well as increased mass and improved distribution of alveolar-capillary erythrocytes. Recruitment is essential for maintaining a normal diffusion-to-perfusion (D/Q̇) ratio and achieving adequate oxygenation of end-capillary blood leaving the lung (7). Conventional interpretation of DLCO implicitly assumes an unchanged Q̇; this assumption is unwarranted and can be misleading. For example, lower cardiac output associated with diabetic heart disease decreases apparent DLCO even when alveolar diffusion is normal. Conversely, elevated cardiac output associated with obesity increases apparent DLCO and could mask the impairment of alveolar diffusion. Thus, the adequacy of alveolar-capillary recruitment and gas transfer cannot be optimally assessed without knowledge of both DLCO and Q̇. We hypothesized that restriction of lung volume, DLCO, and microvascular recruitment develops in type 2 diabetes independent of Q̇ and that abnormalities correlate with disease duration, glycemic control, and extrapulmonary microangiopathy and are compounded by obesity. We simultaneously measured lung volume, Q̇, DLCO, membrane diffusing capacity, and VC using a noninvasive rebreathing technique in type 2 diabetic patients from rest to heavy exercise. Measurements were compared with reference values obtained in nondiabetic control subjects and adjusted for Q̇. Alveolar-capillary recruitment was assessed from the slopes of the increase in DLCO, diffusing capacity of the lung for nitric oxide (DLNO) and VC with respect to Q̇.

etic patients from rest to heavy exercise. Measurements were compared with reference values obtained in nondiabetic control subjects and adjusted for Q̇. Alveolar-capillary recruitment was assessed from the slopes of the increase in DLCO, diffusing capacity of the lung for nitric oxide (DLNO) and VC with respect to Q̇. RESEARCH DESIGN AND METHODS The institutional review board approved all protocols; written informed consent was obtained from all subjects. Nonsmoking type 2 diabetic patients (n = 69) without overt cardiopulmonary disease were recruited from the University of Texas Southwestern Diabetes Treatment Center. Thirty-seven patients were treated with insulin; 25 were also taking an oral hypoglycemic agent. Thirty-one patients were taking oral agents only. Thirty subjects were taking antihypertensive medication, and 32 were taking antihyperlipidemia medication. Five subjects were remote smokers (∼10 pack-years); the average time since smoking cessation was 14 years. Forty-five healthy nondiabetic nonsmokers served as simultaneous control subjects. Adjusted reference values were derived from 75 cumulative nonobese (BMI <30 kg/m2) control subjects.

ntihyperlipidemia medication. Five subjects were remote smokers (∼10 pack-years); the average time since smoking cessation was 14 years. Forty-five healthy nondiabetic nonsmokers served as simultaneous control subjects. Adjusted reference values were derived from 75 cumulative nonobese (BMI <30 kg/m2) control subjects. Apparatus Standard spirometry was performed (Vmax229; Sensormedics, Yorba Linda, CA). Subjects exercised on a bicycle ergometer (Ergometrics-800; Sensormedics) while breathing through a respiratory valve (8500; Hans Rudolph, Kansas City, MO) and solenoid-controlled switching assembly (GH3315; Precision Dynamics, San Fernando, CA, and EV-3–12; Clippard, Cincinnati, OH). The expiratory circuit opened via a mixing chamber to either room air or a bag-in-a box reservoir containing the test gas mixture. Expired ventilation was measured using a turbine flowmeter (VMM2; Interface Associates, Aliso Viejo, CA). Oxygen and CO2 concentrations were measured by mass spectrometry (MGA-1100; PerkinElmer). Electrocardiogram and transcutaneous oxygen saturation (N-180; Nelcor, Carlsbad, CA) were monitored continuously.

s mixture. Expired ventilation was measured using a turbine flowmeter (VMM2; Interface Associates, Aliso Viejo, CA). Oxygen and CO2 concentrations were measured by mass spectrometry (MGA-1100; PerkinElmer). Electrocardiogram and transcutaneous oxygen saturation (N-180; Nelcor, Carlsbad, CA) were monitored continuously. Rebreathing technique The technique is well established (8,9). The bag-in-a box reservoir contained a mixture of 0.3% methane, 0.3% carbon monoxide (CO), 0.8% acetylene, and either 30 or 90% oxygen in a balance of nitrogen. When needed, nitric oxide (NO) (∼40 ppm) was added immediately before testing. At a selected end-expiration the valves switched electronically, allowing the subject to inspire one bolus of test gas to total lung capacity and then rebreathe this bolus in and out of an anesthetic bag for 12–16 s while gas concentrations at the mouth were monitored. Methane, acetylene, and CO concentrations were measured by an infrared analyzer (Sensors, Saline, MI); the NO concentration was measured by chemiluminescence (NOA280; Sievers Instruments, Boulder, CO).

rebreathe this bolus in and out of an anesthetic bag for 12–16 s while gas concentrations at the mouth were monitored. Methane, acetylene, and CO concentrations were measured by an infrared analyzer (Sensors, Saline, MI); the NO concentration was measured by chemiluminescence (NOA280; Sievers Instruments, Boulder, CO). Systemic microangiopathy Retinopathy was assessed by funduscopic examination. The presence of microaneurysm, hemorrhage, exudate, or neovascularization or previous laser treatment was considered positive. Microalbuminuria was assessed from a nonfasting urine sample: ≥30 μg/mg creatinine was considered abnormal. Nerve conduction was studied in the Electrodiagnostics Laboratory of University Diabetes Treatment Center. The ulnar sensory and peroneal motor nerves were stimulated, and the compound nerve or muscle action potential was recorded to assess conduction velocity, latency, and amplitude in comparison with established reference values (10). Individual nerves were abnormal if at least one of these parameters was outside the normal threshold. Neuropathy was conservatively defined as abnormalities in both motor and sensory nerves. Protocol On the first visit, medical history was reviewed and physical examination was performed. A venous blood sample was drawn to measure hematocrit, Hb, and A1C concentrations. A urine sample was collected, and nerve conduction was measured. Spirometry, maximal voluntary ventilation, and DLCO at rest were measured. Maximal oxygen uptake was determined by an incremental protocol (20–30 W every 3 min) until volitional termination.

ple was drawn to measure hematocrit, Hb, and A1C concentrations. A urine sample was collected, and nerve conduction was measured. Spirometry, maximal voluntary ventilation, and DLCO at rest were measured. Maximal oxygen uptake was determined by an incremental protocol (20–30 W every 3 min) until volitional termination. On a second visit, studies were performed at rest and at 30, 60, and 90% of the predetermined maximal workload, with each sustained for 3 min followed by the rebreathing maneuver. Duplicate measurements were performed with the test gas containing 30 or 90% oxygen in balanced order. Before rebreathing the test gas containing 90% oxygen, subjects prebreathed 100% oxygen for ∼30 s until alveolar oxygen tension (PAo2) reached ∼600 mmHg. Subjects rested between workloads until heart rate and ventilation returned to baseline. On a third visit, the exercise protocol was repeated but without NO in the test gas. The presence of NO in the test gas mixture does not alter the measurements during the brief (12–16 s) rebreathing period (11).

reached ∼600 mmHg. Subjects rested between workloads until heart rate and ventilation returned to baseline. On a third visit, the exercise protocol was repeated but without NO in the test gas. The presence of NO in the test gas mixture does not alter the measurements during the brief (12–16 s) rebreathing period (11). Data analysis The analysis was established previously (8,9). Lung volume (body temperature and pressure saturated, in liters) was estimated by methane dilution. Q̇, DLNO, and DLCO were determined from end-tidal disappearance of acetylene, NO, and CO, respectively. Conductance of membrane and hemoglobin binding contribute about equally to DLCO. DLNO was used as a direct index of membrane diffusing capacity. Because NO is rapidly scavenged by hemoglobin, resistance to alveolar NO uptake resides mainly within the tissue/erythrocyte membrane, and DLNO is directly related to diffusing capacity of alveolar membrane for carbon monoxide (DMCO) (DLNO = 2.42.DMCO) (9). From DLCO and the DMCO derived from DLNO, VC was calculated by the standard equation: \documentclass[10pt]{article} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in

Data analysis The analysis was established previously (8,9). Lung volume (body temperature and pressure saturated, in liters) was estimated by methane dilution. Q̇, DLNO, and DLCO were determined from end-tidal disappearance of acetylene, NO, and CO, respectively. Conductance of membrane and hemoglobin binding contribute about equally to DLCO. DLNO was used as a direct index of membrane diffusing capacity. Because NO is rapidly scavenged by hemoglobin, resistance to alveolar NO uptake resides mainly within the tissue/erythrocyte membrane, and DLNO is directly related to diffusing capacity of alveolar membrane for carbon monoxide (DMCO) (DLNO = 2.42.DMCO) (9). From DLCO and the DMCO derived from DLNO, VC was calculated by the standard equation: \documentclass[10pt]{article} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \begin{equation*}\frac{1}{{\mathrm{DL}}_{{\mathrm{CO}}}}=\frac{1}{{\mathrm{DM}}_{{\mathrm{CO}}}}+\frac{1}{{\Theta}_{{\mathrm{CO}}}{\cdot}{\mathrm{Vc}}}\end{equation*}\end{document} where CO uptake by 1 ml of whole blood (θCO) is dependent on mean PAo2 and the Hb concentration: \documentclass[10pt]{article} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in

\begin{document} \begin{equation*}\frac{1}{{\mathrm{DL}}_{{\mathrm{CO}}}}=\frac{1}{{\mathrm{DM}}_{{\mathrm{CO}}}}+\frac{1}{{\Theta}_{{\mathrm{CO}}}{\cdot}{\mathrm{Vc}}}\end{equation*}\end{document} where CO uptake by 1 ml of whole blood (θCO) is dependent on mean PAo2 and the Hb concentration: \documentclass[10pt]{article} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \begin{equation*}\frac{1}{{\Theta}_{CO}}=(0.73+0.0058.P_{A}O_{2}){\cdot}\frac{14.6}{[Hb]}\end{equation*}\end{document} DMCO and VC were used to express DLCO at a constant Hb concentration (14.6 g/dl) and PAo2 (120 mmHg).

\begin{document} \begin{equation*}\frac{1}{{\mathrm{DL}}_{{\mathrm{CO}}}}=\frac{1}{{\mathrm{DM}}_{{\mathrm{CO}}}}+\frac{1}{{\Theta}_{{\mathrm{CO}}}{\cdot}{\mathrm{Vc}}}\end{equation*}\end{document} where CO uptake by 1 ml of whole blood (θCO) is dependent on mean PAo2 and the Hb concentration: \documentclass[10pt]{article} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{pmc} \usepackage[Euler]{upgreek} \pagestyle{empty} \oddsidemargin -1.0in \begin{document} \begin{equation*}\frac{1}{{\Theta}_{CO}}=(0.73+0.0058.P_{A}O_{2}){\cdot}\frac{14.6}{[Hb]}\end{equation*}\end{document} DMCO and VC were used to express DLCO at a constant Hb concentration (14.6 g/dl) and PAo2 (120 mmHg). Duplicate measurements were averaged and expressed as absolute values and as percentages of reference values from nondiabetic control subjects. End-expiratory lung volume (EELV) and end-inspiratory lung volume (EILV) were adjusted for sex, age, and height (men: EELV = 5.72 × height + 0.02 × age − 7.24, EILV = 11.32 × height − 13.23; women: EELV = 3.45 × height + 0.02 × age − 3.84, EILV = 4.89 × height + 0.02 × age − 3.79). DLCO, DLNO, and VC were adjusted for sex, age, body surface area, and Q̇ using multivariate regression analysis (8,11). Individual DLCO, DLNO, and VC measurements were analyzed with respect to Q̇; slope of the linear regression provides an index of alveolar-capillary recruitment (7). Data were compared by ANOVA with a post hoc test by Fisher's protected least significant difference. Differences were significant at P ≤ 0.05.

ysis (8,11). Individual DLCO, DLNO, and VC measurements were analyzed with respect to Q̇; slope of the linear regression provides an index of alveolar-capillary recruitment (7). Data were compared by ANOVA with a post hoc test by Fisher's protected least significant difference. Differences were significant at P ≤ 0.05. RESULTS In type 2 diabetic patients, the prevalence of retinopathy was 32%, the prevalence of microalbuminuria was 38%, and the prevalence of nerve conduction defects was 28%. A1C exceeded 8.0% in 54% of patients; average A1C was slightly lower in obese (BMI >30 kg/m2) than in nonobese patients (Table 1). Hematological indexes were normal. Forced vital capacity (FVC) was significantly (8–11%) lower regardless of BMI. Forced expiratory volume in 1 s (FEV1), and maximal voluntary ventilation were normal. Peak heart rate exceeded 80% of the predicted maximum; peak workload and peak oxygen uptake were ∼25% below the predicted maximum. Ventilation and tidal volume at peak exercise were ∼20% lower in patients compared with control subjects.

ced expiratory volume in 1 s (FEV1), and maximal voluntary ventilation were normal. Peak heart rate exceeded 80% of the predicted maximum; peak workload and peak oxygen uptake were ∼25% below the predicted maximum. Ventilation and tidal volume at peak exercise were ∼20% lower in patients compared with control subjects. Mixing efficiency during rebreathing and transcutaneous oxygen saturation was normal in all subjects. Mean alveolar NO concentration (5–7 ppb) was similar among groups. In patients, EELV and EILV were ∼15% below normal regardless of BMI (Fig. 1A). At the highest sustained workload, Q̇ in patients was below normal (Table 2). Unadjusted DLCO, DLNO, and VC measured upon exercise were modestly lower in patients compared with control subjects (Table 2). When expressed as a percentage of reference values adjusted for Q̇, DLCO, DLNO, and VC were within the normal range in nonobese patients but remained significantly reduced (16–18%) in obese patients (Fig. 1B). The relationship between DLNO and DMCO was normal (not shown). The slopes of the linear increase in DLCO and DLNO with respect to Q̇ were similar among groups. The slope of the linear increase in VC with respect to Q̇ was 20–25% below normal in patients regardless of BMI (Table 2).

(16–18%) in obese patients (Fig. 1B). The relationship between DLNO and DMCO was normal (not shown). The slopes of the linear increase in DLCO and DLNO with respect to Q̇ were similar among groups. The slope of the linear increase in VC with respect to Q̇ was 20–25% below normal in patients regardless of BMI (Table 2). In male and female patients, A1C >8.0% correlated with significantly lower DLCO, DLNO, VC, and EILV (Fig. 2A), and microalbuminuria correlated with lower DLCO, DLNO, and EILV (Fig. 2B) compared with patients without these complications. In male but not female patients, the presence of neuropathy was associated with significantly lower DLCO, DLNO, VC, and EILV (Fig. 2C), whereas retinopathy correlated with a significantly lower DLNO (Fig. 2B). There was no significant correlation of lung function to age or to disease duration in either sex.

plications. In male but not female patients, the presence of neuropathy was associated with significantly lower DLCO, DLNO, VC, and EILV (Fig. 2C), whereas retinopathy correlated with a significantly lower DLNO (Fig. 2B). There was no significant correlation of lung function to age or to disease duration in either sex. CONCLUSIONS This is the first study to quantify pulmonary microvascular reserves in type 2 diabetes. The main findings were as follows. 1) Lung volume was moderately reduced regardless of sex or obesity. 2) Peak Q̇, DLCO, DLNO, and VC were reduced upon exercise. 3) Adjustment for sex, age, and Q̇ normalized DLCO, DLNO, and VC in nonobese type 2 diabetic patients, but the adjusted parameters remained reduced in obese patients. 4) The slope of the increase in VC with respect to Q̇ was reduced regardless of obesity, consistent with diminished recruitment of alveolar capillaries. These results highlight the need to consider Q̇ when interpreting DLCO and its components. 5) Alveolar microvascular indexes were significantly related to glycemic control and extrapulmonary microangiopathy in a sex-specific manner.

regardless of obesity, consistent with diminished recruitment of alveolar capillaries. These results highlight the need to consider Q̇ when interpreting DLCO and its components. 5) Alveolar microvascular indexes were significantly related to glycemic control and extrapulmonary microangiopathy in a sex-specific manner. Lung volume Hyperglycemia and insulin resistance are associated with lower FVC and FEV1 (4,12). A restrictive pattern in middle-aged nondiabetic adults is predictive of subsequent type 2 diabetes (13). Some studies do not show differences in adjusted rates of longitudinal change in spirometry between diabetic and nondiabetic subjects (4), whereas others found that declining FEV1 and lung volume are directly related to glycemic control and mortality (3). In type 1 diabetes, a lower lung volume is associated with abnormal elastic recoil (14) and elevated work of breathing at exercise (2). A stiff chest wall with limited joint mobility (15) may be caused by abnormal connective tissue metabolism as well as collagen cross-linking in thoracic and lung tissue. Autonomic neuropathy involving respiratory muscles may impair thoracic mobility. A similar pathogenesis may cause volume restriction in type 2 diabetes. In elderly men, adiposity and metabolic syndrome are associated with a restrictive spirometric pattern (16). Mechanical loading of the thorax due to adiposity could exacerbate lung volume restriction. Abnormal fat infiltration and connective tissue deposition (17) within the lung parenchyma may further reduce lung volume and compliance.

iposity and metabolic syndrome are associated with a restrictive spirometric pattern (16). Mechanical loading of the thorax due to adiposity could exacerbate lung volume restriction. Abnormal fat infiltration and connective tissue deposition (17) within the lung parenchyma may further reduce lung volume and compliance. Diffusion and alveolar-capillary recruitment Normally, lung volume and Q̇ are the major determinants of DLCO, DLNO, and VC (8,11). Upon exercise, DLCO, DLNO, and VC increase 40–60% in a linear relationship with respect to perfusion (7). In nonobese patients, the lower lung volume and lower Q̇ at exercise fully explain the 10–25% reduction in measured DLCO DLNO, and VC. Because lung volume was similarly reduced in obese and nonobese patients, the persistently lower DLCO DLNO, and VC in obese patients after adjustment for Q̇ suggest additional factors, e.g., infiltrative fat or connective tissue deposition within alveolar tissue, that cause diffusion impairment.

ured DLCO DLNO, and VC. Because lung volume was similarly reduced in obese and nonobese patients, the persistently lower DLCO DLNO, and VC in obese patients after adjustment for Q̇ suggest additional factors, e.g., infiltrative fat or connective tissue deposition within alveolar tissue, that cause diffusion impairment. One major effect of type 2 diabetes is decreased alveolar microvascular perfusion, leading to proportionately lower DLCO, DLNO, and VC at rest or exercise. Obesity further impairs DLCO, DLNO, and VC, but the effect is partially offset by an obesity-associated increase in cardiac output (18). These results highlight the need to consider perfusion when interpreting lung diffusion. The magnitude of diffusion impairment in type 2 diabetes is milder than that observed in type 1 diabetes (1,2); differences could relate to longer disease duration in our earlier type 1 diabetes study (>15 years) compared with that for type 2 diabetes (∼8 years) in this study. True disease duration is often uncertain in type 2 diabetes, and we did not observe a significant relationship between type 2 diabetes duration and lung function. Also, type 1 diabetes is uniformly severe, whereas the severity of type 2 diabetes is heterogeneous. Nonetheless, a consistent inverse relationship between lung function and glycemia emerged in type 2 diabetes as in type 1 diabetes (2).

erve a significant relationship between type 2 diabetes duration and lung function. Also, type 1 diabetes is uniformly severe, whereas the severity of type 2 diabetes is heterogeneous. Nonetheless, a consistent inverse relationship between lung function and glycemia emerged in type 2 diabetes as in type 1 diabetes (2). In moderate/localized lung disease, DLCO, DLNO, and VC are reduced at a given Q̇, but the ability to recruit the remaining alveolar microvasculature is preserved; recruitment mitigates the reduction in DLCO to maintain arterial oxygen saturation (9,19). In contrast, few lung units are recruitable in diffuse pulmonary fibrosis: DLCO and its components are not only reduced at rest but fail to rise as Q̇ increases (19); inadequate recruitment causes the diffusion-to-perfusion (D/Q̇) ratio to fall with exercise, leading to severe arterial hypoxemia (7). Thus, multivariate analysis of lung diffusion should include simultaneously measured Q̇ as a dynamic determinant. Impairment of alveolar-capillary recruitment in type 2 diabetes regardless of obesity suggests parenchymal changes that impede opening or distention of alveolar capillaries, possibly caused by connective tissue deposition within alveolar walls that has been observed in experimental diabetes (17); obesity may exaggerate these changes.

lveolar-capillary recruitment in type 2 diabetes regardless of obesity suggests parenchymal changes that impede opening or distention of alveolar capillaries, possibly caused by connective tissue deposition within alveolar walls that has been observed in experimental diabetes (17); obesity may exaggerate these changes. Relation to systemic microangiopathy Lung function in type 2 diabetes is worse in a sex-specific manner in the presence of extrapulmonary end-organ complications, suggesting that nonenzymatic protein glycation, which predicts long-term progression of retinopathy and nephropathy, also predisposes to lung restriction. Sex-specific susceptibility to diabetes complications is well known. For example, diabetic foot lesion has a poorer prognosis in men than in women (20). The DNA polymorphism that promotes angiotensinogen gene expression increases the risk of nephropathy in diabetic men but not women (21). The risk of cardiovascular disease is higher in diabetic women than in men (22). Diabetes-related oxidative stress and reduction in antioxidant activity is greater in women than in men (23). Lifestyle, genetics, sex hormones, vascular endothelial function, advanced glycation end products, and intrinsic sex differences in lung structure may influence sex susceptibility to complications.

in men (22). Diabetes-related oxidative stress and reduction in antioxidant activity is greater in women than in men (23). Lifestyle, genetics, sex hormones, vascular endothelial function, advanced glycation end products, and intrinsic sex differences in lung structure may influence sex susceptibility to complications. Clinical implications Unlike the smaller microvasculature in the retina, heart, or peripheral nervous system, alveolar microvasculature is extensive. The oxygen transport capacity of the lung is twice that of the cardiovascular system or skeletal muscle. In chronic lung disease, lung volume and DLCO could decline ∼50% without an individual incurring dyspnea at rest. Because of the large physiological reserves and because peak cardiac output is concurrently reduced, diabetic pulmonary dysfunction remains “subclinical.” Nonetheless, a modest loss of alveolar-capillary reserves can be quantified by noninvasive methods independent of physical fitness and correlates with glycemia as well as systemic microangiopathy. It remains to be determined whether alveolar microvascular indexes track longitudinal microangiopathy in a “clean” organ that is not ravaged by diabetes or its treatment. Loss of alveolar reserves could exaggerate aging-related functional decline (5) and predispose to overt sequelae in conjunction with renal and heart failure or primary lung disease. For example, diabetes significantly increases mortality in women with cystic fibrosis (24). Residence at high altitude where alveolar hypoxia imposes the primary limitation to oxygen transport is associated with higher prevalence of diabetic end-organ complications (25). These issues regarding physiological reserves are also important for the chronic use of inhaled insulin, which causes an early reduction in lung function (26). Finally, these data suggest that weight loss in obese type 2 diabetic patients could improve alveolar microvascular function.

etic end-organ complications (25). These issues regarding physiological reserves are also important for the chronic use of inhaled insulin, which causes an early reduction in lung function (26). Finally, these data suggest that weight loss in obese type 2 diabetic patients could improve alveolar microvascular function. This study is supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01 DK063242. We also acknowledge support of the General Clinical Research Center, M01 RR00633. We thank the staff of the University Diabetes Treatment Center for patient liaison and Brenda Brightman for performing nerve conduction studies.

Latent autoimmune diabetes in adults (LADA) consists of a subgroup (∼10%) of adult patients initially diagnosed with type 2 diabetes, who show signs of β-cell autoimmunity and eventually develop insulin requirement (1,2). Signs of β-cell autoimmunity, such as the well-characterized insulin autoantibodies, glutamate decarboxylase (GAD65), and the tyrosine phosphatase–like protein insulinoma-associated protein-2, indicate significant damage of the β-cells and subsequent development of insulin requirement in these patients (1). While autoantibodies to insulin and insulinoma-associated protein-2 antibody (Ab) are inversely correlated with age at onset, GAD65Ab shows no, and in some studies even a positive, correlation with age at onset and is therefore a particularly attractive marker for autoimmune diabetes in the adult population (3,4). Moreover, GAD65Ab can be detected years after the clinical onset of the disease, indicating that these autoantibodies may be permanent markers for the autoimmune response (5,6). Notably, not all LADA patients progress to insulin requirement, raising the possibility that the autoimmune response in these patients resembles that in autoantibody-positive healthy individuals, with no significant risk for development of insulin requirement (7,8). A better understanding of the autoimmune response is necessary to predict insulin requirement in LADA patients, which is important to prevent escalation of blood glucose levels and subsequent complications.

autoantibody-positive healthy individuals, with no significant risk for development of insulin requirement (7,8). A better understanding of the autoimmune response is necessary to predict insulin requirement in LADA patients, which is important to prevent escalation of blood glucose levels and subsequent complications. In previous studies, we have investigated the humoral immune response toward GAD65 as a reflection of islet cell destruction (9). It remains unclear whether the autoimmune response in LADA patients and type 1 diabetic patients differs or whether only the duration of the prodomal period distinguishes between the two groups (10). Therefore, we compared the GAD65-specific humoral autoimmune response in type 1 diabetic patients with that in LADA patients who had or had not progressed to insulin requirement. RESEARCH DESIGN AND METHODS Patients and sera Sera of GAD65Ab-positive type 1 diabetic patients were collected at the Saitama Social Insurance Hospital, Urawa City, Japan (n = 119). All type 1 diabetic patients required insulin treatment at the time of diabetes diagnosis. Sera were collected between 1989 and 2005 and were taken at various times after onset of disease (0–27 years of disease duration [median 1 year]).

e collected at the Saitama Social Insurance Hospital, Urawa City, Japan (n = 119). All type 1 diabetic patients required insulin treatment at the time of diabetes diagnosis. Sera were collected between 1989 and 2005 and were taken at various times after onset of disease (0–27 years of disease duration [median 1 year]). Patients classified as LADA patients were admitted to the Saitama Social Insurance Hospital, Urawa City, Japan. Diagnosis of LADA was made according to the commission of Immunology of Diabetes Society (2) (patients were diagnosed with type 2 diabetes and tested positive for GAD65Ab with an onset age ≥30 years). None of these patients required isulin treatment within the first 6 months after the initial diagnosis. We differentiated two groups of LADA patients, based on their insulin requirements. Nonprogressed LADA patients (n = 56) did not require insulin treatment for over 5 years after diagnosis with type 2 diabetes. Six of these samples were collected at Keio University. Some of the samples were taken earlier (see Table 1); however, all patients were followed to ensure that they did not require insulin treatment for over 5 years past type 2 diabetes diagnosis. Progressed LADA patients (n = 58) developed insulin requirement after the initial LADA classification and had low fasting serum C-peptide levels (≤0.4 ng/ml). Insulin treatment was started at an A1C of ≥8% despite usage of the maximum dose of glibenclamide (5 mg) and observation of a strict diet.

2 diabetes diagnosis. Progressed LADA patients (n = 58) developed insulin requirement after the initial LADA classification and had low fasting serum C-peptide levels (≤0.4 ng/ml). Insulin treatment was started at an A1C of ≥8% despite usage of the maximum dose of glibenclamide (5 mg) and observation of a strict diet. Longitudinal samples were obtained from nine individuals (five male subjects, median age 34 years) who were classified as LADA patients and developed insulin requirement during follow-up. Local institutional ethics committee approval was obtained before collection of all serum samples. Informed consent was obtained from all patients or their legal guardians. The age at onset of diabetes (type 1 or type 2), GAD65Ab titer, duration of diabetes, requirement of insulin, and other clinical relevant information are summarized in Table 1.

mmittee approval was obtained before collection of all serum samples. Informed consent was obtained from all patients or their legal guardians. The age at onset of diabetes (type 1 or type 2), GAD65Ab titer, duration of diabetes, requirement of insulin, and other clinical relevant information are summarized in Table 1. GAD65Ab titer determination GAD65Ab positivity of the serum samples was initially evaluated using a commercial radioimmunoprecipitation assay (Cosmic, Tokyo, Japan) and the manufacturer's suggested cutoff level of 1.5 units/ml. GAD65Ab positivity was confirmed using a radioligand binding assay (RBA) (described below). The World Health Organization standard for GAD65Ab (11) and negative samples were included in every assay to correct for interassay variation and to express immunoglobulin binding levels as units per milliliter (units/ml). Cutoff levels for positivity (34 units/ml) were calculated as the 98th percentile from a healthy control group (n = 50). Samples with a GAD65Ab >1,000 units/ml in the initial screen were diluted to determine their half-maximal binding concentration. Subsequent epitope mapping experiments were carried out at this half-maximal binding concentration. In the Diabetes Antibody Standardization Program 2005 workshop, the GAD65Ab analysis ranked at 80% sensitivity and 91% specificity.

n the initial screen were diluted to determine their half-maximal binding concentration. Subsequent epitope mapping experiments were carried out at this half-maximal binding concentration. In the Diabetes Antibody Standardization Program 2005 workshop, the GAD65Ab analysis ranked at 80% sensitivity and 91% specificity. Recombinant Fab used in this study Monoclonal antibodies used in this study were previously described (12). Recombinant Fabs (rFabs) were produced in Escherichia coli 25F2 cells as previously described (9). Briefly, DPA and DPD were derived from a type 1 diabetic patient and recognize epitopes located at amino acids 483–499 plus 556–586 and 96–173, respectively. Monoclonal antibody b96.11, derived from a patient with autoimmune polyendocrine syndrome type 2, recognizes a conformational epitope involving amino acids located in both the middle and the C-terminus of the molecule (13). Monoclonal antibody MICA-3, isolated from a patient with type 1 diabetes, recognizes an epitope located at amino acid residues 451–585. Epitope-specific RBA Recombinant human [35S]-GAD65 was produced in an in vitro–coupled transcription/translation system with SP6 RNA polymerase and nuclease-treated rabbit reticulocyte lysate (Promega, Madison, WI) as described previously (14). The in vitro–translated [35S]-antigen was kept at −70°C and used within 2 weeks.

pe-specific RBA Recombinant human [35S]-GAD65 was produced in an in vitro–coupled transcription/translation system with SP6 RNA polymerase and nuclease-treated rabbit reticulocyte lysate (Promega, Madison, WI) as described previously (14). The in vitro–translated [35S]-antigen was kept at −70°C and used within 2 weeks. The capacity of the rFab to inhibit GAD65 binding by human serum GAD65Ab was tested in a competitive epitope-specific RBA using protein A Sepharose (Zymed Laboratories) as described (9). The rFab were added at a concentration sufficient to compete binding of the originating intact mAb to GAD65 by at least 80% (0.7–1 μg/ml). The background competition for each rFab was established in competition experiments with normal control sera. The background was subtracted before calculation of percent binding. The cutoff for specific competition was determined as >10% by using a negative control rFab D1.3 (a kind gift from Dr. J. Foote, Arrowsmith Technologies, Seattle), specific to an irrelevant target, hen-egg lysozyme, at 5 μg/ml. Statistical analyses Binding of GAD65Ab to GAD65 in the presence of rFab was expressed as follows: counts per minute of [35]S-GAD65 bound in the presence of rFab/counts per minute of [35S]-GAD65 bound in the absence of rFab × 100.

The capacity of the rFab to inhibit GAD65 binding by human serum GAD65Ab was tested in a competitive epitope-specific RBA using protein A Sepharose (Zymed Laboratories) as described (9). The rFab were added at a concentration sufficient to compete binding of the originating intact mAb to GAD65 by at least 80% (0.7–1 μg/ml). The background competition for each rFab was established in competition experiments with normal control sera. The background was subtracted before calculation of percent binding. The cutoff for specific competition was determined as >10% by using a negative control rFab D1.3 (a kind gift from Dr. J. Foote, Arrowsmith Technologies, Seattle), specific to an irrelevant target, hen-egg lysozyme, at 5 μg/ml. Statistical analyses Binding of GAD65Ab to GAD65 in the presence of rFab was expressed as follows: counts per minute of [35]S-GAD65 bound in the presence of rFab/counts per minute of [35S]-GAD65 bound in the absence of rFab × 100. All samples were analyzed in triplicate determinations, and the intra-assay average coefficient of variation was 5% (range 13–0.04). Median ages, GAD65Ab titers, and competition levels between groups were analyzed using the nonparametric ANOVA (Kruskal-Wallis) followed by Dunn's multiple comparisons test. Competition levels within each group were tested for significance using the nonparametric Wilcoxon matched-pair test. A P value <0.05 was considered significant.

D65Ab titers, and competition levels between groups were analyzed using the nonparametric ANOVA (Kruskal-Wallis) followed by Dunn's multiple comparisons test. Competition levels within each group were tested for significance using the nonparametric Wilcoxon matched-pair test. A P value <0.05 was considered significant. RESULTS Autoantibody status and clinical parameters The type 1 diabetic cohort had a significantly lower median age compared with the nonprogressed and progressed LADA patients (P < 0.0001) (Table 1). No significant difference between the median ages of the nonprogressed LADA patients and the progressed LADA patients was observed. We emphasize that samples from progressed LADA patients were taken after they developed insulin requirement. No significant differences in GAD65Ab levels between the three groups were observed.

gnificant difference between the median ages of the nonprogressed LADA patients and the progressed LADA patients was observed. We emphasize that samples from progressed LADA patients were taken after they developed insulin requirement. No significant differences in GAD65Ab levels between the three groups were observed. GAD65Ab response in relation to insulin requirement All serum samples were analyzed for their binding to GAD65 in the presence of GAD65-specific rFab DPA, b96.11, DPD, and MICA-3 (Fig. 1). We observed significant reduction in median binding to GAD65 in the presence of rFab DPA, DPD, b96.11, and MICA-3 in all groups. No correlation between GAD65Ab titer and reduction of binding conferred by any of the rFab was observed in the type 1 diabetic and progressed LADA patients. In the nonprogressed LADA patients, we observed a significant correlation of GAD65Ab titer and reduction of binding conferred by rFab b96.11 (P = 0.005) and DPD (P = 0.004) (data not shown). No correlation between GAD65Ab epitope specificity and sex or age was observed in any of the groups. To determine whether the epitope recognition differed between the groups, we compared the differences in reduction in median binding to GAD65 conferred by the different rFab (Fig. 1). We found that the reduction in binding conferred by rFab b96.11 was significantly more pronounced in type 1 diabetic patients as compared with progressed and nonprogressed LADA patients (P < 0.01 and P < 0.001, respectively).

differences in reduction in median binding to GAD65 conferred by the different rFab (Fig. 1). We found that the reduction in binding conferred by rFab b96.11 was significantly more pronounced in type 1 diabetic patients as compared with progressed and nonprogressed LADA patients (P < 0.01 and P < 0.001, respectively). GAD65Ab response in relation to GAD65Ab titer Based on our above findings of a correlation between GAD65Ab titer and epitope recognition in the nonprogressed LADA patients and our previous observation that high GAD65Ab titers predict progression to insulin requirement (15), we divided the analysis between samples with GAD65Ab titers > and <1,000 units/ml (Fig. 1, inset). We found that sera exhibiting high GAD65Ab titer samples in both LADA groups showed strong inhibition of GAD65 binding by rFab b96.11,which was similar to that observed in sera obtained from type 1 diabetic patients. For both insulin-requiring patient groups (type 1 diabetes and progressed LADA), no significant differences in inhibition levels observed in the presence of rFab b96.11 were observed when comparing sera with high and low GAD65Ab titers. However, in nonprogressed LADA patients, binding levels in the presence of rFab b96.11 in sera with high GAD65Ab titers were significantly lower compared with sera with low GAD65Ab titers (P < 0.001). Consequently, samples with GAD65Ab titers below the 1,000 units/ml cutoff showed significant differences in the GAD65 binding in the presence of rFab b96.11 between type 1 diabetic patients and progressed (P < 0.01) and nonprogressed LADA patients (P < 0.001).

wer compared with sera with low GAD65Ab titers (P < 0.001). Consequently, samples with GAD65Ab titers below the 1,000 units/ml cutoff showed significant differences in the GAD65 binding in the presence of rFab b96.11 between type 1 diabetic patients and progressed (P < 0.01) and nonprogressed LADA patients (P < 0.001). GAD65Ab response in relation to disease duration We tested whether the GAD65Ab epitope specificities may change longitudinally toward progression to insulin requirement. Therefore, we correlated epitope specificities with disease duration (initial diabetes diagnosis) in nonprogressed LADA patients and time from initial diabetes diagnosis to insulin requirement in progressed LADA patients. No correlation with GAD65Ab titer or GAD65Ab epitope specificity was observed, indicating no longitudinal changes over time. Longitudinal samples obtained from LADA patients (n = 9) during their progression to insulin requirement were analyzed for their epitope specificities (Fig. 2). While some patients showed longitudinal changes over time (Fig. 2A–C), no overall trend in the change of epitope specificities was obvious.

changes over time. Longitudinal samples obtained from LADA patients (n = 9) during their progression to insulin requirement were analyzed for their epitope specificities (Fig. 2). While some patients showed longitudinal changes over time (Fig. 2A–C), no overall trend in the change of epitope specificities was obvious. rFab concentration needed for maximal inhibition is identical in the three groups Disparities in inhibition levels of GAD65 binding exerted by a rFab could be caused by different affinities, or differences in binding specificities. Therefore we established the rFab b96.11 concentration necessary to achieve maximal inhibition in all three groups (Fig. 3). The median rFab concentration to reach 50% inhibition (EC50) was 0.39 nmol/l for all three groups. Serum samples whose GAD65 binding was not inhibited by rFab concentrations of 2.5 nmol/l, were also not inhibited at 12.5 nmol/l rFab. This confirms that the assay conditions are optimal, as the rFab concentration used (12 nmol/l) exceeded the rFab concentration necessary to achieve maximal competition. These results also suggest that the observed differences between the groups were not caused by different binding capacities to the GAD65Ab epitope defined by b96.11.

hat the assay conditions are optimal, as the rFab concentration used (12 nmol/l) exceeded the rFab concentration necessary to achieve maximal competition. These results also suggest that the observed differences between the groups were not caused by different binding capacities to the GAD65Ab epitope defined by b96.11. CONCLUSIONS Our results confirmed previous observations of different GAD65-specific humoral immune responses in type 1 diabetic patients and nonprogressed LADA patients (9,16). Progressed LADA patients exhibited a GAD65Ab epitope pattern intermediate between type 1 diabetic patients and nonprogressed LADA patients. This may indicate that the GAD65Ab response matures in patients as they progress toward insulin requirement. We tested this possibility by analyzing longitudinal samples obtained from a small group of LADA patients during their development of insulin requirement. While some of the patients showed changes in their epitope binding specificities, no overall trend was observed.

atients as they progress toward insulin requirement. We tested this possibility by analyzing longitudinal samples obtained from a small group of LADA patients during their development of insulin requirement. While some of the patients showed changes in their epitope binding specificities, no overall trend was observed. In the progressed LADA patients the disease duration before insulin requirement varies from 0.5 to 27 years. We analyzed whether LADA patients who progressed faster to disease showed different GAD65Ab epitope specificities from patients that progressed slower. However, no correlation between GAD65Ab epitope specificities and length of the prodomal period was observed. These findings together with our earlier observations of longitudinal changes in GAD65Ab epitope specificities in healthy adult individuals during their progression to type 2 diabetes (17) lead to our hypothesis that the autoimmune response in LADA patients remains constant after type 2 diabetes onset.

od was observed. These findings together with our earlier observations of longitudinal changes in GAD65Ab epitope specificities in healthy adult individuals during their progression to type 2 diabetes (17) lead to our hypothesis that the autoimmune response in LADA patients remains constant after type 2 diabetes onset. Some of the nonprogressed LADA patients showed a very long disease duration without developing insulin requirement (up to 27 years since initial diabetes diagnosis). While the presence of GAD65Ab in LADA patients is considered as a risk factor for subsequent insulin requirement (1), one autoantibody alone confers only a low risk for progression in the general population (7). One could therefore assume that some LADA patients show signs of β-cell autoimmunity but are unlikely to develop insulin requirement. To test this hypothesis, we analyzed GAD65Ab epitope specificities in correlation with disease duration. However, no correlation between disease duration and epitope specificity was observed. These data are in agreement with the longitudinal study of LADA patients in the U.K. Prospective Diabetes Study 77, reporting stagnant GAD65Ab epitope reactivities (18).

5Ab epitope specificities in correlation with disease duration. However, no correlation between disease duration and epitope specificity was observed. These data are in agreement with the longitudinal study of LADA patients in the U.K. Prospective Diabetes Study 77, reporting stagnant GAD65Ab epitope reactivities (18). The observed differences in GAD65Ab epitope specificities were particularly pronounced in the samples with medium to low GAD65Ab titers, while high GAD65Ab titer samples in the three groups recognized the type 1 diabetes–associated b96.11 epitope to similar degrees. This may indicate that the type 1 diabetes–associated autoimmune response is more emphasized in LADA patients with high GAD65Ab titers. While these unexpected findings need to be confirmed in a larger study cohort, previous studies (15,19) report that LADA patients with high GAD65Ab titers progress to insulin requirement more often than LADA patients with low GAD65Ab titers. Moreover, a recent study (20) reported that LADA patients with high GAD65Ab titers resemble type 1 diabetic patients with respect to clinical characteristics, genetic susceptibility, and other autoimmune components. However, no correlation between GAD65Ab titer and aggressiveness of β-cell autoimmunity was found in the recent U.K. Prospective Diabetes Study 77 (18). These differences may be caused by different distribution of GAD65Ab titers, as our LADA cohorts included serum samples with very high GAD65Ab titers, while the sera in the U.K. Prospective Diabetes Study 77 cohort showed more moderate GAD65Ab titers.

ity was found in the recent U.K. Prospective Diabetes Study 77 (18). These differences may be caused by different distribution of GAD65Ab titers, as our LADA cohorts included serum samples with very high GAD65Ab titers, while the sera in the U.K. Prospective Diabetes Study 77 cohort showed more moderate GAD65Ab titers. The observed differences in GAD65Ab epitope specificities between high and low GAD65Ab titer samples within the nonprogressed LADA patients suggest a heterogeneous autoimmune response in this group. The disease progression in high titer LADA patients with type 1 diabetes–like GAD65Ab epitope specificity needs to be analyzed in future studies to address this hypothesis. We conclude that the autoimmune responses in LADA and type 1 diabetic patients show different GAD65-specific immune responses, particularly in the samples with moderate GAD65Ab titers. The particular GAD65Ab characteristics remain stable and do not mature during progression to insulin requirement, which may suggest a distinct autoimmune response in the pathogenesis for LADA patients. The study was performed as independent research sponsored by National Institutes of Health Grant DK53456, as well as DK53004, DK26190 (to Å.L.), and DK17047.

Fulminant type 1 diabetes is a subtype of type 1 diabetes characterized by extremely rapid onset that can be classified as type 1B diabetes (1). Although frequent flu-like symptoms prior to onset suggest the contribution of virus infection in the etiology of fulminant type 1 diabetes, both environmental and genetic factors are largely unknown. Susceptibility to classic type 1A diabetes is determined by multiple genes within the HLA region and non-HLA genes, including INS-VNTR, CTLA4, and PTPN22 (2). Among them, CTLA4 is associated with autoimmunity, cancer, allergy, and infectious disease. In the CTLA4 region, a number of variants, such as +49G>A and CT60, have shown type 1 diabetes association (3). Although the association between class II HLA and fulminant type 1 diabetes has been reported (4), the contribution of the non-HLA genes to the susceptibility to fulminant type 1 diabetes has not been investigated. In this study, we examined the genetic contribution of the CTLA4 gene to fulminant type 1 diabetes compared with classic type 1A diabetes.

class II HLA and fulminant type 1 diabetes has been reported (4), the contribution of the non-HLA genes to the susceptibility to fulminant type 1 diabetes has not been investigated. In this study, we examined the genetic contribution of the CTLA4 gene to fulminant type 1 diabetes compared with classic type 1A diabetes. RESEARCH DESIGN AND METHODS We examined 55 patients with fulminant type 1 diabetes (49% female, median age at onset 35.0 years), 91 patients with classic type 1A diabetes (57% female, median age at onset 17.0 years), and 369 healthy control subjects. Diagnostic criteria for fulminant type 1 diabetes are described elsewhere (1). The criteria for the recruitment of type 1A diabetic patients were presence of diabetic ketosis at onset, duration of hyperglycemic symptoms <3 months prior to initiation of insulin therapy, and positive for at least one of the anti-islet autoantibodies. This study was approved by the appropriate ethics committees, and informed consent was obtained from all subjects. Genotyping of two single nucleotide polymorphisms in the CTLA4 gene, +49G>A (rs231775) and CT60 (rs3087243), was performed as reported previously (5). Serum concentration of sCTLA4 was measured by enzyme-linked immunosorbent assay using human soluble CTLA4 (sCTLA4) kit (MedSystems Diagnostics, Vienna, Austria), according to the manufacturer's protocol. Sera from type 1 diabetic patients were obtained at disease onset and stored at −20°C until use.

ted previously (5). Serum concentration of sCTLA4 was measured by enzyme-linked immunosorbent assay using human soluble CTLA4 (sCTLA4) kit (MedSystems Diagnostics, Vienna, Austria), according to the manufacturer's protocol. Sera from type 1 diabetic patients were obtained at disease onset and stored at −20°C until use. The significance of differences in the distribution of genotypes between case and control subjects was determined by χ2 test or Fisher's exact probability test. Comparisons of the sCTLA4 levels were made by ANOVA with phenotypic group alone and ANOVA with phenotypic group and CTLA4 genotype. P < 0.05 was considered to be statistically significant. RESULTS The +49G>A variant was associated with classic type 1A diabetes but not with fulminant type 1 diabetes (Table 1). In contrast, the contribution of CT60 to disease is distinct from that of +49G>A. The frequency of the CT60AA genotype in fulminant type 1 diabetic patients was significantly higher than in control subjects (P = 0.021) and type 1A diabetic patients (P = 0.031). CT60GG was associated with type 1A diabetes (P = 0.008). Because of the strong association of HLA-DR4 in both patient groups (1), the effect of CTLA4 on type 1 diabetes susceptibility relative to HLA-DR4 was also examined. Among DR4-positive individuals, the frequency of the CT60AA genotype was significantly increased in patients with fulminant type 1 diabetes (P = 0.005). However, stratification of patients by the presence or absence of HLA-DR4 did not affect the association between the +49GG genotype and type 1A diabetes (Table 1).

ed. Among DR4-positive individuals, the frequency of the CT60AA genotype was significantly increased in patients with fulminant type 1 diabetes (P = 0.005). However, stratification of patients by the presence or absence of HLA-DR4 did not affect the association between the +49GG genotype and type 1A diabetes (Table 1). It has been reported that the CT60G allele might be associated with lower production of sCTLA4 mRNA (3). We therefore determined serum sCTLA4 levels. The mean sCTLA4 levels in fulminant type 1 diabetic patients (0.56 ± 0.24 ng/ml [mean ± SD], n = 36) was significantly lower than those in type 1A diabetic patients (0.94 ± 0.87 ng/ml, n = 45) and control subjects (0.89 ± 0.76 ng/ml, n = 23) (P = 0.043). A mixed-model ANOVA using phenotypic group (fulminant type 1 diabetes, type 1A diabetes, and control) and CT60 genotype (GG and GA+AA) as factorial fixed effects revealed no differences in sCTLA4 levels between CT60 genotypes (P = 0.76) or phenotype/genotype interactions (P = 0.40).

9 ± 0.76 ng/ml, n = 23) (P = 0.043). A mixed-model ANOVA using phenotypic group (fulminant type 1 diabetes, type 1A diabetes, and control) and CT60 genotype (GG and GA+AA) as factorial fixed effects revealed no differences in sCTLA4 levels between CT60 genotypes (P = 0.76) or phenotype/genotype interactions (P = 0.40). CONCLUSIONS CTLA4, which delivers inhibitory signals to T-cell activation, is expressed on the surface of activated T-cells and regulatory T-cells, and the lack of CTLA4 results in uncontrolled T-cell–mediated lymphoproliferative disease (6). Furthermore, CTLA4 also has a significant biological role in attenuating T-cell responses in the context of an inflammatory environment, such as infection with a pathogen (7). We showed that CTLA4 is associated with an increased risk of fulminant type 1 diabetes and that its contribution is distinct from classic type 1A diabetes. As reported previously (5,8), a significant association between classic type 1A diabetes and +49GG and CT60GG genotype was also found in the present study. However, the CT60AA genotype contributes to the susceptibility to fulminant type 1 diabetes. Moreover, it is implicated that HLA-DR4 influences the association of fulminant type 1 diabetes with the CT60AA genotype.

ation between classic type 1A diabetes and +49GG and CT60GG genotype was also found in the present study. However, the CT60AA genotype contributes to the susceptibility to fulminant type 1 diabetes. Moreover, it is implicated that HLA-DR4 influences the association of fulminant type 1 diabetes with the CT60AA genotype. In this study, we also revealed that serum sCTLA4 level in fulminant type 1 diabetic patients were significantly lower than those in type 1A diabetic patients and control subjects. Although it remains unknown how sCTLA4 regulates T-cell activation, recombinant sCTLA4 inhibits T-cell proliferation in vitro. Furthermore, sCTLA4 is constitutively expressed in nonstimulated T-cells, and its expression is downregulated after T-cell activation (9). Therefore, the decreased levels of sCTLA4 might indicate a lower potential of T-cell inhibition in fulminant type 1 diabetes, which might be caused by functional defects leading to reduced production of sCTLA4 or ongoing activation of the immune system eventually leading to decreased levels of sCTLA4. In conclusion, the present study implicates that CTLA4 confers susceptibility to fulminant type 1 diabetes. Furthermore, the different contributions of CTLA4 to susceptibility to fulminant and classic type 1A diabetes indicate that the underlying immune process–primed β-cell injury might be distinct between these subtypes of type 1 diabetes. This study was partly supported by a grant from the Ministry of Education, Culture, Science, Sports and Technology of Japan. We thank Shinobu Mitsui for excellent technical assistance.

Diabetic neuropathy is a common diabetes complication that may result in serious consequences such as pain, foot ulcers, and amputations. Although optimal glycemic control is considered an effective preventive measure, intervention studies in advanced stages of diabetic neuropathy have been almost uniformly unsuccessful (1). Only arrest of progression of diabetic neuropathy could be achieved in patients after pancreas transplantation (2). To assess nerve regeneration following pancreas transplantation, Kennedy et al. (3) proposed the use of skin biopsies with quantification of intraepidermal nerve fiber density (IENFD). Previously, we documented severe IENFD reduction in lower-limb skin biopsies performed at the time of pancreas transplantation (4). Here we present assessment of IENFD following a mean of 2.5 years of normoglycemia. RESEARCH DESIGN AND METHODS A total of 22 patients with type 1 diabetes undergoing simultaneous pancreas/kidney transplantation (SPK) and 14 healthy control subjects participated in the study. For details of the procedure and study subjects, please see the online appendix (available at http://dx.doi.org/10.2337/dc07-2409). The study was approved by the local ethics committee, and informed consent was obtained from all subjects.

transplantation (SPK) and 14 healthy control subjects participated in the study. For details of the procedure and study subjects, please see the online appendix (available at http://dx.doi.org/10.2337/dc07-2409). The study was approved by the local ethics committee, and informed consent was obtained from all subjects. Skin biopsies were performed using a 3-mm punch (Stiefel Laboratories, Sligo, Ireland) from the distal thigh (two samples at a distance of 1 cm, one assessed in Prague and the other in Würzburg) and the proximal calf (one sample, assessed in Prague) at the time of SPK and at 30 ± 5 (mean ± SD) months post transplant. Biopsies from control subjects were taken from corresponding regions. After fixation (4% paraformaldehyde for 3 h at 4°C, then cryoprotection with 10% sucrose in 0.1 mol/l PBS) and freezing (in isopentane cooled by liquid nitrogen), 40-μm sections were immunoreacted with a rabbit polyclonal antibody to the panaxonal marker protein gene product (PGP) 9.5 (DakoCytomation, Glostrup, Denmark), followed by mouse anti-rabbit IgG conjugated with rhodamine or Cy3 (Jackson Immuno Research, West Grove, PA). Samples were imaged with an Olympus microscope BX 51 (Olympus Optical, Hamburg, Germany) in Prague and with a Zeiss Axiophot 2 (Carl Zeiss, Göttingen, Germany) in Würzburg. Three sections per patient were examined. The mean number of intraepidermal nerve fibers (IENFs) per millimeter epidermis was derived using the software Olympus DP-SOFT (Software Imaging Systems, Münster, Germany) and Image Pro Plus 4.0 (Media Cybernetics, Leiden, Netherlands), respectively. Established counting rules were followed (5). Changes >1 IENF/mm were considered meaningful. In addition, the subepidermal nerve plexus was classified semiquantitatively in Würzburg as “normal,” “reduced,” or “absent.” Clinical neuropathy evaluation in the patients included vibration perception threshold (VPT) tests (Bio-Thesiometer; Bio-Medical Instrument, Newbury, OH) and autonomic function testing (AFT) (VariaPulse TF3; Sima Media, Olomouc, Czech Republic) (6). The Mann-Whitney U test and Wilcoxon's signed-rank test were used for inter- and intra-group comparisons, respectively.

cluded vibration perception threshold (VPT) tests (Bio-Thesiometer; Bio-Medical Instrument, Newbury, OH) and autonomic function testing (AFT) (VariaPulse TF3; Sima Media, Olomouc, Czech Republic) (6). The Mann-Whitney U test and Wilcoxon's signed-rank test were used for inter- and intra-group comparisons, respectively. RESULTS Normoglycemia with insulin independence and satisfactory renal graft function was achieved in 18 patients (male/female 10/8, aged 47 ± 10 years, with diabetes duration 29 ± 9 years and P-creatinine 1.3 ± 0.4 mg/dl at follow-up; online appendix Table A1). At baseline, significantly increased VPTs, reduced AFT results (online appendix Table A2), and severe reduction in IENFD in both regions were present in SPK recipients (Table 1 and online appendix Figure A1). At follow-up 21–40 months (median 29) after SPK, increases in IENFD of the thigh samples were seen in three patients, with results verified in both Prague and Würzburg (median 4.1, range 1.9–10.2 IENF/mm). The subepidermal plexus was reduced or absent in all but one patient. A change in category from “reduced” to “normal” occurred in two patients with improvement of IENFD but in none of the other patients. No significant changes occurred in neurological function or IENFD of the transplanted group as a whole.

.2 IENF/mm). The subepidermal plexus was reduced or absent in all but one patient. A change in category from “reduced” to “normal” occurred in two patients with improvement of IENFD but in none of the other patients. No significant changes occurred in neurological function or IENFD of the transplanted group as a whole. CONCLUSIONS Previous reports of neuropathy follow-up in pancreas or islet transplant recipients were mostly based on clinical examination, electrophysiology, and AFT. Most recently, stabilization of electrophysiological parameters could be shown over a 6-year period in 18 patients with islet transplantation after kidney transplantation (7). An innovative noninvasive approach, corneal confocal microscopy, was proposed by researchers from Manchester (8). Using this method, a significant improvement of corneal nerve fiber density and length was detected within 6 months of SPK (9).

od in 18 patients with islet transplantation after kidney transplantation (7). An innovative noninvasive approach, corneal confocal microscopy, was proposed by researchers from Manchester (8). Using this method, a significant improvement of corneal nerve fiber density and length was detected within 6 months of SPK (9). We did not encounter a similarly significant early regenerative response of lower-limb nerve fibers after SPK. While a type II error cannot be excluded and more advanced diabetic neuropathy could have been present, other reasons may be also responsible. The length-related pattern of diabetic neuropathy and varying regenerative capacity of nerve fibers from different body regions could play a role. Moreover, the subepidermal plexus from which epidermal reinnervation should occur was reduced in most patients. We observed some improvement of nerve fiber counts in the biopsies from the more proximal lower-thigh area in three patients. While this subgroup did not differ in clinical characteristics including time from SPK, a still longer period of normoglycemia might be needed to achieve nerve fiber regeneration in the lower limbs of the remaining patients. Of note, in the case of diabetic nephropathy, reversal of renal lesions was seen after more than 5 years of normoglycemia following pancreas transplantation (10).

uding time from SPK, a still longer period of normoglycemia might be needed to achieve nerve fiber regeneration in the lower limbs of the remaining patients. Of note, in the case of diabetic nephropathy, reversal of renal lesions was seen after more than 5 years of normoglycemia following pancreas transplantation (10). Irreparable damage of lower-limb nerves might also be present in some advanced cases. Although generally producing an immense improvement of the recipient's clinical condition and long-term prognosis, SPK does not eliminate risks connected with diabetic neuropathy. Matricali et al. (11) recently reported on a high rate of Charcot foot complications at a mean of 1.8 years posttransplant. Foot ulcers and gangrene, while often co-initiated by vascular disease and infection, are not uncommon throughout the postoperative period. Such complications have occurred in 62 of 200 pancreas transplantation recipients at our center since 1994. Supplementary Material Online-Only Appendix This study was supported by the NR 7929-4/2004 grant, Internal Grant Agency, Czech Ministry of Health; Bundesministerium für Bildung und Forschung (Deutsches Forschungsnetzwerk, Neuropathischer Schmerz); and intramural funds, University of Würzburg.

Intramyocellular triglyceride (TG) content in skeletal muscle is increased in patients with type 1 diabetes compared with that in control subjects (1), suggesting a major role of metabolic dysregulation in the induction of abnormal intramyocellular lipid accumulation in type 1 diabetes. Recent optimalization of magnetic resonance (MR) spectroscopy techniques allowed us to study myocardial TG content and myocardial function in vivo and to document an inverse relationship between myocardial TG content and myocardial function in healthy subjects (2). Therefore, we hypothesized that episodes of metabolic dysregulation due to insufficient insulin provision may adversely affect myocardial TG content and myocardial function.

myocardial function in vivo and to document an inverse relationship between myocardial TG content and myocardial function in healthy subjects (2). Therefore, we hypothesized that episodes of metabolic dysregulation due to insufficient insulin provision may adversely affect myocardial TG content and myocardial function. RESEARCH DESIGN AND METHODS MR imaging and MR spectroscopy were performed twice in 10 C-peptide–negative, nonsmoking patients with type 1 diabetes (five male) using insulin treatment by insulin pump therapy (continuous subcutaneous insulin infusion). No participant showed evidence of cardiovascular disease. The study was designed to mimic hyperglycemic dysregulation in daily life. Therefore, the subjects were instructed to retain their daily routine. One study was done after a 3-day period in which subjects aimed at optimal blood glucose levels measured by a continuous glucose monitoring system (Medtronic). The second study was performed after ∼50% reduction in basal and bolus insulin infusions during 24 h, compared with the first study, in order to maintain hyperglycemia with glucose levels between 15 and 20 mmol/l. For both occasions, patients were instructed to maintain the same caloric intake for 3 days before examination. The sequence between the euglycemic and hyperglycemic occasions was randomly assigned. Before MR examination, blood samples (postprandial) were taken. An ethics committee approved the study, and subjects signed informed consent.

patients were instructed to maintain the same caloric intake for 3 days before examination. The sequence between the euglycemic and hyperglycemic occasions was randomly assigned. Before MR examination, blood samples (postprandial) were taken. An ethics committee approved the study, and subjects signed informed consent. MR spectroscopy measurements (1.5-T; Philips) were obtained using a point-resolved, spatially localized spectroscopic pulse sequence to acquire single voxel (8 ml) spectra. For the heart, data acquisition was double triggered (electrocardiogram triggering and navigator echoes [3]). For the liver, voxel sites were matched at the study occasions. Lipid resonances of myocardial and hepatic TG were summed and calculated as a percentage of the unsuppressed water signal ([TG/water] × 100). To assess left ventricular (LV) systolic function, the heart was imaged in the short axis orientation. To assess LV diastolic function, a phase contrast sequence with velocity encoding was performed to measure blood flow across the mitral valve (4). Analysis was performed using MASS and FLOW (Medis) to quantify LV ejection fraction and flow velocities in early diastole (early filling phase [E]) and at atrial contraction (atrial filling phase [A], E/A ratio, and E deceleration). Data were compared by paired t test and are shown as mean ± SEM. P < 0.05 was considered to reflect significant differences.

To assess left ventricular (LV) systolic function, the heart was imaged in the short axis orientation. To assess LV diastolic function, a phase contrast sequence with velocity encoding was performed to measure blood flow across the mitral valve (4). Analysis was performed using MASS and FLOW (Medis) to quantify LV ejection fraction and flow velocities in early diastole (early filling phase [E]) and at atrial contraction (atrial filling phase [A], E/A ratio, and E deceleration). Data were compared by paired t test and are shown as mean ± SEM. P < 0.05 was considered to reflect significant differences. RESULTS Patient characteristics at baseline and during hyperglycemia are shown in Table 1. During partial insulin deprivation, hyperglycemic dysregulation was present in all patients (mean plasma 24-h glucose was 8.4 ± 0.6 mmol/l during the control study, which increased to 15.9 ± 0.8 mmol/l during partial insulin deprivation [P < 0.001]) and associated with an increase in plasma levels of nonesterified fatty acids from 0.31 ± 0.05 to 0.46 ± 0.07 mmol/l (P = 0.015). Myocardial TG content was 0.31 ± 0.04% at baseline and did not change during hyperglycemic dysregulation (0.34 ± 0.06%; P = 0.587). E/A ratio was unaffected (1.9 ± 0.2 at baseline vs. 1.9 ± 0.3).

ssociated with an increase in plasma levels of nonesterified fatty acids from 0.31 ± 0.05 to 0.46 ± 0.07 mmol/l (P = 0.015). Myocardial TG content was 0.31 ± 0.04% at baseline and did not change during hyperglycemic dysregulation (0.34 ± 0.06%; P = 0.587). E/A ratio was unaffected (1.9 ± 0.2 at baseline vs. 1.9 ± 0.3). CONCLUSIONS This is the first study to document the effects of short-term hyperglycemic dysregulation on myocardial TG content and LV function in patients with type 1 diabetes. The present study shows that hyperglycemic dysregulation for 24 h, as frequently observed in patients with type 1 diabetes, does not modulate myocardial TG content or myocardial function, despite considerable metabolic dysregulation. We hypothesized that short-term partial insulin deprivation results in changes in myocardial TG content, possibly associated with changes in myocardial function. Stiffness of intermediate-sized arteries is rapidly increased in patients with type 1 diabetes during hyperglycemia, whereas larger arteries seem unaffected (5). Moreover, in healthy subjects myocardial function and TG content rapidly adapt to changes in metabolic state (2). Interestingly, this adaptation could not be evoked by short-term hyperglycemic dysregulation, suggesting that the heart is protected from these short-term effects. Nonetheless, we cannot exclude the possibility that prolongation of the duration of partial insulin deprivation might have resulted in changes in myocardial function and TG content.

ation could not be evoked by short-term hyperglycemic dysregulation, suggesting that the heart is protected from these short-term effects. Nonetheless, we cannot exclude the possibility that prolongation of the duration of partial insulin deprivation might have resulted in changes in myocardial function and TG content. Patients with type 1 diabetes have considerably altered myocardial glucose and fatty acid metabolism. Myocardial fatty acid utilization is increased, whereas myocardial glucose uptake is considerably lower in diabetic patients compared with that in control subjects (6). These adaptations protect the heart-to-substrate overflow of the myocardium. Accordingly, the present study interestingly shows that myocardial TG content in diabetic patients was not different from the values we observed in studies in healthy subjects (2,3). However, fatty acid kinetics during hyperglycemia cannot be derived from the present data, and we cannot exclude changes in myocardial fatty acid oxidation. Hepatic TG content was measured to study the effects of insulin deprivation on tissue-specific TG distribution because we previously found discrepant effects of interventions on heart and liver (2). However, in the present study design insulin deprivation did not result in altered hepatic TG content. In conclusion, short-term partial insulin deprivation resulting in hyperglycemic dysregulation, which is frequently observed in patients with type 1 diabetes, does not modulate myocardial or hepatic TG content or LV function, despite considerable metabolic dysregulation.

Hepatic TG content was measured to study the effects of insulin deprivation on tissue-specific TG distribution because we previously found discrepant effects of interventions on heart and liver (2). However, in the present study design insulin deprivation did not result in altered hepatic TG content. In conclusion, short-term partial insulin deprivation resulting in hyperglycemic dysregulation, which is frequently observed in patients with type 1 diabetes, does not modulate myocardial or hepatic TG content or LV function, despite considerable metabolic dysregulation. We thank Marja Dijk-Schaap and Nathalie Masurel for their assistance with the study.

Heme oxidase (HO) is a rate-limiting enzyme in heme degradation that leads to the generation of free iron, biliverdin, and carbon monoxide. Biliverdin is subsequently converted to bilirubin via the action of biliverdin reductase, and free iron is promptly sequestered into ferritin. There are two genetically distinct isozymes of HO: the inducible HO-1 and a constitutively expressed HO-2. HO-1 is a cytoprotective enzyme upregulated in mammals mostly dependent on transcriptional activation of the HO-1 gene to diverse cellular stress. The relationship of HO to atherosclerotic vascular disease was suggested initially in 1994 by an observational study reporting that low serum concentrations of bilirubin are associated with increased risk of coronary artery disease (CAD) (1). The human HO-1 gene has been mapped to chromosome 22q12, and a (GT)n dinucleotide repeat has been identified in the proximal promoter region (2). The (GT)n repeat is highly polymorphic and modulates gene transcription by oxidant challenge (3). We and others have demonstrated that a longer (GT)n repeat exhibits lower transcriptional activity and is associated with susceptibility to CAD in high-risk patients (4,5).

n identified in the proximal promoter region (2). The (GT)n repeat is highly polymorphic and modulates gene transcription by oxidant challenge (3). We and others have demonstrated that a longer (GT)n repeat exhibits lower transcriptional activity and is associated with susceptibility to CAD in high-risk patients (4,5). Bilirubin, a natural product of heme catabolism by HO, has been recognized to be an antioxidant and can inhibit lipid peroxidation (6). There is accumulating evidence that individuals with high-normal or just greater than normal plasma bilirubin levels have a lesser incidence of CAD and carotid plaque formation (7,8). HO-1 is also of critical contribution to iron homeostasis. The association between body iron status and the risk of cardiovascular disease was first postulated by Sullivan in the early 1980s (9) and thereafter by a number of epidemiological studies (10). Because HO-1 promoter polymorphism can conceivably affect the development of CAD, in the present study, the associations of the HO-1 promoter polymorphism with bilirubin levels, markers of iron status, and the development of CAD were examined.

early 1980s (9) and thereafter by a number of epidemiological studies (10). Because HO-1 promoter polymorphism can conceivably affect the development of CAD, in the present study, the associations of the HO-1 promoter polymorphism with bilirubin levels, markers of iron status, and the development of CAD were examined. RESEARCH DESIGN AND METHODS The study population consisted of 986 unrelated adult patients who consecutively underwent coronary angiography in the Cardiology Division at Taipei Veterans General Hospital from August 1999 to October 2000. CAD was documented by angiographic evidence of ≥75% stenosis of at least one major coronary artery or a history of prior angioplasty, coronary artery bypass surgery, or myocardial infarction by history validated by electrocardiographic changes. The non-CAD group consisted of subjects who had normal coronary arteries as documented by angiography (<20% intraluminal obstruction) and had neither a history of atherosclerosis nor clinical or laboratory evidence of atherosclerosis in other vascular beds. This study protocol was approved by the review committee of Taipei Veterans General Hospital, and all participants gave their written informed consent.

angiography (<20% intraluminal obstruction) and had neither a history of atherosclerosis nor clinical or laboratory evidence of atherosclerosis in other vascular beds. This study protocol was approved by the review committee of Taipei Veterans General Hospital, and all participants gave their written informed consent. Analysis of length variability of (GT)n repeats in HO-1 gene promoter Genomic DNA was extracted from leukocytes by the conventional procedure. The 5′-flanking region containing (GT)n repeats of the HO-1 gene was amplified by PCR with a FAM-labeled sense primer, 5′-AGAGCCTGCAGCTTCTCAGA-3′, and an antisense primer, 5′-ACAAAGTCTGGCCATAGGAC-3′, as described previously (4). The PCR products were mixed together with a GenoType TAMRA DNA ladder (size range 50–500 bp; GibcoBRL) and analyzed with an automated DNA sequencer (ABI Prism 377). Each size of the (GT)n repeat was calculated using GeneScan Analysis software (PE Applied Biosystems).

CAAAGTCTGGCCATAGGAC-3′, as described previously (4). The PCR products were mixed together with a GenoType TAMRA DNA ladder (size range 50–500 bp; GibcoBRL) and analyzed with an automated DNA sequencer (ABI Prism 377). Each size of the (GT)n repeat was calculated using GeneScan Analysis software (PE Applied Biosystems). Baseline measurements Hypertension was defined as measured systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg. Diabetes was diagnosed on the basis of the World Health Organization criteria. Patients with hypercholesterolemia were defined as those having a total cholesterol level of >240 mg/dl or those who were receiving lipid-lowering therapy. Laboratory measurements were made on 12-h fasting venous blood samples. The biochemical indicators of iron status in this study included the serum iron concentration, the serum ferritin levels, the serum total iron-binding capacity (TIBC), and the serum transferrin saturation. Serum iron was measured with a colorimetric assay. Serum ferritin and TIBC values were assessed with an immunometric assay (Boehringer Mannheim). Transferrin saturation was calculated as the ratio of serum iron to TIBC.

ferritin levels, the serum total iron-binding capacity (TIBC), and the serum transferrin saturation. Serum iron was measured with a colorimetric assay. Serum ferritin and TIBC values were assessed with an immunometric assay (Boehringer Mannheim). Transferrin saturation was calculated as the ratio of serum iron to TIBC. Statistical analysis All statistical analyses were conducted using the SPSS statistical package (version 10.0; SPSS, Chicago, IL). Distributions of continuous variables in groups were expressed as means ± SD and compared by t test for two groups or ANOVA using the least significant difference method for post hoc multivariate comparison of the means for more than three groups. Values of serum ferritin were log transformed because of their skewed distributions. Categorical variables were analyzed by χ2 test or Fisher's exact test. The association of CAD status with the allele frequency was assessed with consideration of confounding effects by known coronary risk factors, such as age, sex, diabetes, hypercholesterolemia, hypertension, and smoking habits. After preliminary bivariate analysis using the t test and χ2 test, multiple logistic regression analysis with forward stepwise selection was performed to evaluate the effect of genotype on CAD after controlling for other established risk factors of CAD. Significance was accepted at P < 0.05. All of the study participants were Chinese from northern Taiwan and had similar ethnic backgrounds.

ple logistic regression analysis with forward stepwise selection was performed to evaluate the effect of genotype on CAD after controlling for other established risk factors of CAD. Significance was accepted at P < 0.05. All of the study participants were Chinese from northern Taiwan and had similar ethnic backgrounds. RESULTS The allele frequencies of (GT)n microsatellite polymorphism in the HO-1 promoter region were highly polymorphic, ranging from 16 to 38 (4). Because the proportion of allele frequencies of either <27 or >27 GT repeats was ∼50%, we classified the alleles into two subgroups: the lower component, with repeat number <27, was designated as “class S,” and the upper component, with ≥27 GT repeats, was designated as “class L.” These patients were then classified as having an S/S, S/L, or L/L genotype according to each of their HO-1 alleles. Table 1 of the online appendix (available at http://dx.doi.org/10.2337/dc07-2126) shows the distribution of HO-1 promoter genotypes in all subjects and in those with hypertension (n = 639), diabetes (n = 263), or hypercholesterolemia (n = 179) and those who currently smoked (n = 260) stratified by the status of CAD. No significant difference in genotypic frequencies between the two groups (CAD vs. non-CAD) in the whole study population was observed. However, diabetes was found to have a significant interaction with genotypes: the proportions of S/S, S/L, and L/L genotypes were 36.5, 47.6, and 15.9%, respectively, in diabetic subjects without CAD and 18.5, 51.5, and 30.0%, respectively, in diabetic subjects with CAD.

CAD) in the whole study population was observed. However, diabetes was found to have a significant interaction with genotypes: the proportions of S/S, S/L, and L/L genotypes were 36.5, 47.6, and 15.9%, respectively, in diabetic subjects without CAD and 18.5, 51.5, and 30.0%, respectively, in diabetic subjects with CAD. The characteristics of the whole study population and subjects with diabetes stratified by HO-1 genotype are presented in Table 1. Across the three genotypes, only serum bilirubin and ferritin concentrations were significantly different in both the whole study population and subjects with diabetes. There were no significant differences in age, sex, percentages of risk factors, levels of serum cholesterol, triglycerides, fasting blood glucose, or markers of iron status including serum iron, TIBC, and transferrin saturation values.

gnificantly different in both the whole study population and subjects with diabetes. There were no significant differences in age, sex, percentages of risk factors, levels of serum cholesterol, triglycerides, fasting blood glucose, or markers of iron status including serum iron, TIBC, and transferrin saturation values. The mean serum bilirubin level was higher in carriers of the S allele (0.85 ± 0.32 mg/dl) than in those with the L/L genotype (0.79 ± 0.25 mg/dl) (P = 0.013) in the whole study population, and the difference was more pronounced (0.81 ± 0.24 vs. 0.70 ± 0.22 mg/dl, P = 0.001) in subjects with diabetes. Serum ferritin levels were highest in subjects with the L/L genotype, intermediate in those with the S/L genotype, and lowest in those with the S/S genotype in the whole study population and in subjects with diabetes. When subjects with the L/L genotype and those carrying the S allele were compared, the ferritin level was significantly higher in subjects with the L/L genotype (127 ± 99 or 4.54 ± 0.88 μg/l for log ferritin) than in carriers of the S allele (114 ± 107 or 4.33 ± 0.98 μg/l for log ferritin) (P = 0.008 for log ferritin) in the whole study population. Among subjects with diabetes, this difference was again much greater (148 ± 104 vs. 111 ± 96 μg/l for ferritin, P = 0.031 or 4.76 ± 0.72 vs. 4.28 ± 1.05 for log ferritin, P = 0.001).

in carriers of the S allele (114 ± 107 or 4.33 ± 0.98 μg/l for log ferritin) (P = 0.008 for log ferritin) in the whole study population. Among subjects with diabetes, this difference was again much greater (148 ± 104 vs. 111 ± 96 μg/l for ferritin, P = 0.031 or 4.76 ± 0.72 vs. 4.28 ± 1.05 for log ferritin, P = 0.001). The baseline characteristics of the whole study population and subjects with diabetes stratified by the status of CAD are summarized in Table 2. When all subjects were considered, patients with CAD were older and had a higher percentage of male sex, higher fasting blood glucose and triglyceride levels, and a lower HDL value compared with those without CAD, as expected. Serum bilirubin levels were significantly lower (0.81 ± 0.30 vs. 0.87 ± 0.32 mg/dl, P = 0.006) and a trend toward a higher serum ferritin level was observed in patients with CAD (126 ± 124 vs. 110 ± 95 μg/l, P = 0.061). There was no difference in serum iron value, TIBC, or transferrin saturation between subjects with versus without CAD. On the other hand, with respect to demographic characteristics, the two groups of diabetic patients with versus without CAD only differed in percentages of male sex. Diabetic patients with CAD had significantly lower serum bilirubin levels (0.76 ± 0.23 vs. 0.86 ± 0.32 mg/dl, P = 0.040) and higher serum ferritin levels (141 ± 139 vs. 104 ± 102 μg/l or 4.54 ± 1.01 vs. 4.16 ± 1.10 μg/l for log ferritin, P = 0.024 for log ferritin).

without CAD only differed in percentages of male sex. Diabetic patients with CAD had significantly lower serum bilirubin levels (0.76 ± 0.23 vs. 0.86 ± 0.32 mg/dl, P = 0.040) and higher serum ferritin levels (141 ± 139 vs. 104 ± 102 μg/l or 4.54 ± 1.01 vs. 4.16 ± 1.10 μg/l for log ferritin, P = 0.024 for log ferritin). The relations between serum bilirubin and ferritin levels, HO-1 genotypes, and CAD are shown in Fig. 1. Among subjects with diabetes, serum bilirubin levels in patients with CAD who had the L/L genotype were significantly lower than those in carriers of the S allele, regardless of their CAD status, whereas differences in serum bilirubin levels between carriers of the L/L genotype with and without CAD were not statistically significant. On the other hand, log ferritin values in patients with CAD who had the L/L genotype were significantly higher than those in carriers of the S allele with or without CAD. Differences in log ferritin values between patients with and without CAD who had the L/L genotype, although substantial, did not reach statistical significance. Differences in both serum bilirubin and ferritin levels in nondiabetic subjects were much less prominent than those in subjects with diabetes.

or without CAD. Differences in log ferritin values between patients with and without CAD who had the L/L genotype, although substantial, did not reach statistical significance. Differences in both serum bilirubin and ferritin levels in nondiabetic subjects were much less prominent than those in subjects with diabetes. We then performed multivariate analyses to further examine the links between serum bilirubin and ferritin levels, HO-1 genotypes, and CAD in diabetic patients. After we controlled for conventional risk factors, carriers of the L/L genotype showed significantly enhanced susceptibility to CAD compared with those carrying the S allele, resulting in an odds ratio (OR) of 2.81 (95% CI 1.22–6.47, P = 0.015) (Table 3). As a next step, we investigated the association of serum bilirubin and ferritin levels with CAD separately. When the HO-1 genotype was not included in the model, a 0.1 mg/dl increase in bilirubin levels decreased CAD risk by 16% and 1 log unit elevation in ferritin values increased CAD risk by 41%. After we included both the HO-1 promoter genotype and bilirubin levels in the logistic regression model, the OR of HO-1 effect fell to 2.65 and became less significant (95% CI 1.05–6.69, P = 0.040). When both the HO-1 promoter genotype and ferritin values were included, the OR of HO-1 effect decreased to 2.31 and was of borderline significance (95% CI 0.97–5.49, P = 0.058). With adjustment of both serum bilirubin and ferritin values, the OR of HO-1 effect was reduced further to 1.71 and became nonsignificant (95% CI 0.75–3.90, P = 0.203) (Table 3).

e and ferritin values were included, the OR of HO-1 effect decreased to 2.31 and was of borderline significance (95% CI 0.97–5.49, P = 0.058). With adjustment of both serum bilirubin and ferritin values, the OR of HO-1 effect was reduced further to 1.71 and became nonsignificant (95% CI 0.75–3.90, P = 0.203) (Table 3). CONCLUSIONS Decreased HO-1 expression has been shown in humans and experimental animals with diabetes (11,12), and an inverse relationship between the HO-1 activity and vascular complications associated with diabetes was demonstrated (13). In line with these findings, our previous study (4) revealed that the length polymorphism in the HO-1 gene promoter is correlated with susceptibility to CAD in diabetic patients. In the present study, we further demonstrated that this effect might be conveyed through its influence on bilirubin and ferritin. The concept that HO-1 may be causally related to cardiovascular diseases in humans has been suggested by studies assessing the (GT)n dinucleotide length polymorphism in the 5′-flanking sequence of the human HO-1 gene. By using HO-1 promoter/luciferase reporter genes carrying different lengths of (GT)n repeats, we demonstrated previously that the more (GT)n repeats in the promoter region, the less transcriptional activity of the HO-1 gene in rat aortic smooth muscle cells (4); a similar result was also shown earlier in Hep3B cells (3).

using HO-1 promoter/luciferase reporter genes carrying different lengths of (GT)n repeats, we demonstrated previously that the more (GT)n repeats in the promoter region, the less transcriptional activity of the HO-1 gene in rat aortic smooth muscle cells (4); a similar result was also shown earlier in Hep3B cells (3). Bilirubin is a natural product of heme catabolism by HO. Here we demonstrated that there is an association between HO-1 promoter polymorphism and serum bilirubin levels, which are correlated with the development of CAD. The mean serum bilirubin level was significantly higher in carriers of the S allele than in those with the L/L genotype. In a previous case-control study of individuals with early familial CAD, higher serum bilirubin concentrations within the normal range were associated with a significant and marked reduction in CAD risk (7). In the prospective Framingham Offspring Study, higher serum bilirubin concentrations were associated with a decreased incidence of ischemic heart disease (8). Considering the antioxidant and antiatherogenic properties of bilirubin, the beneficial influence on serum bilirubin in carriers of the S allele might exert a protective effect against the development of CAD.

higher serum bilirubin concentrations were associated with a decreased incidence of ischemic heart disease (8). Considering the antioxidant and antiatherogenic properties of bilirubin, the beneficial influence on serum bilirubin in carriers of the S allele might exert a protective effect against the development of CAD. HO releases free ferrous (Fe2+) iron from heme. The toxic effect of free iron has been linked to oxidative stress through the Fenton reaction, in which Fe2+ oxidizes H2O2, leading to the generation of hydroxyl radicals (14), which in turn initiate lipid peroxidation. The amount of free ferrous iron is normally maintained at a very low level in humans. Of all the iron in the body (4 g), approximately two-thirds is found in association with hemoglobin in the ferrous form, and the majority of the remainder is stored as ferritin. In 1981, Sullivan (9) suggested that a state of iron depletion was potentially protective against coronary heart disease. Although the majority of animal research and the in vitro human studies support a role of iron in the pathogenesis of atherosclerosis, prospective human studies have provided inconsistent results in terms of clinical cardiovascular outcomes (10). Some investigators have hypothesized that iron may be primarily involved in the early stage of atherosclerosis, and focusing on cardiovascular morbidity and mortality (reflecting later stages of the disease) may not give insight into the potential mechanistic role of iron (15). Likewise, one recent study demonstrated that reduction of body iron stores by phlebotomy in patients with peripheral arterial disease produced a significant improvement in cardiovascular outcomes in patients aged <60 years but not in those at an older age (and thus with more advanced atherosclerosis) (16).

hat lead to the development of type 2 diabetes. Consistent with this, we previously demonstrated that the plasma glucose concentration at 1 h during the OGTT has a stronger correlation with surrogate measures of hepatic and muscle insulin resistance and β-cell dysfunction compared with the 2-h plasma glucose value (9). In summary, the plasma glucose concentration at 1 h during the OGTT is a strong predictor of future risk for type 2 diabetes. A cutoff point at 155 mg/dl plus the ATP III criteria for the metabolic syndrome can be used to stratify nondiabetic subjects into three risk groups—low, intermediate, and high risk—independent of the 2-h plasma glucose concentration.

f iron (15). Likewise, one recent study demonstrated that reduction of body iron stores by phlebotomy in patients with peripheral arterial disease produced a significant improvement in cardiovascular outcomes in patients aged <60 years but not in those at an older age (and thus with more advanced atherosclerosis) (16). In the present study, for the first time, we demonstrated that there is an association between HO-1 promoter polymorphism and serum ferritin concentrations, a measure of the body's iron stores, and an association between ferritin concentrations and the development of CAD in diabetic subjects. The mechanisms by which HO-1 polymorphism confers the variance in ferritin values remain to be elucidated. Nevertheless, a few animal studies and clinical data provided some indirect clues. A mouse model deficient in mammalian HO-1 (Hmox1) developed pathological accumulation of tissue iron stores associated with an increase in serum ferritin levels (17). HO-1 deficiency is very rare in humans. The first autopsy case of HO-1 deficiency was a 6-year-old boy who presented with growth retardation, anemia, elevated serum levels of ferritin and heme, low serum bilirubin concentrations, and hyperlipidemia. Fatty streaks and fibrous plaques were noted in his aorta (18). Moreover, treatment of healthy volunteers, patients with primary biliary cirrhosis, and patients with idiopathic hemochromatosis substantially with HO inhibitors increased serum ferritin concentrations (19). We hence postulated that the lower expression level of HO-1 imposed by the L allele under higher oxidative stress, as in the setting of diabetes, increases iron load in the vascular system, which may contribute to the development of atherosclerosis in such a virulent way.

s increased serum ferritin concentrations (19). We hence postulated that the lower expression level of HO-1 imposed by the L allele under higher oxidative stress, as in the setting of diabetes, increases iron load in the vascular system, which may contribute to the development of atherosclerosis in such a virulent way. The present study has strengths and limitations. Strengths include the large number of patients and the fact that all subjects had coronary arteriography and measures of bilirubin and ferritin. Furthermore, the homogeneous ethnic background possibly reduces variability in measurements. Among the study limitations is that the sample is primarily Chinese, making generalization to the other ethic groups uncertain. Furthermore, the present study design was cross-sectional, and we cannot infer causality. In summary, we have demonstrated that the microsatellite polymorphism in the promoter of HO-1 gene imposes modulation on serum bilirubin and ferritin levels, which might be associated with the development of CAD among diabetic subjects. Supplementary Material Online-Only Appendix

The prevalence of diabetes is increasing in Hispanic and Chinese Americans (1,2), groups comprised largely of immigrants. Immigration and subsequent behavior changes may contribute to the development of diabetes. Acculturation has been broadly defined as “the process by which individuals adopt the attitudes, values, customs, beliefs, and behaviors of another culture” (3). More recently, there has been recognition of the multidimensional aspects of acculturation (4) and the fact that the health effects of acculturation vary by country of origin and the health behavior or outcome being studied (5). Prior studies have suggested a relationship between acculturation, lifestyle behaviors, and other risk factors that may result in higher cardiovascular risk for immigrants in the U.S. (6,7). However, the associations between immigration, acculturation, and diabetes among U.S. immigrants have not been as well studied.

. Prior studies have suggested a relationship between acculturation, lifestyle behaviors, and other risk factors that may result in higher cardiovascular risk for immigrants in the U.S. (6,7). However, the associations between immigration, acculturation, and diabetes among U.S. immigrants have not been as well studied. Studies that have looked at the association between acculturation and diabetes have found differing results, depending on the immigrants’ country of origin. Among Japanese Americans, studies suggest that increasing acculturation is associated with higher diabetes risk (8,9). One study of Arab Americans found that a lack of acculturation is a risk factor for diabetes (10). Data on the association between acculturation and diabetes in Hispanics have not been consistent, and few studies have examined differences by country of origin (11). Understanding the consequences of acculturation for diabetes and its risks factors would have important implications for preventing diabetes in a large and growing portion of the U.S. population.

ion and diabetes in Hispanics have not been consistent, and few studies have examined differences by country of origin (11). Understanding the consequences of acculturation for diabetes and its risks factors would have important implications for preventing diabetes in a large and growing portion of the U.S. population. The main objective of this study was to examine the hypothesis that diabetes prevalence among Hispanic and Chinese participants in the Multi-Ethnic Study of Atherosclerosis (MESA) differs by acculturation status. Based on prior studies showing that acculturation is associated with greater BMI among Asians and Hispanics (12,13), we hypothesized that greater acculturation would be associated with a higher diabetes prevalence among Hispanics and Chinese in MESA and that BMI would be part of the mechanism. We also explored the roles of physical activity and diet in mediating this association and examined whether associations between acculturation and diabetes differed by race/ethnicity and country of origin among Hispanics.

abetes prevalence among Hispanics and Chinese in MESA and that BMI would be part of the mechanism. We also explored the roles of physical activity and diet in mediating this association and examined whether associations between acculturation and diabetes differed by race/ethnicity and country of origin among Hispanics. RESEARCH DESIGN AND METHODS Data source We used cross-sectional data from MESA, a 10-year longitudinal study with the goal of identifying risk factors for subclinical atherosclerosis and transition from subclinical disease to clinical events (14). The MESA cohort includes 6,814 men and women aged 45–84 years at baseline who were recruited from six field centers: Baltimore, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles, California; New York, New York; and St. Paul, Minnesota. Only individuals free of clinical cardiovascular disease at baseline were eligible. Approximately 40% of the cohort are non-Hispanic white, 30% are non-Hispanic black, 20% are Hispanic, and 10% are Chinese. Only Hispanic and Chinese participants were included in this study because the non-Hispanic white and black groups had very few immigrants and little variation in acculturation. Cubans, Puerto Ricans, and other Hispanics were represented at four of six field centers, whereas Mexican-origin Hispanics were located at three of the field centers. All Dominicans were located at a single MESA field center (New York). Chinese participants were recruited from Los Angeles and Chicago. The baseline visit for the cohort took place between July 2000 and September 2002.

ur of six field centers, whereas Mexican-origin Hispanics were located at three of the field centers. All Dominicans were located at a single MESA field center (New York). Chinese participants were recruited from Los Angeles and Chicago. The baseline visit for the cohort took place between July 2000 and September 2002. Dependent and independent variables Data were collected in a standardized manner at all study sites by trained personnel; blood assays were processed at central laboratories (14). Questionnaires were administered as part of the baseline visit in English, Spanish, or Chinese. Questionnaires were translated by certified translators and reviewed by bilingual study investigators, staff at different sites, and a multicultural research office at one of the sites.

rocessed at central laboratories (14). Questionnaires were administered as part of the baseline visit in English, Spanish, or Chinese. Questionnaires were translated by certified translators and reviewed by bilingual study investigators, staff at different sites, and a multicultural research office at one of the sites. The main dependent variable in this analysis was diabetes, which was defined as fasting glucose ≥126 mg/dl and/or use of antidiabetes medications, a definition based on the 2003 American Diabetes Association criteria (15). Our main independent variable was acculturation score. MESA has information on three crude proxies of acculturation: nativity, language spoken at home, and years in the U.S. Nativity was categorized as U.S. born or foreign born. U.S.-born individuals were those who were born in the U.S. All others (including individuals born in Puerto Rico) were classified as foreign born. Language spoken at home was categorized as speaks English only, speaks English and Chinese or English and Spanish, or only speaks a non-English language at home. Among the foreign born, years in the U.S. was categorized as living in the U.S. ≥20 years, living in the U.S. 10–19 years, and living in the U.S. <10 years.

. Language spoken at home was categorized as speaks English only, speaks English and Chinese or English and Spanish, or only speaks a non-English language at home. Among the foreign born, years in the U.S. was categorized as living in the U.S. ≥20 years, living in the U.S. 10–19 years, and living in the U.S. <10 years. We constructed an acculturation score for each participant based on these proxy markers. A score of 0–3 was assigned for nativity combined with years in the U.S. (3 = U.S. born, 2 = foreign born and lived in the U.S. ≥20 years, 1 = foreign born and lived in the U.S. 10–19 years, and 0 = foreign born and lived in the U.S. <10 years). A score of 0–2 was assigned to language spoken at home (2 = English, 1 = English and Chinese or English and Spanish, and 0 = non-English languages). These scores were summed to obtain the acculturation score, ranging from 0 (least acculturated) to 5 (most acculturated). We used the summary acculturation score, rather than the individual variables because a single acculturation score takes into account the fact that these characteristics are often clustered within an individual and their combination may give a more accurate representation of acculturation than each indicator independently. Mexican-origin Hispanics were categorized into four groups based on the distribution of the summary acculturation score: scores of 0–1, 2, 3–4, and 5. Because far fewer non–Mexican-origin Hispanic and Chinese participants were highly acculturated, the acculturation score was collapsed into three categories in these groups (0–1, 2, and 3–5 for non–Mexican-origin Hispanics and 0, 1, 2, and 3–5 for Chinese).

the summary acculturation score: scores of 0–1, 2, 3–4, and 5. Because far fewer non–Mexican-origin Hispanic and Chinese participants were highly acculturated, the acculturation score was collapsed into three categories in these groups (0–1, 2, and 3–5 for non–Mexican-origin Hispanics and 0, 1, 2, and 3–5 for Chinese). Sociodemographic covariates included race/ethnicity, age, sex, and socioeconomic status (SES). Race/ethnicity was based on participants’ responses to the ethnicity and race questions included in the year 2000 U.S. census. If a participant self-identified as Hispanic, he or she was then asked, “which of the following best describes you (you may choose from more than one group)?” Participants could choose from Mexican, Chicano, Mexican American, Dominican, Puerto Rican, Cuban, or Other (asked to specify). Mexican, Chicano, and Mexican-American subjects were all classified as Mexican-origin Hispanics, and the rest were categorized as non–Mexican-origin Hispanics for this analysis. Based on self-reported subgroup, our sample included 708 Mexican-origin Hispanics and 547 non–Mexican-origin Hispanics. Among the non–Mexican origin Hispanics, 131 were Dominicans, 157 were Puerto Rican, 239 were from South/Central America, and 47 were Cuban.

tegorized as non–Mexican-origin Hispanics for this analysis. Based on self-reported subgroup, our sample included 708 Mexican-origin Hispanics and 547 non–Mexican-origin Hispanics. Among the non–Mexican origin Hispanics, 131 were Dominicans, 157 were Puerto Rican, 239 were from South/Central America, and 47 were Cuban. SES was measured by income and education. Participants were asked to select their family income from a list of 13 categories and education from a list of 8 categories; these were collapsed into fewer categories for our analysis. The questionnaire also inquired whether participants used a primary care clinic, emergency room, or another place for routine health care services. In addition, health insurance status was ascertained (private health insurance, HMO, Medicaid, Medicare, veteran's health care, or none).

fewer categories for our analysis. The questionnaire also inquired whether participants used a primary care clinic, emergency room, or another place for routine health care services. In addition, health insurance status was ascertained (private health insurance, HMO, Medicaid, Medicare, veteran's health care, or none). Behavioral factors were also considered. BMI (weight in kilograms divided by the square of height in meters) was measured and used as a continuous variable. Physical activity was self-reported using a semiquantitative questionnaire adapted from the Cross-Cultural Activity Participation Study (14). For the purposes of this study, physical activity was defined as the number of MET minutes per week spent doing intentional leisure-time exercise. We used leisure-time exercise because this variable appeared to be a representation of physical activity that was well defined, is readily understood, and has been associated with physiological measures (16). Total dietary calories (kilocalories per day), carbohydrates (grams per day), fat (grams per day), and fiber (grams per day) were estimated from the MESA food frequency questionnaire, which was modified from the Insulin Resistance Atherosclerosis study in which comparable validity was observed for non-Hispanic white, African American, and Hispanic individuals (14). These food frequency questionnaires were modified to include foods typically eaten by Chinese individuals (14).

frequency questionnaire, which was modified from the Insulin Resistance Atherosclerosis study in which comparable validity was observed for non-Hispanic white, African American, and Hispanic individuals (14). These food frequency questionnaires were modified to include foods typically eaten by Chinese individuals (14). Statistical analysis Participant characteristics by acculturation score were compared using ANOVA for continuous variables and a χ2 test for categorical variables. A test for linear trend was performed using linear regression (continuous variables) and the Cochran-Armitage test (binary variables). Relative risk regression was used to estimate the prevalence ratio of diabetes associated with acculturation for Mexican-origin Hispanics, non–Mexican-origin Hispanics, and Chinese separately, with adjustment for potential confounders or mediators. That is, the relative prevalence of diabetes was modeled as a function of acculturation score (entered as dummy variables) using a generalized linear model with log link and binomial error distribution. In cases in which the model failed to converge with the log-binomial model, a Poisson model was used, and robust error variances were estimated (17). In model 1, adjustments were made for age and sex. Model 2 included the variables in model 1 plus SES. To investigate potential mediators between acculturation and diabetes, models were fitted by adding to the variables in model 2: BMI (model 3), diet (model 4), and physical activity (model 5). Model 6 included all variables in model 2 plus BMI, diet variables (total calories in kilocalories, total fat in percent kilocalories, total carbohydrate in percent kilocalories, total fiber in grams per 1,000 calories), and physical activity in MET minutes per week. Interactions between acculturation score (dummy variables) and sex were tested separately for Mexican-origin Hispanics, non–Mexican-origin Hispanics, and Chinese by including cross-product terms in the regression models along with age and SES. No interactions were statistically significant at P = 0.05. All analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC).

e tested separately for Mexican-origin Hispanics, non–Mexican-origin Hispanics, and Chinese by including cross-product terms in the regression models along with age and SES. No interactions were statistically significant at P = 0.05. All analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC). RESULTS Of the 2,299 Hispanic and Chinese MESA participants, 1,992 remained for analyses: 147 were excluded because of missing nutrient data, 4 because of missing diabetes information, and 2 because of missing other critical data; 1 could not be classified with respect to language spoken at home; and 153 were missing data on years in the U.S. Of the Mexican-origin Hispanics, 53% were U.S. born, whereas only 10% of non–Mexican-origin Hispanics and 4% of Chinese were U.S. born (P < 0.001) (Table 1). Non–Mexican-origin Hispanics were more likely to speak Spanish at home than Mexican-origin Hispanics. Nearly 90% of Chinese participants spoke Chinese at home. Thirty-nine percent of Mexican-origin Hispanics, 70% of non–Mexican-origin Hispanics, and 88% of Chinese had low acculturation (acculturation score of 0–1 or 2). Chinese participants had slightly higher incomes and were more highly educated than both Hispanic groups (P < 0.001). The prevalence of diabetes varied significantly: 21% of Mexican-origin Hispanics, 14% of non–Mexican-origin Hispanics, and 13% of Chinese participants had diabetes (P < 0.001).

ation score of 0–1 or 2). Chinese participants had slightly higher incomes and were more highly educated than both Hispanic groups (P < 0.001). The prevalence of diabetes varied significantly: 21% of Mexican-origin Hispanics, 14% of non–Mexican-origin Hispanics, and 13% of Chinese participants had diabetes (P < 0.001). As expected, Mexican-origin Hispanics, non–Mexican-origin Hispanics, and Chinese participants with higher acculturation had greater incomes, education, and health insurance coverage (P < 0.001 for all variables within each ethnic group) (Table 2). Among Mexican-origin Hispanics, the prevalence of diabetes was lowest (19.5%) in the most acculturated group (acculturation score = 5); however, the overall trend was not significant. Among non–Mexican-origin Hispanics, the prevalence of diabetes was greater among the groups with higher acculturation (16% in those with an acculturation score of 2 and 14% in those with a score of 3–5) compared with those in the least acculturated group (7%) (P for trend 0.072). Among Chinese, there was no trend in diabetes prevalence by acculturation. Higher acculturation was associated with a higher BMI in Mexican-origin Hispanics (P = 0.019), non–Mexican-origin Hispanics (P = 0.053), and Chinese (P < 0.001). Highly acculturated Mexican-origin Hispanics, non–Mexican-origin Hispanics, and Chinese also reported significantly more physical activity than those in the lower acculturation groups. Among Mexican-origin Hispanics, higher acculturation was associated with consuming significantly fewer calories (P < 0.001). Among the Chinese participants, greater acculturation was associated with consuming more calories, more carbohydrates, and less fat (P < 0.001).

vity than those in the lower acculturation groups. Among Mexican-origin Hispanics, higher acculturation was associated with consuming significantly fewer calories (P < 0.001). Among the Chinese participants, greater acculturation was associated with consuming more calories, more carbohydrates, and less fat (P < 0.001). Among Hispanics, associations between acculturation and diabetes differed by country of origin (P for interaction 0.03). Among Mexican-origin Hispanics, there was no clear association between acculturation levels and diabetes prevalence (Table 3). In contrast, among non–Mexican-origin Hispanics, the highest acculturated group had a higher prevalence of diabetes (prevalence rate [PR] 2.49 [95% CI 1.14–5.44]) than those in the least acculturated group, independent of sociodemographics (Table 3, model 2). This association was slightly reduced after additional adjustment for BMI (2.08 [0.97–4.47]) (Table 3, model 3). Adjustment for diet had a similar effect as adjustment for BMI (SES- and diet-adjusted PR 2.08 [0.97–4.47] for highest versus lowest acculturation group). Adjustment for physical activity did not modify estimates adjusted for SES. Associations between acculturation and BMI were further reduced with adjustments for BMI, diet, and physical activity (1.59 [0.75–3.39] for highest versus lowest acculturation category) (Table 3, model 6). Among Chinese participants, there was no significant association between acculturation score and diabetes prevalence.

ociations between acculturation and BMI were further reduced with adjustments for BMI, diet, and physical activity (1.59 [0.75–3.39] for highest versus lowest acculturation category) (Table 3, model 6). Among Chinese participants, there was no significant association between acculturation score and diabetes prevalence. CONCLUSIONS We have found that higher levels of acculturation are associated with greater prevalence of diabetes in non–Mexican-origin Hispanics aged 45–84 years who are free of clinical cardiovascular disease. This association was only partly mediated by BMI or diet. In contrast, acculturation was not associated with an increased or decreased risk of diabetes prevalence among Chinese and Mexican-origin Hispanic MESA participants.

n non–Mexican-origin Hispanics aged 45–84 years who are free of clinical cardiovascular disease. This association was only partly mediated by BMI or diet. In contrast, acculturation was not associated with an increased or decreased risk of diabetes prevalence among Chinese and Mexican-origin Hispanic MESA participants. Data on the association between acculturation and diabetes in Hispanics have not been consistent, and few studies have examined differences by country of origin. In the San Antonio Heart Study, higher acculturation was associated with a significantly lower prevalence of obesity and diabetes among Mexican American women and men, independent of SES (18). Two other studies, based on data from the National Health and Nutrition Examination Survey (6) and data from the Hispanic Health and Nutrition Examination Survey (19), showed that the prevalence of diabetes was greater among Mexican-Americans in the middle group of acculturation (19). Only one study of which we are aware reported that a higher level of acculturation (as measured by language and country of origin) was associated with higher diabetes prevalence in Mexican Hispanics, after adjustment for age and sex (20). Our results are therefore consistent with those of most researchers who have looked at acculturation and diabetes among Mexican Americans and have found either no association or lower diabetes in more acculturated individuals.

with higher diabetes prevalence in Mexican Hispanics, after adjustment for age and sex (20). Our results are therefore consistent with those of most researchers who have looked at acculturation and diabetes among Mexican Americans and have found either no association or lower diabetes in more acculturated individuals. In contrast to results for Mexican-origin Hispanics we found that higher acculturation levels may be a risk factor for diabetes in non–Mexican-origin Hispanic groups. Very few studies have examined effects of acculturation on diabetes in non–Mexican-origin Hispanics. Among the Hispanics in the Hispanic Health and Nutrition Examination Survey, Mexicans and Puerto Ricans had a higher prevalence of diabetes than Cubans (21). There was no significant association between acculturation and diabetes prevalence; however, the results are not reported by Hispanic subgroup (21).

in Hispanics. Among the Hispanics in the Hispanic Health and Nutrition Examination Survey, Mexicans and Puerto Ricans had a higher prevalence of diabetes than Cubans (21). There was no significant association between acculturation and diabetes prevalence; however, the results are not reported by Hispanic subgroup (21). Several factors could explain differences in the association between acculturation and diabetes in the Mexican-origin and non–Mexican-origin Hispanics in MESA. Prior studies suggested that the behavioral consequences of acculturation differ for Hispanic subgroups because of differences in social and cultural context, the reasons for immigration, and connection to the country of origin (5). We did find that acculturation had a different relationship with BMI and dietary intake across subgroups. Greater acculturation was associated with greater BMI in Hispanics (especially in Mexican-origin Hispanics). Higher acculturation in Mexican-origin Hispanics was associated with a diet that was significantly lower in calories, but this was not true for non–Mexican-origin Hispanics. Our findings are consistent with a growing body of evidence that there may be significant heterogeneity in the association between acculturation, health behaviors, and chronic disease prevalence.

was associated with a diet that was significantly lower in calories, but this was not true for non–Mexican-origin Hispanics. Our findings are consistent with a growing body of evidence that there may be significant heterogeneity in the association between acculturation, health behaviors, and chronic disease prevalence. Studies in Japanese and Chinese Americans show a more consistent relationship between acculturation and diabetes prevalence, with a higher diabetes prevalence among Asians who are acculturated to a more Western lifestyle (9,22,23). Asian Americans may be more sensitive than Hispanic populations to the changes that occur with acculturation, such as increasing BMI. For example, some Asian groups seem to develop diabetes and glucose intolerance at a lower BMI than other racial/ethnic minorities (2). However, we did not find an association between acculturation and diabetes in Chinese participants. This finding may reflect the lack of variability in acculturation among the Chinese in MESA.

ple, some Asian groups seem to develop diabetes and glucose intolerance at a lower BMI than other racial/ethnic minorities (2). However, we did not find an association between acculturation and diabetes in Chinese participants. This finding may reflect the lack of variability in acculturation among the Chinese in MESA. Acculturation to a Western lifestyle is associated with higher BMI (13), which in turn is associated with a greater risk of diabetes (21). Adjustment for BMI or diet partially attenuated the relationship between acculturation and diabetes observed in non–Mexican-origin Hispanics. However, a substantial increased risk associated with acculturation remained after adjustment for these variables, although it was not statistically significant. Adjustment for physical activity did not significantly change the association between diabetes and acculturation from the age-, sex-, and SES-adjusted estimates. Prior studies have shown a lower prevalence of diabetes in more physically active populations (24). The lack of association between physical activity and diabetes in this study may be due to limitations of the physical activity measure, which only included leisure-time activity. In any case, the fact that physical activity levels were actually greater in more acculturated than in less acculturated Hispanics implies that the type of physical activity we investigated (leisure-time) is not a mediator of any acculturation effects on diabetes. The fact that associations of acculturation with diabetes remained after adjustment (although they were not statistically significant) suggests that other mediators, including stress-related processes implicated in the development of diabetes (25), need to be investigated.

ator of any acculturation effects on diabetes. The fact that associations of acculturation with diabetes remained after adjustment (although they were not statistically significant) suggests that other mediators, including stress-related processes implicated in the development of diabetes (25), need to be investigated. There are several limitations to this study. This is a cross-sectional analysis, which limits causal inferences; although it is unlikely that diabetes leads to acculturation, a diagnosis of diabetes may lead to changes in some of the behavioral variables associated with diabetes, such as diet and physical activity. The majority of studies on health and acculturation vary in how they measure acculturation, and this variation may account for different results across studies. In this study, we used nativity, language, and years in the U.S. as proxies for acculturation, and these variables do not fully capture the complex process of acculturation and its health effects. The MESA sample is not a nationally representative sample, and it is unclear whether these findings can be generalized to other populations in the U.S. Because of sample size limitations, we were unable to further separate out the non–Mexican Hispanics by country of origin and were also unable to examine whether the association between acculturation and diabetes varied by sex. Among Chinese participants, there was little variability in acculturation; this limited the statistical power to detect any meaningful association between acculturation and diabetes. Future studies should explore how acculturation and its health effects vary across different racial/ethnic groups and countries of origin with larger sample sizes.

, there was little variability in acculturation; this limited the statistical power to detect any meaningful association between acculturation and diabetes. Future studies should explore how acculturation and its health effects vary across different racial/ethnic groups and countries of origin with larger sample sizes. In this study, we found an association between higher acculturation levels and diabetes prevalence in middle-aged and elderly non–Mexican-origin Hispanics. This risk was partly explained by the higher BMI or higher calorie diet associated with acculturation. Acculturation should be considered when risk factors for diabetes in immigrant populations are studied. Adequate investigation of acculturation will require the development of valid instruments in different subgroups of the population. We thank the other investigators, the staff, and the participants in MESA for their valuable contributions. An earlier version of this article was presented at the American Heart Association Scientific Sessions 2006, 12–15 November 2006, Chicago, Illinois.

Left ventricular hypertrophy (LVH), a cardinal manifestation of preclinical cardiovascular disease, strongly predicts myocardial infarction, stroke, and cardiovascular death in patients with hypertension (1) or coronary artery disease (2), as well as in the general population (3). In the Framingham Study, electrocardiographic (ECG) evidence of LVH (ECG-LVH) was associated with a twofold increase in mortality over that resulting from hypertension alone (4).

stroke, and cardiovascular death in patients with hypertension (1) or coronary artery disease (2), as well as in the general population (3). In the Framingham Study, electrocardiographic (ECG) evidence of LVH (ECG-LVH) was associated with a twofold increase in mortality over that resulting from hypertension alone (4). Studies have consistently shown that antihypertensive therapy may effectively limit the incidence of ECG-LVH, regardless of the treatments used to reduce blood pressure (5). However, the Heart Outcomes Prevention Education (HOPE) (6) and the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) trials (7) showed that, in patients with ECG-LVH at inclusion, the ACE inhibitor ramipril and the angiotensin receptor blocker (ARB) losartan, respectively, regressed LVH more effectively than drugs that do not directly interfere with the renin-angiotensin-aldosterone system (RAAS). The finding in both trials that this benefit was significant even after adjustments for the small differences in blood pressure between the two treatment groups provided consistent evidence that RAAS inhibitor therapy has a specific cardioprotective effect that exceeds expectations based on changes in blood pressure alone. However, the HOPE trial (6) was not powered to assess the treatment effect on new-onset LVH in the subgroup with no ECG-LVH at baseline, and the LIFE trial (7) included only patients with LVH. Thus, whether RAAS inhibitor therapy may also prevent new-onset LVH in subjects with normal left ventricular mass to start with is unknown. To formally explore this issue, we compared the effect of ACE versus non-ACE inhibitor therapy on incident ECG-LVH in patients from the Bergamo Nephrologic Diabetes Complications Trial (BENEDICT) (8–11) who had no ECG-LVH at inclusion.

ew-onset LVH in subjects with normal left ventricular mass to start with is unknown. To formally explore this issue, we compared the effect of ACE versus non-ACE inhibitor therapy on incident ECG-LVH in patients from the Bergamo Nephrologic Diabetes Complications Trial (BENEDICT) (8–11) who had no ECG-LVH at inclusion. RESEARCH DESIGN AND METHODS BENEDICT (8) was a prospective, randomized, double-blind, parallel group study that evaluated the possibility of preventing the onset of persistent microalbuminuria in 1,204 patients with type 2 diabetes (World Health Organization criteria) and arterial hypertension (systolic or diastolic blood pressure >130 or 85 mmHg or concomitant antihypertensive therapy) but normal urinary albumin excretion rate (<20 μg/min in at least two of three consecutive overnight urine collections) who were randomly assigned to at least 3 years of treatment with one of the following study drugs: 1) a nondihydropyridine calcium channel blocker (240 mg/day verapamil SR), 2) an ACE inhibitor (2 mg/day trandolapril), 3) a fixed-dose combination (180 mg/day verapamil SR plus 2 mg/day trandolapril [VeraTran]), and 4) placebo. The target blood pressure after random assignment and throughout the whole study period was <130/80 mmHg for all of the treatment groups. Other antihypertensive drugs (with the exception of RAAS inhibitors and nondihydropyridine calcium channel blockers different from the study drugs) could be used to achieve and maintain the target blood pressure according to predefined guidelines (8). The study protocol was in accordance with the Declaration of Helsinki and was approved by the institutional review board at each center and by the safety committee of BENEDICT. All patients gave written informed consent. Patients were eligible to enter the ECG-LVH substudy if they had no ECG evidence of LVH at baseline and had at least 1 year of follow-up.

ith the Declaration of Helsinki and was approved by the institutional review board at each center and by the safety committee of BENEDICT. All patients gave written informed consent. Patients were eligible to enter the ECG-LVH substudy if they had no ECG evidence of LVH at baseline and had at least 1 year of follow-up. Aims The primary aim of analyses was to compare the incidence of ECG-LVH in patients randomly assigned to ACE inhibitor or non-ACE inhibitor therapy who had no ECG evidence of LVH at baseline. Secondarily, we evaluated the relationships between incidence of ECG-LVH and baseline and follow-up variables, including treatable risk factors such as blood pressure and A1C. ECG-LVH and other outcome variables The main outcome variable was ECG-LVH defined as Sokolow-Lyon (SV1 + RV5/6) voltage ≥3.5 mV (12) and/or Cornell (RaVL + SV3) voltage ≥2.0 mV (women) or ≥2.4 mV (men) (13). Secondary outcomes were Sokolow-Lyon and Cornell voltages considered as continuous variables. Standard 12-lead ECGs were recorded at 25 mm/s and 1 mV/cm calibration at baseline and every year thereafter. They were centrally and independently evaluated by two investigators who were blinded to treatment allocation and patient data. ECGs with inconsistent readings were evaluated by a third independent cardiologist, and his diagnosis was considered for data analysis. Trough systolic and diastolic (Korotkoff phase I/V) blood pressure was measured in the morning before treatment administration (8).

re blinded to treatment allocation and patient data. ECGs with inconsistent readings were evaluated by a third independent cardiologist, and his diagnosis was considered for data analysis. Trough systolic and diastolic (Korotkoff phase I/V) blood pressure was measured in the morning before treatment administration (8). Data were reported in dedicated case report forms and doubly entered in an ad hoc database that was eventually merged with the BENEDICT database. Before analyses, all data were monitored by the Monitoring Unit of the Clinical Research Center for Rare Diseases “Aldo & Cele Daccò” of the Mario Negri Institute for Pharmacological Research.

in dedicated case report forms and doubly entered in an ad hoc database that was eventually merged with the BENEDICT database. Before analyses, all data were monitored by the Monitoring Unit of the Clinical Research Center for Rare Diseases “Aldo & Cele Daccò” of the Mario Negri Institute for Pharmacological Research. Sample size We assumed a 10% incidence of ECG-LVH in the non-ACE inhibitor group and a 60% reduction (from 10 to 4%) in the ACE inhibitor group. The expected incidence of ECG-LVH was assumed to be higher than that in the HOPE control subjects (6) because, different from the HOPE trial, all BENEDICT patients were hypertensive and diabetic at inclusion (8), and they also had additional risk factors such as older age, systolic hypertension, obesity, and, conceivably, insulin resistance. The 60% risk reduction was assumed on the basis of experimental evidence that ACE inhibition fully prevents LVH if treatment is started before the induction of arterial hypertension (14). Because our patients were already hypertensive at study entry and had other risk factors, we considered a conservative assumption of 60% risk reduction as appropriate. Thus, we calculated that 400 patients per group gave the study a 90% power to detect as statistically significant (α = 0.05, two-tailed test) the expected between-group difference in incidence of ECG-LVH.

udy entry and had other risk factors, we considered a conservative assumption of 60% risk reduction as appropriate. Thus, we calculated that 400 patients per group gave the study a 90% power to detect as statistically significant (α = 0.05, two-tailed test) the expected between-group difference in incidence of ECG-LVH. Statistical analyses The analyses were performed by the Laboratory of Biostatistics of the Clinical Research Center. Between-group comparisons were performed on continuous variables by unpaired t test or Wilcoxon's rank-sum test and on categorical variables by a χ2 test or Fisher's exact test. Within-group comparisons were performed on continuous variables by paired t test or Wilcoxon's signed-rank test and on categorical variables by the McNemar test. Predefined (8) baseline covariates were considered: site, age, sex, smoking status (patients who had never smoked versus former smokers and current smokers), diastolic blood pressure, and log-transformed urinary albumin excretion (median of three readings) at baseline.

and on categorical variables by the McNemar test. Predefined (8) baseline covariates were considered: site, age, sex, smoking status (patients who had never smoked versus former smokers and current smokers), diastolic blood pressure, and log-transformed urinary albumin excretion (median of three readings) at baseline. The main study results were reported by a Cox regression model. For graphic representation, Kaplan-Meier curves were plotted for each group considered. Exploratory analyses were also conducted, including follow-up systolic and diastolic blood pressure measurement and absolute reductions from baseline for follow up of blood pressure measurement to help data interpretation. All statistical analyses were performed using SAS (version 9.1; SAS Institute, Cary, NC). P < 0.05 was considered statistically significant. No P value adjustment was carried out for multiple comparisons. RESULTS Of 905 patients with readable ECG at baseline and at least 1 year of follow-up, 816 (433 receiving ACE and 383 receiving non-ACE inhibitor therapy) had no ECG evidence of LVH. Of these, 799 patients (423 receiving ACE and 376 receiving non-ACE inhibitor therapy) were available for analyses (see study profile in Online Appendix 2 [available at http://dx.doi.org/10.2337/dc08-0371]). Baseline characteristics and Sokolow-Lyon and Cornell voltages were similar between treatment groups with the exception of the percentage of never smokers and of serum cholesterol levels that were higher in the non-ACE inhibitor group (Table 1).

in Online Appendix 2 [available at http://dx.doi.org/10.2337/dc08-0371]). Baseline characteristics and Sokolow-Lyon and Cornell voltages were similar between treatment groups with the exception of the percentage of never smokers and of serum cholesterol levels that were higher in the non-ACE inhibitor group (Table 1). Incidence of ECG-LVH Over a median (interquartile range) of 36 (24–48) months of follow-up, LVH developed in 44 patients (5.5%), 13 (3.1%) receiving ACE and 31 (8.2%) receiving non-ACE inhibitor therapy (Fig. 1). The unadjusted hazard ratio (HR) [95% CI] for ECG-LVH was 0.34 [0.18–0.65] (P = 0.0012). The HR (0.35 [0.18–0.68]) was statistically significant even after adjustment for predefined baseline characteristics (P = 0.0018) and baseline and follow-up systolic and diastolic blood pressure, as well as systolic and diastolic blood pressure reduction versus baseline (Table 2). Compared with baseline, both Sokolow-Lyon and Cornell voltages significantly decreased at different years on follow-up in the study group as a whole and in the subgroup receiving ACE inhibitor therapy (Table 3). In the non-ACE inhibitor group, changes in Sokolow-Lyon voltage were not significant, and changes in Cornell voltage achieved statistical significance only at 2 and 3 years (Table 3).

y decreased at different years on follow-up in the study group as a whole and in the subgroup receiving ACE inhibitor therapy (Table 3). In the non-ACE inhibitor group, changes in Sokolow-Lyon voltage were not significant, and changes in Cornell voltage achieved statistical significance only at 2 and 3 years (Table 3). Blood pressure and metabolic control Follow-up systolic (138.4 ± 9.4 mmHg) and diastolic (80.8 ± 5.2 mmHg) blood pressure was 12.7 and 7.1 mmHg lower than at baseline, respectively (P < 0.0001 for both). Follow-up systolic (137.2 ± 9.2 vs. 139.7 ± 9.5 mmHg, respectively, P = 0.045) and diastolic (80.0 ± 5.2 vs. 81.7 ± 5.0 mmHg, P = 0.005) blood pressure was lower in the ACE than in the non-ACE inhibitor group. Follow-up A1C levels were lower in the ACE than in the non-ACE inhibitor group (5.76 ± 1.17 vs 5.91 ± 1.23%, P = 0.03), whereas blood glucose was similar in the two treatment groups (156.1 ± 36.2 vs. 161.3 ± 39.1 mg/dl, P = 0.31) (see blood pressure, A1C, and blood glucose profiles in Online Appendix 3 [available at http://dx.doi.org/10.2337/dc08-0371]).

he ACE than in the non-ACE inhibitor group (5.76 ± 1.17 vs 5.91 ± 1.23%, P = 0.03), whereas blood glucose was similar in the two treatment groups (156.1 ± 36.2 vs. 161.3 ± 39.1 mg/dl, P = 0.31) (see blood pressure, A1C, and blood glucose profiles in Online Appendix 3 [available at http://dx.doi.org/10.2337/dc08-0371]). In multivariable analyses, known diabetes duration, baseline BMI, and Sokolow-Lyon and Cornell voltages were associated with the incidence of ECG-LVH on follow-up, whereas baseline and follow-up systolic and diastolic blood pressure, A1C, and blood glucose as well as systolic and diastolic blood pressure, A1C, and blood glucose changes versus baseline had no predictive value. No significant correlation was found between changes in Sokolow-Lyon or Cornell voltages on follow-up and concomitant changes in blood pressure, A1C, and blood glucose (versus baseline).

ood glucose as well as systolic and diastolic blood pressure, A1C, and blood glucose changes versus baseline had no predictive value. No significant correlation was found between changes in Sokolow-Lyon or Cornell voltages on follow-up and concomitant changes in blood pressure, A1C, and blood glucose (versus baseline). Comparative analyses between randomization arms LVH developed in 4 patients receiving trandolapril (1.9%), 9 receiving VeraTran (4.2%), 16 receiving verapamil (8.9%), and 15 receiving placebo (7.6%). After adjustment for predefined covariates, the risk of LVH was significantly lower with trandolapril than with verapamil (HR 0.22 [95% CI 0.07–0.65)], P = 0.007) or placebo (0.25 [0.08–0.78], P = 0.017) and with VeraTran compared with verapamil (0.42 [0.19–0.97], P = 0.0.043). Risk reduction with VeraTran compared with placebo was not significant (0.49 [0.21–1.16], P = 0.10). Risk was not significantly different between VeraTran and trandolapril (1.97 ([0.60–6.49], P = 0.26) and between verapamil and placebo (1.21 [0.58–2.52], P = 0.61). CONCLUSIONS In the present study, ACE inhibition with trandolapril significantly reduced the incidence of ECG-LVH in patients with arterial hypertension and type 2 diabetes compared with non-ACE inhibitor therapy. The protective effect of trandolapril against ECG-LVH was already evident at 1 year after random assignment and progressively increased on follow-up. Sokolow-Lyon and Cornell voltages consistently decreased with trandolapril therapy, whereas they did not change appreciably with non-ACE inhibitor therapy.

bitor therapy. The protective effect of trandolapril against ECG-LVH was already evident at 1 year after random assignment and progressively increased on follow-up. Sokolow-Lyon and Cornell voltages consistently decreased with trandolapril therapy, whereas they did not change appreciably with non-ACE inhibitor therapy. The reduced incidence of ECG-LVH with trandolapril was not explained by the small differences in blood pressure control between the two treatment groups, because risk reduction was highly significant even after adjustments were made for blood pressure control achieved on follow-up and for blood pressure reductions observed versus baseline. Baseline factors potentially involved in LVH development and progression, such as age, BMI, blood pressure, and Sokolow-Lyon and Cornell voltages were similar in the two groups. A low-salt diet was prescribed for all patients, and no patient performed vigorous physical activities. Altogether, the above data indicate that trandolapril had a specific protective effect against the development of LVH that was additional to the benefit of blood pressure reduction.

es were similar in the two groups. A low-salt diet was prescribed for all patients, and no patient performed vigorous physical activities. Altogether, the above data indicate that trandolapril had a specific protective effect against the development of LVH that was additional to the benefit of blood pressure reduction. Both hemodynamic and nonhemodynamic factors most likely contributed to the cardioprotective effect of trandolapril therapy. ACE inhibitors increase vessel wall compliance and reduce arterial wave reflection amplitude and thus reduce aortic and left ventricular blood pressure even more consistently than peripheral artery blood pressure, that is, the brachial blood pressure as measured at the arm (15). Thus, at comparable peripheral blood pressure, ACE inhibitors may reduce central pressures and left ventricular afterload more effectively than antihypertensive drugs that do not directly interfere with the RAAS (16). These hemodynamic effects probably contributed to the regression of LVH observed with RAAS inhibitor therapy in the HOPE (6) and LIFE trials (7) and might also explain why in BENEDICT trandolapril prevented LVH more effectively than non-ACE inhibitor therapy, even at comparable brachial blood pressure (9).

e RAAS (16). These hemodynamic effects probably contributed to the regression of LVH observed with RAAS inhibitor therapy in the HOPE (6) and LIFE trials (7) and might also explain why in BENEDICT trandolapril prevented LVH more effectively than non-ACE inhibitor therapy, even at comparable brachial blood pressure (9). ACE inhibitors may prevent LVH also through direct inhibition of cardiac RAAS. Angiotensin II promotes the growth of myocytes independently of loading conditions (17), and ACE inhibitors may prevent the hypertrophic effect of angiotensin II even at doses that do not affect the blood pressure (14). Renin expression and tissue-converting enzyme activity are increased in hypertrophied hearts of spontaneously hypertensive rats (18,19), and angiotensin II type 1 (AT1) receptors are overexpressed in cardiomyocytes of mice with cardiac hypertrophy and normal blood pressure (20). Consistently, in rats, cardiac expression of the AT1 receptor antisense transgene attenuated cardiac hypertrophy without affecting the blood pressure (21). However, recent data that cardiac AT1 receptors cannot sustain angiotensin II-dependent hypertension and cardiac hypertrophy after knocking-out renal AT1 receptors (22) suggest that kidney RAAS activation is also involved and that its inhibition may explain at least part of the cardioprotective effect of ACE inhibitors and ARBs. Unlike ARBs, ACE inhibitors may directly prevent myocardial hypertrophy also by increasing local bradykinin bioavailability through inhibition of the myocardial kallikrein-kinin pathway (23).

on is also involved and that its inhibition may explain at least part of the cardioprotective effect of ACE inhibitors and ARBs. Unlike ARBs, ACE inhibitors may directly prevent myocardial hypertrophy also by increasing local bradykinin bioavailability through inhibition of the myocardial kallikrein-kinin pathway (23). Vascular stiffness and RAAS activation (24) are common in patients with type 2 diabetes, and this may explain the high prevalence and severity of LVH in this population, even at “normal” blood pressure (8,9). The synergistic effect on left ventricular structure of arterial hypertension and of the above hemodynamic and metabolic abnormalities most likely explained the relatively high incidence of ECG-LVH we observed in our present control subjects receiving non-ACE inhibitor therapy. On the other hand, all of the above risk factors for LVH can be controlled by ACE inhibitor therapy. Indeed, in addition to inhibiting angiotensin II and aldosterone production, ACE inhibitors may also ameliorate arterial compliance (25). Thus, the coexistence of several abnormalities that can be ameliorated by RAAS inhibitors may also explain the remarkable protective effect of trandolapril against ECG-LVH we observed here, an effect that appears to exceed that of ramipril in the HOPE trial (6).

inhibitors may also ameliorate arterial compliance (25). Thus, the coexistence of several abnormalities that can be ameliorated by RAAS inhibitors may also explain the remarkable protective effect of trandolapril against ECG-LVH we observed here, an effect that appears to exceed that of ramipril in the HOPE trial (6). Post hoc analyses according to the four original randomization arms showed that trandolapril alone prevented LVH more effectively than verapamil alone or placebo, whereas trandolapril combined with verapamil was more cardioprotective than verapamil alone. Trandolapril alone tended to be more effective than trandolapril combined with verapamil, whereas the effect of verapamil was similar to that of placebo. Within the limitations of the limited power and multiple comparisons, these data show, consistent with the results of primary outcome analyses, that the ACE inhibitor is the effective component of the trandolapril-verapamil combination and that verapamil has no specific protective effect against LVH. A possible limitation of the present study is that, as in previous large studies (4,6,7), LVH was assessed by electrocardiography. This probably resulted in reduced precision of outcome data (ECG may underestimate LVH in obese subjects) that reduced the power of the analyses but probably was not likely to introduce a systematic bias. Actually, a posterior power analyses showed that the probability of a false-positive finding was <0.2%.

rdiography. This probably resulted in reduced precision of outcome data (ECG may underestimate LVH in obese subjects) that reduced the power of the analyses but probably was not likely to introduce a systematic bias. Actually, a posterior power analyses showed that the probability of a false-positive finding was <0.2%. In summary, hypertensive patients with type 2 diabetes receiving trandolapril had a significantly lower incidence of ECG-LVH by Sokolow-Lyon and Cornell voltage than control subjects receiving non-ACE inhibitor therapy, an effect that was highly significant even after adjustments for blood pressure control achieved in the two treatment groups throughout the study. These findings and the observation that ECG-LVH strongly predicts cardiovascular morbidity and mortality in people with hypertension and type 2 diabetes support the use of early ACE inhibitor therapy for effective prevention of left ventricular hypertrophy and, conceivably, cardiovascular morbidity and mortality (6,7) in this population. Supplementary Material Online-Only Appendix Abbott (Ludwigshafen, Germany) partially supported the study. We thank Maria Ganeva and Laura Gallizioli for technical support, the doctors and nurses of the BENEDICT study group for patient care, and Manuela Passera for help in preparing the manuscript.

Gestational diabetes mellitus (GDM) is a well-known reproductive risk factor for subsequent type 2 diabetes (1). Other reproductive factors such as preeclampsia are associated with insulin resistance during pregnancy and may also increase the subsequent risk for diabetes. Furthermore, some (2–4) but not all (5) studies suggest that pregnancy itself is a risk factor for future type 2 diabetes. For example, a population-based study of 1,186 elderly women showed that, even after accounting for age, obesity, and family history of diabetes, parity was associated with an increased risk of type 2 diabetes, with an odds ratio (OR) of 1.16 per pregnancy (95% CI 1.04–1.20) (3). An even larger study comprising 2,310 women with type 2 diabetes reported that parity greater than six was associated with a relative risk (RR) of diabetes of 1.56 (95% CI 1.27–1.91); however, the estimate of the RR decreased to 1.19 (0.97–1.48) after adjustment for current age (2). The applicability of these results is limited by the homogeneity of the population (registered nurses with relatively high socioeconomic status and 98% Caucasian) and the use of the older fasting plasma glucose cutoff for diabetes of >7.8 mmol/l (>140 mg/dl) rather than the current, more sensitive value of 7.0 mmol/l (126 mg/dl) (6).

ty of these results is limited by the homogeneity of the population (registered nurses with relatively high socioeconomic status and 98% Caucasian) and the use of the older fasting plasma glucose cutoff for diabetes of >7.8 mmol/l (>140 mg/dl) rather than the current, more sensitive value of 7.0 mmol/l (126 mg/dl) (6). The prevalence of dysglycemia (type 2 diabetes, impaired glucose tolerance [IGT], and impaired fasting glucose [IFG]) is increasing; however, reproductive risk factors are often underrecognized. In particular, their association with the more recently recognized forms of glucose dysregulation, IGT and IFG, have not yet been well studied. The detection of dysglycemia could be improved if risk factors were better known. Moreover, if reproductive factors such as parity and preeclampsia are risk factors for dysglycemia, they could be used to refine screening approaches. The goal of this research was to identify reproductive risk factors for dysglycemia in a contemporary, multiethnic group of women.

ved if risk factors were better known. Moreover, if reproductive factors such as parity and preeclampsia are risk factors for dysglycemia, they could be used to refine screening approaches. The goal of this research was to identify reproductive risk factors for dysglycemia in a contemporary, multiethnic group of women. RESEARCH DESIGN AND METHODS This is a study of 14,661 women screened as possible participants in the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) trial (7,8), a large, international, multicenter, randomized, double-blind, controlled diabetes prevention trial. Participants were volunteers recruited from 21 countries on five continents from a wide variety of sources including first-degree relatives of diabetic individuals in diabetes clinics, ads in newspapers, pharmacies, national diabetes associations, newsletters, clinics, community announcements, screening programs, and targeted mailings. Institutional research ethics boards at each site approved the DREAM trial. Assessment After an overnight (8–18 h) fast, participants consumed 75 g anhydrous glucose and provided fasting and 2-h blood samples for local measurement of plasma glucose. At the same time they completed a 12-page questionnaire regarding baseline characteristics, medications, and personal and family history, and women completed a 1-page reproductive questionnaire regarding regularity of menstrual cycles, fertility, how many children they had given birth to, and complications of pregnancy including GDM, preeclampsia, or eclampsia.

uestionnaire regarding baseline characteristics, medications, and personal and family history, and women completed a 1-page reproductive questionnaire regarding regularity of menstrual cycles, fertility, how many children they had given birth to, and complications of pregnancy including GDM, preeclampsia, or eclampsia. Definitions Type 2 diabetes was defined as fasting plasma glucose ≥7.0 mmol/l (≥126 mg/dl) or plasma glucose ≥11.1 mmol/l (≥200 mg/dl) 2 h after a 75-g oral glucose load. IGT was defined as plasma glucose 7.8–11.0 mmol/l (140–199 mg/dl) 2 h after a 75-g oral glucose load. IFG was defined as fasting glucose of 6.1–6.9 mmol/l (110–124 mg/dl). Dysglycemia was defined as IFG, IGT, or type 2 diabetes. Parity was defined as the number of infants a woman had borne. Irregular menses was defined as six or fewer menstrual cycles per year between the ages of 18 and 45 years not including pregnancy and was used as a surrogate for polycystic ovary syndrome. Early menopause was defined as the permanent cessation of menstrual periods before age 45. Income range tables with five strata were developed specific to each country in which recruitment occurred; low socioeconomic status was defined as the lowest strata for that country. In Canada, for example, that included a household income ≤$29,999 and for the U.S. it was ≤$15,400. Non-European ancestry was defined as anyone indicating any ancestry other than European at the time of their clinic visit.

recruitment occurred; low socioeconomic status was defined as the lowest strata for that country. In Canada, for example, that included a household income ≤$29,999 and for the U.S. it was ≤$15,400. Non-European ancestry was defined as anyone indicating any ancestry other than European at the time of their clinic visit. Statistical analysis Women were classified as those with and without dysglycemia. Continuous variables were compared using a t test, and categorical variables were compared with a χ2 test. Logistic regression was used to calculate age-adjusted ORs and 95% CI for reproductive risk factors for dysglycemia. Factors that were statistically significant at P < 0.10 in the age-adjusted analysis were included in the multivariate logistic regression to determine their independent relationship with prevalent dysglycemia. This model was rerun for each tertile of BMI (and P values for heterogeneity were calculated) to determine whether the risk of dysglycemia for each risk factor varied with BMI. All P values are reported as two-tailed. All analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC).

nt dysglycemia. This model was rerun for each tertile of BMI (and P values for heterogeneity were calculated) to determine whether the risk of dysglycemia for each risk factor varied with BMI. All P values are reported as two-tailed. All analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC). RESULTS Table 1 presents the base- line characteristics of the women with (n = 6, 298) and without (n = 8, 363) dysglycemia, whose 2-h plasma glucose concentrations were 10.0 ± 3.1 versus 5.7 ± 1.1 mmol/l, respectively (P < 0.0001), with fasting values of 6.3 ± 1.4 versus 5.0 ± 1.1 mmol/l (P < 0.0001, respectively). Women with dysglycemia were significantly older than those without dysglycemia (55.1 vs. 50.0 years, respectively, P < 0.0001). After adjustment for age, most of the reproductive factors remained significantly associated with dysglycemia, including the number of children a woman gave birth to (OR 1.05 per child [95% CI 1.04–1.06]), a history of preeclampsia/eclampsia (1.19 [1.07–1.32]), irregular menses (1.2 [1.09–1.38]), GDM (1.58 [1.40–1.78]), low socioeconomic status (1.09 [1.01–1.17]), and non-European ancestry (1.10 [1.02–1.17]) (Table 1).

lycemia, including the number of children a woman gave birth to (OR 1.05 per child [95% CI 1.04–1.06]), a history of preeclampsia/eclampsia (1.19 [1.07–1.32]), irregular menses (1.2 [1.09–1.38]), GDM (1.58 [1.40–1.78]), low socioeconomic status (1.09 [1.01–1.17]), and non-European ancestry (1.10 [1.02–1.17]) (Table 1). In a multivariate model that included age, ancestry, and fertility-related factors, dysglycemia was significantly associated with the number of children a woman gave birth to (OR 1.03 per child [95% CI 1.01–1.05]), age (1.05 per year [1.04–1.05]), non-European ancestry (1.09 [1.01–1.17]), a history of preeclampsia/eclampsia (1.14 [1.02–1.27]), irregular menses (1.21 [1.07–1.36]), and GDM (1.53 [1.35–1.74]) (Fig. 1). Early menopause (1.24 [0.98–1.53]) and low socioeconomic status (0.95 [0.88–1.02]) were no longer significantly associated with dysglycemia. To determine whether there was an interaction between BMI and the reproductive risk factors, particularly GDM, we reran the multiple regression model for each tertile of BMI, with 27.1 and 32.2 defining lower, middle, and upper levels. The relationships between GDM and current dysglycemia did not differ significantly across tertiles of BMI (Pheterogeneity = 0.84), nor for any of the other risk factors (Pheterogeneity ≥0.10 for all). The ORs for GDM and current dysglycemia for each tertile of BMI (<27.1, 27.1–32.2, and >32.2) were 1.81 [95% CI 1.42–2.30], 1.44 [1.14–1.81], and 1.47 [1.20–1.81], respectively.

not differ significantly across tertiles of BMI (Pheterogeneity = 0.84), nor for any of the other risk factors (Pheterogeneity ≥0.10 for all). The ORs for GDM and current dysglycemia for each tertile of BMI (<27.1, 27.1–32.2, and >32.2) were 1.81 [95% CI 1.42–2.30], 1.44 [1.14–1.81], and 1.47 [1.20–1.81], respectively. CONCLUSIONS This large, multiethnic study of middle-aged women showed that a history of GDM is independently associated with prevalent dysglycemia, confirming that pregnancy is a “stress test for life” (9,10). This observation may be understood in light of the fact that the occurrence of GDM is clear evidence of an impaired ability to maintain normoglycemia under the metabolic stress of pregnancy and is consistent with previous reports (11–14) from smaller studies. Hence, in young women of child-bearing age, reproductive factors, particularly GDM, can be used to counsel patients about their future risk of dysglycemia regardless of future BMI, whereas in middle-aged women a history of reproductive risk factors may be useful as a screening tool for dysglycemia. This study also showed that a history of a variety of reproductive risk factors, including irregular menses, parity, and preeclampsia, was independently associated with dysglycemia and was not explained by age, ethnicity, or socioeconomic status. One potential explanation is the association of many of these factors with insulin resistance, including preeclampsia (15,16), pregnancy (17), and, even in nonobese women, polycystic ovary syndrome (PCOS) (of which irregular menses is a key component) (18–21).

s not explained by age, ethnicity, or socioeconomic status. One potential explanation is the association of many of these factors with insulin resistance, including preeclampsia (15,16), pregnancy (17), and, even in nonobese women, polycystic ovary syndrome (PCOS) (of which irregular menses is a key component) (18–21). This is the only study, to our knowledge, examining reproductive risk factors for dysglycemia involving a broad population, allowing for wide applicability of results. Participation spanned all socioeconomic strata and involved 21 countries on five continents. Another strength is the large sample size (14, 661 women), by far the largest study in the literature on reproductive risk factors and dysglycemia, which allowed for control of multiple confounding factors. Another strength is the fact that reproductive risk factors that identified women at increased risk of dysglycemia were elicited with simple screening questions, which are part of a routine history, and do not require serology or imaging investigations such as pelvic ultrasound. Participants were asked, for instance, about irregular menses as a surrogate for PCOS, as PCOS remains undiagnosed in many patients or they are unfamiliar with the medical term. Although this approach has the potential for misclassifying some patients who had had fertility-related risk factors as being unaffected, it suggests that the associations between fertility-related risk factors and dysglycemia are probably even stronger than those observed.

or they are unfamiliar with the medical term. Although this approach has the potential for misclassifying some patients who had had fertility-related risk factors as being unaffected, it suggests that the associations between fertility-related risk factors and dysglycemia are probably even stronger than those observed. Limitations of the study include the fact that the participants were asked to recall events, such as pregnancies, which in many instances occurred several decades earlier. However, by gathering this baseline information before the administration of the oral glucose tolerance test, recall bias was limited. We did not have access to the participants’ medical charts and relied on patient history, which may not always be reliable and may underestimate some of the above associations. In summary, in this large, multiethnic study of middle-aged women without a previous diagnosis of diabetes, prevalent dysglycemia was independently associated with a history of several reproductive risk factors, particularly GDM. Moreover, the relationship between prior GDM and current dysglycemia persisted across BMI strata. The DREAM Trial was funded by the Canadian Institutes of Health Research, sanofi-aventis, GlaxoSmithKline, and King Pharmaceuticals through the University Industry Grant Program. S.Y. holds a Heart and Stroke Foundation Chair in Cardiovascular Research. S.S.A. holds the Eli Lilly Canada–May Cohen Chair in Women's Health. H.C.G. holds the Population Health Institute Chair in Diabetes Research sponsored by aventis.

Glycemic control has long been known to play a critical role in the development of the complications of diabetes. Since the results of the Diabetes Control and Complications Trial (DCCT) first demonstrated that intensive efforts to lower glycemia resulted in a reduction in the rate at which subjects with type 1 diabetes developed microvascular complications (1), clinicians and their patients have attempted to normalize glucose control through many different therapeutic modalities. Exogenous insulin has been used to achieve target glycemia almost universally in the treatment of patients with type 1 diabetes, but the risk of developing severe hypoglycemia has become a limiting factor for many (2). Pancreas transplantation offers an alternative for selected patients with diabetes who seek to achieve normal levels of glycemia without periodic hypoglycemia. During 2004, >1,400 pancreas transplants were performed in the U.S. (3). The vast majority of these transplants were done using deceased donor organs. However, some centers, including our own, have considered using living donors in situations in which improved outcomes over the use of a deceased donor organ might be expected. Such situations could include the presence of a nondiabetic HLA-identical sibling, a recipient with high panel-reactive antibody levels, or associated morbidities that predict a high risk of mortality while the recipient is on the waiting list.

improved outcomes over the use of a deceased donor organ might be expected. Such situations could include the presence of a nondiabetic HLA-identical sibling, a recipient with high panel-reactive antibody levels, or associated morbidities that predict a high risk of mortality while the recipient is on the waiting list. At the University of Minnesota, the use of living donors in pancreas transplantation dates back to 1977 (4). In this procedure, the distal half of the pancreas is removed from a living donor and placed within the pelvis of the diabetic recipient. Although outcomes for the recipient are at least equivalent to those achieved with a deceased donor organ and perhaps improved for those patients with high panel-reactive antibody levels that prevent an optimal tissue match (4), 25% of the donors were previously found to have glucose intolerance or frank diabetes (non–insulin-dependent) 1 year after hemipancreatectomy (HPx) (5). In addition, even donors with normal glucose tolerance were noted to experience a modest increase in blood glucose and a reduction in insulin and glucagon secretion ≥1 year after HPx (6). Because of these findings, the University of Minnesota changed the criteria used to select hemipancreas donors in 1997 to exclude those with clinical or metabolic features that may be associated with the future development of diabetes.

ood glucose and a reduction in insulin and glucagon secretion ≥1 year after HPx (6). Because of these findings, the University of Minnesota changed the criteria used to select hemipancreas donors in 1997 to exclude those with clinical or metabolic features that may be associated with the future development of diabetes. In this report, we examine the metabolic outcomes in hemipancreatectomized donors selected because they appeared to be at low risk for developing diabetes. Our study was designed to test the hypothesis that the implementation of the revised University of Minnesota criteria would reduce risk of development of abnormal glucose tolerance to less than the rate of 25% reported using the previous criteria (5). We hoped that our observations would be of benefit to pancreas transplant programs considering the development of a living donor program. Further study of this unique population of healthy hemipancreatectomized humans also provides us with a rare opportunity to gain insight into the effect of β-cell mass reduction on the maintenance of normal glucose tolerance. RESEARCH DESIGN AND METHODS Donor selection In December 1996, the University of Minnesota pancreas transplant program revised the selection criteria for living hemipancreas donors to preclude those believed to be at greatest risk for the development of diabetes from undergoing the procedure (7). The revised criteria are shown in Table 1.

METHODS Donor selection In December 1996, the University of Minnesota pancreas transplant program revised the selection criteria for living hemipancreas donors to preclude those believed to be at greatest risk for the development of diabetes from undergoing the procedure (7). The revised criteria are shown in Table 1. Operative procedure The operative procedure for HPx has been described in detail elsewhere (4). The procedure results in resection of the distal pancreas where it overlies the portal vein (50% resection), leaving the pancreatic head and proximal tail intact in the donor. More recently, surgeons have made the pancreas transection slightly to the left of the portal vein, resulting in a 40% pancreatectomy. Follow-up studies In 2006, all donors who underwent HPx between January 1997 and December 2003 were contacted by phone to ascertain their willingness to participate in a study. This contact was followed by a written explanation of the study. All donors contacted were interviewed on the phone about their current health status. Donors who were not taking diabetes medications were asked to undergo the metabolic evaluation detailed below. This protocol was approved by the University of Minnesota Institutional Review Board, and subjects gave written informed consent before their participation.

ewed on the phone about their current health status. Donors who were not taking diabetes medications were asked to undergo the metabolic evaluation detailed below. This protocol was approved by the University of Minnesota Institutional Review Board, and subjects gave written informed consent before their participation. The metabolic evaluation was done after the donors had followed a diet consisting of at least 150 g carbohydrate per day for 3 days. On the day of the study, donors presented to the University of Minnesota General Clinical Research Center or to a local clinic in the morning after a 12-h fast. Oral glucose tolerance tests were performed by administering 75 g glucose (Cardinal Health, McGaw Park, IL) orally over a 5-min period. Blood samples were obtained for later determination of serum glucose levels at −10 and −5 min before glucose was administered and at 30, 60, 90, and 120 min after the administration of the glucose load. Fasting samples were also obtained for A1C, insulin, and anti-GAD antibody. Serum glucose was measured on an Analog glucose analyzer system (Analox Instruments, Hammersmith, London, U.K.). Serum insulin was measured using chemiluminescence (Immulite 2000). Samples for anti-GAD antibody levels were analyzed using an immunoradiometric assay (Associated Regional and University Pathologists, Salt Lake City, UT) (8). Blood samples obtained in clinics located away from the University of Minnesota were sent overnight to the University of Minnesota Medical Center for analysis.

. Samples for anti-GAD antibody levels were analyzed using an immunoradiometric assay (Associated Regional and University Pathologists, Salt Lake City, UT) (8). Blood samples obtained in clinics located away from the University of Minnesota were sent overnight to the University of Minnesota Medical Center for analysis. Data analysis Unless otherwise indicated, results are given as means ± SD. The American Diabetes Association criteria for the diagnosis of diabetes, impaired glucose tolerance, and impaired fasting glucose (9) were used to categorize glucose tolerance in donors. The differences between groups were analyzed by two-tailed statistics using the Mann-Whitney test for unpaired data and the Wilcoxon signed-rank test or Student's t test for paired data. Correlation was determined using the Spearman rank order correlation coefficient. P < 0.05 was considered statistically significant.

The differences between groups were analyzed by two-tailed statistics using the Mann-Whitney test for unpaired data and the Wilcoxon signed-rank test or Student's t test for paired data. Correlation was determined using the Spearman rank order correlation coefficient. P < 0.05 was considered statistically significant. RESULTS Twenty-one individuals underwent HPx for the purpose of organ donation at the University of Minnesota between January 1997 and December 2003. Preoperatively, 17 were noted to have normal fasting glucose values, normal glucose tolerance, normal insulin secretory responses, and an unremarkable personal and family medical history. One individual with a mother who had type 2 diabetes, one individual with a BMI >27 kg/m2, and three individuals with a single glucose value between 150 and 199 mg/dl on the preoperative oral glucose tolerance test were allowed to serve as hemipancreas donors because they perceived their own risk of donation to be less than the recipient's risk of foregoing the transplant. On follow-up, six donors could not be located. On telephone interview, two reported using a single antidiabetic agent (metformin in one and pioglitazone in the other). The remaining 13 donors completed the follow-up metabolic evaluation. Abnormalities in glucose metabolism were identified in 7 of 13 (54%) reporting for follow-up evaluation (Fig. 1). Table 2 shows the preoperative and postoperative results in the 13 hemipancreatectomized donors who were studied.

itazone in the other). The remaining 13 donors completed the follow-up metabolic evaluation. Abnormalities in glucose metabolism were identified in 7 of 13 (54%) reporting for follow-up evaluation (Fig. 1). Table 2 shows the preoperative and postoperative results in the 13 hemipancreatectomized donors who were studied. In the 13 donors who participated in the metabolic evaluation, no differences were noted with respect to age at donation (47 ± 6 vs. 41 ± 14 years) or months since donation (60 ± 10 vs. 55 ± 12 months) between those with normal versus abnormal glucose tolerance at follow-up. Despite the elevated postoperative serum glucose levels in those with abnormal glucose tolerance, fasting serum insulin values were similar to those measured in the donors with normal postoperative glucose tolerance (5.3 ± 3.0 vs. 6.4 ± 3.6 μU/ml, P = 0.58). There was a trend toward a higher BMI in the group with abnormal glucose tolerance compared with group with normal glucose tolerance both before (26.1 ± 3.6 vs. 23.6 ± 2.2 kg/m2, P = 0.14) and after (26.4 ± 4.4 vs. 23.5 ± 2.4 kg/m2, P = 0.23) donation; however, these differences were not statistically significant. An increase in weight over time did not correlate with higher fasting glucose values in these 13 donors (Spearman r2s = 0.21, P = 0.28). All subjects had immeasurable levels of GAD. The subject who was selected despite a family history of diabetes, the subject allowed to donate despite having a BMI >27 kg/m2, and two of the three subjects who were allowed to donate despite achieving a glucose value between 150 and 199 mg/dl on the preoperative oral glucose tolerance tests showed abnormal glucose tolerance on follow-up. HLA typing data are available for all 15 donors contacted. Only two donors were heterozygous for the DRB1*3/4-DQB1*0302 (DQ8) genotype that confers higher risk for type 1 diabetes (10). One of these donors had impaired fasting glucose, and the other was taking an oral antidiabetic agent on follow-up. Table 3 summarizes the characteristics and variables of the 13 hemipancreatectomized donors studied.

terozygous for the DRB1*3/4-DQB1*0302 (DQ8) genotype that confers higher risk for type 1 diabetes (10). One of these donors had impaired fasting glucose, and the other was taking an oral antidiabetic agent on follow-up. Table 3 summarizes the characteristics and variables of the 13 hemipancreatectomized donors studied. CONCLUSIONS The purpose of this study was to determine whether the application of stringent metabolic criteria during the selection of living pancreas donors would reduce the risk of development of abnormal glucose tolerance after HPx to below the 25% rate reported previously (5). Whereas four of the current donors who developed abnormalities in glucose tolerance were allowed to donate despite failing to meet all of the selection criteria, five of the donors who met the strict criteria developed to minimize the risk of post-HPx disturbances in glucose tolerance had either been given a diagnosis of diabetes by their primary physician or had abnormal glucose tolerance on a standard glucose tolerance test 3 to 10 years after organ donation. The preoperative criteria, which are the gold standard in the pretransplant workup for potential hemipancreas donors (11), were not successful in identifying donors at risk for the development of diabetes and impaired glucose tolerance after β-cell mass reduction. Thus, the risk for postdonation glucose intolerance cannot be completely eliminated even with the application of these narrow criteria. Our findings raise concerns about the expanded use of living donors in pancreas and islet transplantation programs.

diabetes and impaired glucose tolerance after β-cell mass reduction. Thus, the risk for postdonation glucose intolerance cannot be completely eliminated even with the application of these narrow criteria. Our findings raise concerns about the expanded use of living donors in pancreas and islet transplantation programs. The metabolic effects of HPx for the purpose of organ donation have been studied previously. Bolinder et al. (12) were the first to demonstrate that fasting blood glucose concentrations were higher after hemipancreas donation than they were before surgery. This observation was quickly followed by that of Kendall et al. (5) who noted that 25% of healthy individuals providing a hemipancreas to a first-degree relative with type 1 diabetes developed abnormal glucose tolerance or diabetes within 1 year of surgery. Subsequent work demonstrated that HPx is associated with reductions in glucose, arginine, and glucose-potentiated arginine-induced insulin secretion (6) as well as an increase in serum proinsulin concentrations (13). β-Cell mass reduction has been presumed to be the cause of these metabolic abnormalities after HPx. This assumption is supported by recent work in humans in which β-cell mass was found to be reduced in an autopsy study of individuals with type 2 diabetes (14), and the magnitude of the reduction appeared to have a curvilinear relationship with fasting glucose (15). These previous studies, as well as the current study, support the hypothesis that a deficit in β-cell mass contributes to the pathogenesis of hyperglycemia and further emphasize the need to maintain β-cell mass to maintain normal glucose tolerance.

appeared to have a curvilinear relationship with fasting glucose (15). These previous studies, as well as the current study, support the hypothesis that a deficit in β-cell mass contributes to the pathogenesis of hyperglycemia and further emphasize the need to maintain β-cell mass to maintain normal glucose tolerance. Interesting new observations suggest that a reduction in β-cell mass may also reduce glucose disposal in the peripheral tissues (16). In dogs studied by Matveyenko et al. (16), a 50% pancreatectomy resulted in impaired fasting glucose or impaired glucose tolerance, a reduction in the pulse mass of glucose-induced insulin secretion, a decrease in hepatic insulin extraction, and a 40% reduction in insulin-stimulated glucose disposal. The findings raise the provocative possibility that β-cell mass reduction may not only have effects on insulin secretion that are important in the pathogenesis of type 2 diabetes but may also play a role in the impaired insulin action present in this disorder in the dog. Whether β-cell mass has an effect on insulin action in humans is not as clear. Hemipancreatectomized human donors have been found to have normal insulin sensitivity as measured by the hyperinsulinemic-euglycemic clamp and by the frequently sampled intravenous glucose tolerance test (17), but the effect of β-cell mass reduction on the pulse mass of insulin secretion and on hepatic extraction has not been examined.

ized human donors have been found to have normal insulin sensitivity as measured by the hyperinsulinemic-euglycemic clamp and by the frequently sampled intravenous glucose tolerance test (17), but the effect of β-cell mass reduction on the pulse mass of insulin secretion and on hepatic extraction has not been examined. Why one group of donors went on to develop abnormal glucose tolerance whereas another group maintained normal glucose tolerance is not clear from our data. There was a trend toward higher BMIs among those who developed abnormal glucose tolerance, but this difference was not statistically significant. In a previous study, obesity was more related to the development of diabetes or glucose intolerance in donors after HPx than family history of type 2 diabetes or age (18). In our cohort, the stringent criteria of 1997 recommended that no one with a BMI >27 kg/m2 be allowed to provide a hemipancreas to a recipient with type 1 diabetes, although one donor with a BMI of 30.8 was permitted to undergo the procedure. Because nearly half of our relatively lean cohort were found to have abnormal glucose tolerance on follow-up, it is possible that even BMI <27 kg/m2 is still too high to ensure the maintenance of normal glucose homeostasis in the setting of β-cell mass reduction.

th a BMI of 30.8 was permitted to undergo the procedure. Because nearly half of our relatively lean cohort were found to have abnormal glucose tolerance on follow-up, it is possible that even BMI <27 kg/m2 is still too high to ensure the maintenance of normal glucose homeostasis in the setting of β-cell mass reduction. Nine of the donors in our group were not biologically related to the recipient of their hemipancreas. Five of these individuals are known to have developed abnormal glucose tolerance on follow-up. It is interesting that the proportion of first-degree relatives known to have developed diabetes or abnormal glucose tolerance (4 of 12) is actually lower than the proportion in those unrelated to their recipient (5 of 9), suggesting that a family history of diabetes may not be predictive of abnormal glucose tolerance after HPx. The present study has certain limitations. Six of the 21 donors operated on between 1997 and 2003 were lost to follow-up and as a result we lack a complete description of the outcomes in this patient group. However, even if all six of these donors were found to have normal glucose tolerance, 43% (9 of 21) of those undergoing HPx during this period were found to have abnormal glucose tolerance on follow-up. Our conclusion that the new criteria do not successfully exclude subjects at risk for developing abnormal glucose tolerance after HPx still stands.

ese donors were found to have normal glucose tolerance, 43% (9 of 21) of those undergoing HPx during this period were found to have abnormal glucose tolerance on follow-up. Our conclusion that the new criteria do not successfully exclude subjects at risk for developing abnormal glucose tolerance after HPx still stands. All of these donors underwent a surgical HPx, but it is unclear whether small differences in technique among surgeons could play a role in donor outcomes. Determining the exact percentage of islet tissue removed is difficult because of the heterogeneous nature of the pancreas. However, Robertson et al. (18) found that probable differences in β-cell mass between donated and retained pancreatic segments did not appear to influence clinical outcomes.

role in donor outcomes. Determining the exact percentage of islet tissue removed is difficult because of the heterogeneous nature of the pancreas. However, Robertson et al. (18) found that probable differences in β-cell mass between donated and retained pancreatic segments did not appear to influence clinical outcomes. In summary, our results indicate that the development of abnormal glucose tolerance is a risk inherent to HPx. Despite the use of stringent criteria to exclude those at risk for developing abnormalities in glucose metabolism, 43% of healthy humans who underwent HPx between 1997 and 2003 at the University of Minnesota (60% of those who reported for examination) had impaired fasting glucose, impaired glucose tolerance, or diabetes on follow-up. Given our inability to accurately predict those who will develop abnormalities in glucose tolerance and the fact that, in most circumstances, whole pancreas transplantation from a deceased donor would be comparably effective (4), we conclude that living donors should be used only in very special circumstances. Such circumstances may arise in the presence of an HLA-identical sibling or a recipient with high panel-reactive antibody levels, which would make tissue matching very difficult, or associated morbidities that predict a high risk of recipient mortality while on the waiting list. As in the case of all living organ donation, the donor should always be expected to experience consequences. In the future, preoperative genotyping might be useful in excluding those potential donors with the greatest risk for developing diabetes, but until then our data demonstrate that even subjects with normal preoperative metabolic characteristics are at significant risk of developing abnormalities in glucose tolerance after HPx.

uture, preoperative genotyping might be useful in excluding those potential donors with the greatest risk for developing diabetes, but until then our data demonstrate that even subjects with normal preoperative metabolic characteristics are at significant risk of developing abnormalities in glucose tolerance after HPx. The authors gratefully acknowledge the support of the Diabetes Trust Fund and the University of Minnesota General Clinical Research Center (M01-RR00400). Parts of this study were presented in abstract form at the 67th annual meeting of the American Diabetes Association, Chicago, Illinois, 22–26 June 2007.

The potential role of the mitochondria in the development of insulin resistance and type 2 diabetes has recently attracted much interest. Muscle biopsies taken from people with type 2 diabetes demonstrate smaller mitochondria and lower activities of oxidative enzymes compared with those of lean individuals without diabetes (1). Insulin-resistant people with a family history of diabetes have reduced basal mitochondrial activity in skeletal muscle compared with insulin-sensitive individuals (2). These observations, in combination with others (3–6), raise the possibility that mitochondrial defects could underlie type 2 diabetes. Defects in oxidative function could possibly help explain the impaired fatty acid oxidation (7) and elevated intramyocellular lipid (IMCL) (8) characteristic of impaired insulin action and type 2 diabetes. The elevated intramuscular lipid may affect insulin signaling in skeletal muscle (5), exacerbating insulin resistance. However, other studies have not observed abnormalities in basal mitochondrial activity in skeletal muscle of people with type 2 diabetes (9). Recent biopsy work has also shown that differences in oxidative enzymes between people with and without type 2 diabetes disappear when corrected for mitochondrial density (10). These data raise the possibility that type 2 diabetes is associated with normal mitochondrial function but that the mitochondrial capacity is reduced. This is an important differentiation, as it holds implications for the therapeutic approach to type 2 diabetes.

disappear when corrected for mitochondrial density (10). These data raise the possibility that type 2 diabetes is associated with normal mitochondrial function but that the mitochondrial capacity is reduced. This is an important differentiation, as it holds implications for the therapeutic approach to type 2 diabetes. People with type 2 diabetes are more sedentary than those without diabetes (11). It is clear that reversing this sedentary lifestyle with physical activity and/or exercise can produce significant improvements in long-term glucose control (12). These benefits could be mediated, at least in part, by changes in mitochondrial function (13). In people with type 2 diabetes, moderate-intensity exercise combined with moderate weight loss produced a significant improvement in insulin sensitivity and mitochondrial density (14). However, such moderate intensity exercise programs are difficult to implement and usually require close supervision. In contrast, unsupervised walking has been shown to produce significant improvements in long-term glucose control and is a sustainable behavior over long periods of time (2 years) (15). Little is known about how low-intensity physical activity interventions such as walking influence muscle metabolism in people with type 2 diabetes.

unsupervised walking has been shown to produce significant improvements in long-term glucose control and is a sustainable behavior over long periods of time (2 years) (15). Little is known about how low-intensity physical activity interventions such as walking influence muscle metabolism in people with type 2 diabetes. This study was designed to 1) determine whether there are differences in basal and stimulated mitochondrial activity in people with type 2 diabetes compared with physical activity–matched control subjects and 2) establish whether an increase in daily physical activity is associated with changes in mitochondrial ATP turnover and changes in lipid oxidation.

ne whether there are differences in basal and stimulated mitochondrial activity in people with type 2 diabetes compared with physical activity–matched control subjects and 2) establish whether an increase in daily physical activity is associated with changes in mitochondrial ATP turnover and changes in lipid oxidation. RESEARCH DESIGN AND METHODS Subject information and initial testing Sedentary people with type 2 diabetes (>2 years duration, A1C <7.5%, stable control on either diet or diet plus sulfonylurea and/or metformin) (n = 10) and age-, weight-, and physical activity–matched people without type 2 diabetes (n = 10) were recruited. Volunteers with heart, liver, kidney, or diabetic foot disease or those undertaking a physical activity program were excluded. Participants were assessed before and after 2 and 8 weeks of increased physical activity. At each time point, physical activity, resting substrate oxidation, fasting plasma glucose, and A1C were assessed. Basal and maximal ATP use and IMCL were quantitated using magnetic resonance techniques. For all metabolic evaluations, participants were transported to the magnetic resonance facility by taxi and data were collected in the fasted state. Participants provided informed consent to join the study, and the study was approved by the local research ethics committee.

antitated using magnetic resonance techniques. For all metabolic evaluations, participants were transported to the magnetic resonance facility by taxi and data were collected in the fasted state. Participants provided informed consent to join the study, and the study was approved by the local research ethics committee. Magnetic resonance acquisition Magnetic resonance data were acquired using a 3T Achieva scanner (Philips, Best, the Netherlands) with an in-built body coil used for imaging, a 14-cm diameter surface coil for phosphorus spectroscopy, and a 10-cm diameter pair of flexible coils (Philips) for proton spectroscopy.

antitated using magnetic resonance techniques. For all metabolic evaluations, participants were transported to the magnetic resonance facility by taxi and data were collected in the fasted state. Participants provided informed consent to join the study, and the study was approved by the local research ethics committee. Magnetic resonance acquisition Magnetic resonance data were acquired using a 3T Achieva scanner (Philips, Best, the Netherlands) with an in-built body coil used for imaging, a 14-cm diameter surface coil for phosphorus spectroscopy, and a 10-cm diameter pair of flexible coils (Philips) for proton spectroscopy. Resting ATP flux This technique has previously been described in nontechnical terms (16). In brief, a saturation transfer sequence was used to measure transfer of magnetization between γ-ATP and inorganic phosphate (Pi) (17). The steady-state magnetization of Pi was measured during selective irradiation of γ-ATP (Mz) and compared with the equilibrium Pi magnetization with the irradiation placed symmetrically downfield from the Pi frequency (TR = 25 s, bandwidth = 3,000 Hz, 2,048 points, 16 averages) (Mo). The fractional reduction of Pi magnetization upon saturation of γ-ATP, (Mo − Mz)/Mo, was used to calculate the pseudo–first order rate constant using the Forsen-Hoffman equation (18). T1* was measured using an inversion recovery experiment (τ1 − 180° − τ2 − 90° − acquire, TR = 25 s, four averages), while saturation of γ-ATP was performed during the delay times τ1 and τ2. Broadband proton decoupling was used. Eight variable τ2 time delays were used ranging from 635 to 9,035 ms. The intraday variability of the method is 6.5% and interday variation 8.0%.

nt (τ1 − 180° − τ2 − 90° − acquire, TR = 25 s, four averages), while saturation of γ-ATP was performed during the delay times τ1 and τ2. Broadband proton decoupling was used. Eight variable τ2 time delays were used ranging from 635 to 9,035 ms. The intraday variability of the method is 6.5% and interday variation 8.0%. Maximal ATP generation Plantar flexion exercise at 30% of the maximum voluntary contraction was performed in the magnetic resonance imaging scanner on a custom-built device. The study protocol consisted of 3 min of plantar flexion at 2 Hz and 3 min of rest, changing pH levels as little as possible (19). Phosphorus spectra were collected at 10-s intervals throughout exercise (NS = 2, bandwidth = 3,000 Hz, 2,048 points, broadband decoupling, and NOE). IMCL Proton magnetic resonance spectroscopy was acquired using a localized PRESS sequence (voxel size 15 × 15 × 20 mm) in soleus with water suppression (TR/TE/NSA = 3,000 ms/37 ms/32,2048 samples, bandwidth 2,000 Hz). Sixteen unsuppressed averages were collected for reference. Quantitation of spectra Analysis of all spectra was performed with jMRUI (version 3.0) (20) using AMARES (21), with custom prior knowledge. Phosphcreatine (PCr) concentrations were calculated by measuring PCr relative to β-ATP, correcting for magnetic saturation and assuming a resting (ATP) of 8.2 mmol/l (19). Maximal ATP production was assessed from postexercise PCr kinetics (19). Proton spectra were analyzed for IMCL and expressed relative to the water reference peak.

(PCr) concentrations were calculated by measuring PCr relative to β-ATP, correcting for magnetic saturation and assuming a resting (ATP) of 8.2 mmol/l (19). Maximal ATP production was assessed from postexercise PCr kinetics (19). Proton spectra were analyzed for IMCL and expressed relative to the water reference peak. Physical activity Physical activity was assessed over 3 days using a validated multisensor armband (22) (SenseWear; Bodymedia, Pittsburgh, PA). Physical activity goals were set, with participants targeting 45 min of extra walking per day, and the benefits to glucose control (diabetes group) and long-term well-being (control group) discussed. Participants were also provided with a pedometer, they recorded the pedometer reading, and they received periodic phone calls from the research team. Indirect calorimetery Expired gases were collected from a constant-flow hood calorimeter (Deltatrac; Datex Ohmeda, Hertfordsire, U.K.) over 30 min. Substrate oxidation rates and energy expenditure were calculated from oxygen consumption and carbon dioxide production values using stoichiometric equations (23). Whole-blood glucose and plasma insulin Whole-blood glucose was measured (YSI glucose analyzer; YSI, Yellow Springs, OH). Plasma insulin was measured using an enzyme-linked immunosorbent assay kit (DAKO, Ely, U.K.). A1C was measured using high-performance liquid chromatography (TOSOH, Tokyo, Japan).

Indirect calorimetery Expired gases were collected from a constant-flow hood calorimeter (Deltatrac; Datex Ohmeda, Hertfordsire, U.K.) over 30 min. Substrate oxidation rates and energy expenditure were calculated from oxygen consumption and carbon dioxide production values using stoichiometric equations (23). Whole-blood glucose and plasma insulin Whole-blood glucose was measured (YSI glucose analyzer; YSI, Yellow Springs, OH). Plasma insulin was measured using an enzyme-linked immunosorbent assay kit (DAKO, Ely, U.K.). A1C was measured using high-performance liquid chromatography (TOSOH, Tokyo, Japan). Statistical analysis Statistical calculations were performed using SPSS version 11 (SPSS, Chicago, IL). Two-way ANOVA (time and treatment) was used to assess metabolic and physiological differences between groups. Statistical significance was accepted at P < 0.05. Data are presented as means ± SE of the mean, unless otherwise stated. RESULTS Baseline group description The characteristics of the type 2 diabetes and control groups are given in Table 1. Groups were matched for age, sex, and weight. The diabetic group was shorter than the control group, resulting in a higher BMI. Habitual physical activity was similarly low in both groups. Participants with type 2 diabetes had good glucose control, demonstrated by a mean A1C of 6.7 ± 0.3% and a fasting whole-blood glucose concentration of 7.1 ± 0.4 mmol/l.

weight. The diabetic group was shorter than the control group, resulting in a higher BMI. Habitual physical activity was similarly low in both groups. Participants with type 2 diabetes had good glucose control, demonstrated by a mean A1C of 6.7 ± 0.3% and a fasting whole-blood glucose concentration of 7.1 ± 0.4 mmol/l. Physical activity The physical activity monitors showed that both groups had low baseline activity levels. There was no significant difference in the number of steps taken per day between the diabetes and control groups at baseline (Fig. 1). Following counseling, both groups demonstrated a sustained increase in the number of steps taken per day (P < 0.05, Fig. 1). At 8 weeks, both groups undertook more steps than baseline; however, the control group had reduced the number of steps from the 2-week point (P < 0.05) (Fig. 1). Resting skeletal muscle metabolites Skeletal muscle metabolites were similar between the two groups (Table 2). Resting skeletal muscle ADP concentrations were similar between people with or without diabetes and did not change after 2 or 8 weeks of increased physical activity (Table 2). There was no difference in resting skeletal muscle pH between people with and without diabetes, and this was not influenced by 2 and 8 weeks of increased physical activity (Table 2). Similarly, the ratio of IMCL to water was comparable between people with and without diabetes at baseline and also did not change after 2 and 8 weeks of increased physical activity (Table 2).

between people with and without diabetes, and this was not influenced by 2 and 8 weeks of increased physical activity (Table 2). Similarly, the ratio of IMCL to water was comparable between people with and without diabetes at baseline and also did not change after 2 and 8 weeks of increased physical activity (Table 2). Mitochondrial activity The type 2 diabetic and control groups showed similar basal and maximal ATP turnover rates (Table 3). There was no relationship between basal and maximal ATP turnover (R2 = 0.194, P > 0.05). Both basal and maximal ATP turnover remained constant following 2 or 8 weeks of increased physical activity in either the type 2 diabetes or control group (Table 3). End-exercise ADP and PCr use during exercise were similar (Table 3), ensuring that the recovery modeling of phosphocreatine was completed from similar metabolic starting points. Indirect calorimetery People with and without type 2 diabetes showed similar levels of carbohydrate (2.82 ± 0.19 vs. 2.61 ± 0.17 mg · kg body wt−1 · min−1) and lipid oxidation (Fig. 1). There were no changes in basal energy expenditure at any time point. Following 8 weeks of increased physical activity, the diabetes group showed an increase in rate of lipid oxidation (Fig. 1) and a decrease in rate of carbohydrate oxidation (to 2.26 ± 0.022 mg · kg body wt−1 · min−1).

−1) and lipid oxidation (Fig. 1). There were no changes in basal energy expenditure at any time point. Following 8 weeks of increased physical activity, the diabetes group showed an increase in rate of lipid oxidation (Fig. 1) and a decrease in rate of carbohydrate oxidation (to 2.26 ± 0.022 mg · kg body wt−1 · min−1). Plasma glucose, insulin, and A1C People with type 2 diabetes had higher A1C (P < 0.01, Table 2) and fasting plasma glucose levels (P < 0.01) (Table 2). Insulin sensitivity, as assessed using homeostasis model assessment, was lower in people with type 2 diabetes than in control subjects (P < 0.05) (Table 2). In the type 2 diabetes group, there was a nonsignificant trend toward lower fasting plasma glucose (P = 0.08) (Table 2). Anthropometric measurements There were no differences in weight between people with and without type 2 diabetes (Table 1). There was no change in weight in the type 2 diabetes group after 2 weeks, although after 8 weeks there was a significant decrease relative to both baseline and 2 weeks (P < 0.05) (Table 2). Control subjects showed no change in weight after 2 or 8 weeks of increased physical activity (Table 2).

ype 2 diabetes (Table 1). There was no change in weight in the type 2 diabetes group after 2 weeks, although after 8 weeks there was a significant decrease relative to both baseline and 2 weeks (P < 0.05) (Table 2). Control subjects showed no change in weight after 2 or 8 weeks of increased physical activity (Table 2). CONCLUSIONS We observed no abnormality in mitochondrial function in people with well-controlled type 2 diabetes compared with physical activity–, age-, and weight-matched control subjects. There was no relationship between basal (fasted) and maximal (recovery from exercise) ATP synthesis, suggesting that the factors influencing basal and maximal ATP synthesis are different. The physical activity intervention markedly increased the number of steps taken per day during the 8-week intervention. Fasting lipid oxidation was increased, but there was no change in ATP turnover or maximal ATP production.

thesis, suggesting that the factors influencing basal and maximal ATP synthesis are different. The physical activity intervention markedly increased the number of steps taken per day during the 8-week intervention. Fasting lipid oxidation was increased, but there was no change in ATP turnover or maximal ATP production. The observation of no abnormality in basal ATP flux contrasts with recent studies. The seminal study (2) indicating the possible importance of mitochondrial function in development of type 2 diabetes observed basal ATP flux in extreme phenotypes of insulin sensitivity. From 150 screened subjects, the most and least insulin-sensitive people underwent assessment of basal ATP flux and IMCL using magnetic resonance techniques. The 14 least insulin-sensitive subjects (with a family history of diabetes) had lower basal ATP flux and higher intramuscular lipid levels than the 10 most insulin-sensitive people. It was suggested that impaired mitochondrial function reduces the ability to oxidize lipids, with accumulation of intramuscular lipid impeding insulin signaling (5), forming a pathway from impaired mitochondrial function to the development of type 2 diabetes. A further study (3) demonstrated impaired ATP flux in people with type 2 diabetes, but the control subjects were unmatched for habitual physical activity. The present data agree with a recent study (9) reporting no difference in basal ATP use between age-, weight-, and physical activity–matched subjects with and without type 2 diabetes using saturation transfer magnetic resonance. This study was only able to demonstrate differences in ATP flux in people with type 2 diabetes under insulin-stimulated conditions. The observation of no abnormality in basal ATP flux in diabetes implies that abnormal basal mitochondrial function is unlikely to be a primary causative factor in type 2 diabetes.

sonance. This study was only able to demonstrate differences in ATP flux in people with type 2 diabetes under insulin-stimulated conditions. The observation of no abnormality in basal ATP flux in diabetes implies that abnormal basal mitochondrial function is unlikely to be a primary causative factor in type 2 diabetes. In contrast to basal ATP synthesis rates, which are primarily influenced by steady-state energy demand, the recovery of PCr from exercise is a robust measure of maximal oxidative ATP turnover (19). This also reflects the recovery from muscular activity that happens frequently throughout the waking day, and any abnormality could bring about marked differences in muscle metabolism that may be associated with type 2 diabetes. The present data show no differences in the recovery of PCr from exercise between people with and without type 2 diabetes when controlled for habitual physical activity, weight, and age. To our knowledge, the present study is the first report to include both measures of basal and maximal ATP turnover in people with or without diabetes. No correlation was found between the basal and maximal ATP turnover rates. The lack of relationship between basal and maximal ATP turnover supports the concept that ATP turnover in these tests is determined by different factors. In the fasted, nonexercise state, the level of insulin-stimulated glucose uptake is likely to dominate requirement for ATP synthesis. This suggestion is based on studies that show a robust relationship between ATP synthesis and insulin-stimulated glucose uptake in skeletal muscle (9). In the postexercise stimulated state, the ability to supply and use oxygen is likely to dominate the rate of ATP turnover. The lack of correlation between the basal and maximal measures of ATP synthesis highlights the importance of examining separately these differing states.

ated glucose uptake in skeletal muscle (9). In the postexercise stimulated state, the ability to supply and use oxygen is likely to dominate the rate of ATP turnover. The lack of correlation between the basal and maximal measures of ATP synthesis highlights the importance of examining separately these differing states. A previous study (3) reported maximal ATP turnover to be reduced in people with type 2 diabetes compared with a BMI-matched control group. However, this study did not objectively control for differences in habitual physical activity, a factor that can influence mitochondrial function (13). People with type 2 diabetes tend to be less physically active than people without diabetes (11). Indeed, the reported PCr recovery data postexercise of people with type 2 diabetes (3) are comparable with both the type 2 diabetes and control groups in the present study, raising the likelihood that the differences in maximal ATP turnover may lay in differences in habitual physical activity. Recent support has also been given to the idea that type 2 diabetes is not necessarily associated with impaired mitochondrial function but may reflect differences in mitochondrial volume. Direct analyses of mitochondria from biopsies taken from people with type 2 diabetes show that any apparent defect in mitochondrial ATP production disappears when corrected for mitochondrial density (10). The reduced oxidative capacity accompanying type 2 diabetes may be the result of a deconditioning phenomenon (24).

Direct analyses of mitochondria from biopsies taken from people with type 2 diabetes show that any apparent defect in mitochondrial ATP production disappears when corrected for mitochondrial density (10). The reduced oxidative capacity accompanying type 2 diabetes may be the result of a deconditioning phenomenon (24). The present study shows that walking an extra 45 min per day over an 8-week period is an insufficient stimulus to induce detectable mitochondrial biogenesis. The physical activity was deliberately chosen to be of low intensity, as walking has been shown to be achievable and sustainable by people with type 2 diabetes (15,25). More intensive and prolonged physical activity and diet does change mitochondrial density and aerobic capacity (14). These changes correlate well with improvements in long-term glucose control and fasting insulin sensitivity. Other biopsy data suggest that the beneficial effect of exercise and moderate weight loss upon mitochondrial density is modest (26). However, the in vitro function of mitochondria is improved, with a disproportionate increase in electron transfer chain activity following intervention. The improvements of mitochondrial function accompanying exercise are not replicated by weight loss alone (27), stressing the importance of exercise in modifying oxidative capacity and maintaining metabolic flexibility. Further work is required to define the long-term effects of practically sustainable physical activity on mitochondrial function in muscle.

ction accompanying exercise are not replicated by weight loss alone (27), stressing the importance of exercise in modifying oxidative capacity and maintaining metabolic flexibility. Further work is required to define the long-term effects of practically sustainable physical activity on mitochondrial function in muscle. In type 2 diabetes changes in mitochondrial capacity are intertwined with changes in lipid oxidation (7). The present data demonstrate physical activity–induced enhancement of resting lipid oxidation, independent of intramuscular lipid levels. Type 2 diabetes is characterized by both abnormal lipid storage and oxidation, with glucose control commonly reported to be negatively related to intramuscular lipid content (2,8,28) via an effect on insulin action (5). Walking for an extra 45 min each day increases skeletal muscle mRNA expression of genes implicated in glucose and lipid metabolism (25). The cumulative effect of a sustained increase in lipid oxidation and decrease in IMCL would be expected to improve blood glucose control (15). The effects of increased physical activity upon circulating triglyceride turnover, and the consequential influence upon insulin action, remain to be determined.

metabolism (25). The cumulative effect of a sustained increase in lipid oxidation and decrease in IMCL would be expected to improve blood glucose control (15). The effects of increased physical activity upon circulating triglyceride turnover, and the consequential influence upon insulin action, remain to be determined. An important aspect of this study is that people with type 2 diabetes were able to sustain a more physically active lifestyle without supervision, and this would be expected to influence metabolic risk and glucose control (29). The study was not powered to detect changes in glucose control in the basal state, as this was not a primary objective. However, measures of glucose control were lower following increased physical activity. Other dynamic testing methods may have been more sensitive to changes in insulin sensitivity than homeostasis model assessment. These data are in line with larger, better powered studies of walking interventions (12). No change in serum triglycerides occurred in either group (data not shown). It is also possible that the various motivations for taking part in this study produced differences in the physical activity behavior. It is notable that the diabetic individuals sustained the level of physical activity better than the nondiabetic control group. The challenge ahead is to better understand how we can engage people with type 2 diabetes in reducing sedentary periods as well as to define how the underlying physiological mechanisms produce these benefits.

at the diabetic individuals sustained the level of physical activity better than the nondiabetic control group. The challenge ahead is to better understand how we can engage people with type 2 diabetes in reducing sedentary periods as well as to define how the underlying physiological mechanisms produce these benefits. The magnetic resonance methods applied here, and previously (2,3,9,17), are not without limitation. As the mitochondria are not isolated, ATP production may be limited by external factors, such as the supply of oxygen or ATP demand. However, with these limitations noted, it is clear that these noninvasive techniques provide a patient-friendly methodology that can be applied serially and complements more detailed in vitro techniques. In summary, resting and maximal ATP turnover are not impaired in people with well-controlled type 2 diabetes, when compared with control subjects matched for physical activity as well as age and weight. Increased daily physical activity in the form of walking is sustainable and improves lipid oxidation independent of mitochondrial activity in people with type 2 diabetes. The study was funded by the Wellcome Trust (grant no. 073561). M.I.T. is supported by a Diabetes U.K. RD Lawrence Fellowship. The authors are most grateful to the volunteers, Sister Jean Gerrard, Louise Morris, and Carol Smith.

Clinical trials have demonstrated that lifestyle intervention and pharmacological therapy in high-risk individuals reduce the incidence of type 2 diabetes (1). Thus, reliable models for identification of individuals at high risk for future type 2 diabetes are essential and have important clinical implications for intervention programs. Subjects with impaired glucose tolerance (IGT) are at increased risk for future type 2 diabetes (2), and the oral glucose tolerance test (OGTT) has become the standard method for identifying individuals at risk for type 2 diabetes. Indeed, all clinical trials that have assessed strategies for type 2 diabetes prevention have recruited subjects with IGT. Although IGT subjects have increased risk for type 2 diabetes, only ∼50% convert to type 2 diabetes within 10 years of follow-up (2), indicating that the future risk for diabetes is not similar among all individuals with IGT. Furthermore, in longitudinal epidemiological studies, ∼40% of subjects who develop type 2 diabetes have normal glucose tolerance (NGT) at baseline, indicating that there is a population of NGT subjects who are at risk for future type 2 diabetes (2). Recently, we demonstrated that subjects with NGT, despite having relatively low risk for type 2 diabetes, can be stratified into low- and high-risk categories based upon the relationship between their postload and fasting plasma glucose (FPG) concentrations (3).

subjects who are at risk for future type 2 diabetes (2). Recently, we demonstrated that subjects with NGT, despite having relatively low risk for type 2 diabetes, can be stratified into low- and high-risk categories based upon the relationship between their postload and fasting plasma glucose (FPG) concentrations (3). Several models have been proposed to improve the predictive ability for future type 2 diabetes (4–7). These models are based upon established risk factors for type 2 diabetes (e.g., obesity, FPG, lipid profile, and blood pressure). All of these risk factors are components of the metabolic or insulin resistance syndrome, which is itself a predictor of future type 2 diabetes in nondiabetic individuals (8). In a recent publication (9), we demonstrated that the 1-h plasma glucose concentration is a better predictor for future type 2 diabetes than either the FPG or 2-h plasma glucose concentration. Furthermore, the addition of the 1-h plasma glucose concentration to a prediction model based on clinical parameters significantly improved the ability of the model to predict future type 2 diabetes (9). In this study, we have used the classification tree model (10) to stratify the risk for future type 2 diabetes in nondiabetic subjects based upon their 1-h plasma glucose concentration during the OGTT and the Adult Treatment Panel (ATP) III criteria for the metabolic syndrome. We demonstrate that a model based on the combination of 1-h plasma glucose concentration during the OGTT and the ATP III criteria for the metabolic syndrome improves the ability to predict the future risk for type 2 diabetes.

n during the OGTT and the Adult Treatment Panel (ATP) III criteria for the metabolic syndrome. We demonstrate that a model based on the combination of 1-h plasma glucose concentration during the OGTT and the ATP III criteria for the metabolic syndrome improves the ability to predict the future risk for type 2 diabetes. RESEARCH DESIGN AND METHODS All subjects were participants of the San Antonio Heart Study (11–13), which is a population-based, epidemiological study of type 2 diabetes and cardiovascular disease. A total of 2,616 eligible participants, who were free of type 2 diabetes at baseline, completed a 7- to 8-year follow-up examination and had their diabetes outcome determined with a repeat OGTT. Of 2,616 participants, 1,610 subjects had plasma glucose measurements at 0, 30, 60, and 120 min during the baseline OGTT and constitute the study population. The study was approved by the institutional review board of University of Texas Health Science Center at San Antonio. All subjects gave their written informed consent before participation.

10 subjects had plasma glucose measurements at 0, 30, 60, and 120 min during the baseline OGTT and constitute the study population. The study was approved by the institutional review board of University of Texas Health Science Center at San Antonio. All subjects gave their written informed consent before participation. Definition of variables and outcomes All studies were performed in a mobile clinic following a 12-h overnight fast. A standard 75-g glucose OGTT was performed, and blood was obtained at 0, 30, 60, and 120 min for determination of plasma glucose and serum insulin concentrations. Plasma glucose and serum lipids were measured with an Abbott Bichromatic Analyzer (South Pasadena, CA). The diagnosis of diabetes was based upon World Health Organization criteria (14). Subjects on insulin or oral antihyperglycemic medications also were considered to have diabetes. The metabolic syndrome was diagnosed according to ATP III criteria (15).

e measured with an Abbott Bichromatic Analyzer (South Pasadena, CA). The diagnosis of diabetes was based upon World Health Organization criteria (14). Subjects on insulin or oral antihyperglycemic medications also were considered to have diabetes. The metabolic syndrome was diagnosed according to ATP III criteria (15). Classification tree Recursively partitioned classification trees (16) were used to model the relationship between the future risk of type 2 diabetes and 1) 1-h plasma glucose concentration during the OGTT and 2) presence or absence of the metabolic syndrome. Sequential partitioning of the individuals based upon their 1-h plasma glucose concentration relative to 155 mg/dl (above or below) and the presence or absence of the metabolic syndrome produced subgroups or compartments of individuals with homogenous risk for future type 2 diabetes. Subgroups with annual risk for future type 2 diabetes <0.5% (<3.5% risk in 7–8 years) were considered as having low risk for future type 2 diabetes. Annual risk between 1 and 2% (7–15% risk in 7–8 years) was considered intermediate risk. Annual risk >4% (>30% risk in 7–8 years) was considered high risk.

ype 2 diabetes. Subgroups with annual risk for future type 2 diabetes <0.5% (<3.5% risk in 7–8 years) were considered as having low risk for future type 2 diabetes. Annual risk between 1 and 2% (7–15% risk in 7–8 years) was considered intermediate risk. Annual risk >4% (>30% risk in 7–8 years) was considered high risk. Statistical methods Variables are presented as the means ± SD. The significance of the mean differences was tested with ANOVA. Differences between categorical variables were tested with the χ2 test. Statistical significance was considered at the level of P < 0.05. Assessment of the predictive discrimination of the various models was made using the receiver-operating characteristic curve by plotting the sensitivity against the corresponding false-positive rate. Statistical analysis was performed with the SPSS software package.

significance was considered at the level of P < 0.05. Assessment of the predictive discrimination of the various models was made using the receiver-operating characteristic curve by plotting the sensitivity against the corresponding false-positive rate. Statistical analysis was performed with the SPSS software package. RESULTS Table 1 presents the anthropometric, laboratory, and clinical characteristics of the study population. Of 1,611 study participants, 1,301 had NGT, 90 had impaired fasting glucose (IFG), and 221 had IGT at baseline, respectively. Fifty-one of 221 subjects with IGT also had IFG and were designated as having combined glucose intolerance (CGI). The conversion rate to type 2 diabetes over the study period (7–8 years) was 5.0, 26.1, 30.9, and 82.3% for NGT, IFG, IGT, and CGI subjects, respectively. We previously demonstrated that the 1-h plasma glucose concentration during the OGTT is a good predictor for future type 2 diabetes (9). A plasma glucose cutoff point of 155 mg/dl has the maximal sum of sensitivity and specificity (0.75 and 0.79 for sensitivity and specificity, respectively) and for predicting future type 2 diabetes. Similarly, the ideal cutoff point for fasting plasma glucose concentration in predicting future type 2 diabetes was 94.5 mg/dl. Therefore, we have used these values as cutoff points to test the prediction of future type 2 diabetes with two tree models.

ecificity, respectively) and for predicting future type 2 diabetes. Similarly, the ideal cutoff point for fasting plasma glucose concentration in predicting future type 2 diabetes was 94.5 mg/dl. Therefore, we have used these values as cutoff points to test the prediction of future type 2 diabetes with two tree models. The first tree model is based upon the glucose tolerance status (17), 1-h plasma glucose value, and presence of the metabolic syndrome. The receiver-operating characteristic for this model was 86.7%. In this model, individuals were divided, according to the American Diabetes Association criteria (17), into four groups (NGT, IFG, IGT, and CGI) based upon their fasting and 2-h plasma glucose concentration. Individuals in each group were further divided into two subgroups based upon their 1-h plasma glucose concentration (above or below 155 mg/dl). Figure 1 depicts the incidence of type 2 diabetes based upon 1-h plasma glucose concentration. Although, as a whole, subjects with NGT had a low risk for type 2 diabetes (5.0%), normal glucose-tolerant subjects with 1-h plasma glucose >155 mg/dl had significantly increased risk (15.3%) for future type 2 diabetes compared with nor subjects with 1-h plasma glucose <155 mg/dl (2.9%) (P < 0.0001). Further division of this group based upon the presence or absence of the metabolic syndrome demonstrated that NGT subjects with 1-h plasma glucose >155 mg/dl and the metabolic syndrome had a 32.1% incidence rate of type 2 diabetes compared with a 9.4% incidence rate for subjects without the metabolic syndrome.

0.0001). Further division of this group based upon the presence or absence of the metabolic syndrome demonstrated that NGT subjects with 1-h plasma glucose >155 mg/dl and the metabolic syndrome had a 32.1% incidence rate of type 2 diabetes compared with a 9.4% incidence rate for subjects without the metabolic syndrome. Subjects with IFG and a 1-h plasma glucose >155 mg/dl had a 37.3% incidence of type 2 diabetes, while IFG subjects with a 1-h plasma glucose concentration <155 mg/dl had a 10.8% incidence rate. Table 2 presents the odds ratio for having diabetes for the various glucose tolerance groups. Subjects with IGT and a 1-h plasma glucose >155 mg/dl had a 35.5% diabetes incidence rate, while IGT subjects with a 1-h plasma glucose <155 mg/dl had a 17.8% diabetes incidence rate.

ion <155 mg/dl had a 10.8% incidence rate. Table 2 presents the odds ratio for having diabetes for the various glucose tolerance groups. Subjects with IGT and a 1-h plasma glucose >155 mg/dl had a 35.5% diabetes incidence rate, while IGT subjects with a 1-h plasma glucose <155 mg/dl had a 17.8% diabetes incidence rate. The second model includes the 1-h plasma glucose concentration, the metabolic syndrome, and fasting plasma glucose concentration. The receiver-operating characteristic for this model was 85.4%. In this model, subjects were divided into two groups based upon their 1-h plasma glucose concentration (above or below 155 mg/dl) and each group was further divided into two subgroups based upon the presence or absence of the metabolic syndrome. Figure 2 depicts the 7- to 8-year risk for type 2 diabetes for each subgroup. In general, nondiabetic subjects with 1-h plasma glucose <155 mg/dl had a low risk (3.9%) for future development of type 2 diabetes compared with subjects with a 1-h plasma glucose >155 mg/dl (31.0%) (P < 0.0001). When subjects with 1-h plasma glucose <155 mg/dl were divided according to the presence or absence of the metabolic syndrome, subjects with a 1-h plasma glucose <155 mg/dl without the metabolic syndrome had a 3.2% risk for future type 2 diabetes, while those with the metabolic syndrome had 7.8% risk for future diabetes. Subjects with a 1-h plasma glucose concentration >155 mg/dl and the metabolic syndrome had a 51.6% risk for future diabetes. Subjects with a 1-h plasma glucose >155 mg/dl without the metabolic syndrome, but with a fasting plasma glucose >95 mg/dl, had a 44.7% risk for future diabetes, while subjects with a 1-h plasma glucose >155 mg/dl without the metabolic syndrome and fasting plasma glucose <95 mg/dl had a 10.8% risk for future type 2 diabetes.

cts with a 1-h plasma glucose >155 mg/dl without the metabolic syndrome, but with a fasting plasma glucose >95 mg/dl, had a 44.7% risk for future diabetes, while subjects with a 1-h plasma glucose >155 mg/dl without the metabolic syndrome and fasting plasma glucose <95 mg/dl had a 10.8% risk for future type 2 diabetes. Because the waist circumference is rarely measured in clinical practice and is part of the ATP III definition of the metabolic syndrome, we also examined the predictive value of triglyceride–to–HDL cholesterol ratio >3.5 in place of the metabolic syndrome (Table 2). Although the metabolic syndrome was a better predictor compared with the triglycerides–to–HDL cholesterol ratio, a model based on 1-h plasma glucose concentration and triglyceride–to–HDL cholesterol ratio could classify subjects to three risk groups: low, intermediate, and high risk (Table 2).

bolic syndrome (Table 2). Although the metabolic syndrome was a better predictor compared with the triglycerides–to–HDL cholesterol ratio, a model based on 1-h plasma glucose concentration and triglyceride–to–HDL cholesterol ratio could classify subjects to three risk groups: low, intermediate, and high risk (Table 2). CONCLUSIONS The American Diabetes Association Consensus Statement has recommended metformin, in addition to diet and exercise, in individuals with IGT/IFG to reduce their risk for future diabetes (18). This recommendation for pharmacologic intervention underscores the need for models that reliably identify individuals at increased risk for future development of type 2 diabetes. The results of this study demonstrate that the plasma glucose concentration at 1 h during the OGTT is a useful tool that can be used to stratify the risk of future type 2 diabetes into three groups: low, intermediate, and high risk. In general, subjects with NGT have low risk for progression to type 2 diabetes (∼0.67% annual rate) (2). However, ∼40% of individuals who develop type 2 diabetes have NGT at baseline (2) and, in the present study, 16.7% of normal glucose-tolerant subjects with a 1-h plasma glucose concentration (OGTT) >155 mg/dl developed type 2 diabetes over a 7- to 8-year period. In this group of normal glucose-tolerant subjects, the annual risk for future type 2 diabetes was significantly greater (2.2% per year) compared with subjects whose 1-h plasma glucose concentration did not exceed 155 mg/dl (0.39% per year, P < 0.00001). Further, NGT subjects with a 1-h plasma glucose >155 mg/dl who fulfilled the ATP III criteria for the metabolic syndrome had a 4.3% annual risk for future type 2 diabetes. Thus, the group of normal glucose-tolerant subjects with 1-h PG >155 mg/dl plus the metabolic syndrome is at very high risk for the development of type 2 diabetes, their risk exceeds that of subjects with IFG or IGT, and their odds ratio for developing diabetes is double that of IGT subjects with a 1-h plasma glucose <155 mg/dl (Table 2). Consistent with the American Diabetes Association Consensus Conference Statement (18), this group of high-risk NGT individuals could benefit from an intervention program employing diet, exercise, and pharmacotherapy (metformin) to reduce future risk for diabetes.

subjects with a 1-h plasma glucose <155 mg/dl (Table 2). Consistent with the American Diabetes Association Consensus Conference Statement (18), this group of high-risk NGT individuals could benefit from an intervention program employing diet, exercise, and pharmacotherapy (metformin) to reduce future risk for diabetes. Subjects with CGI have the greatest risk for future type 2 diabetes, with an annual risk >10% per year, while subjects with isolated IFG or IGT have an intermediate risk between CGI and NGT. However, within the IFG and IGT groups, the 1-h plasma glucose during the OGTT also stratifies the future diabetes risk into intermediate and high risk. Thus, IFG and IGT subjects with a 1-h plasma glucose <155 mg/dl have an annual risk of ∼1.5% compared with an annual risk of ∼5% for IGT and IFG subjects with a 1-h plasma glucose >155 mg/dl. It is noteworthy that every CGI subject had a 1-h plasma glucose concentration >155 mg/dl. Thus, the plasma glucose concentration at 1 h during the OGTT is a strong predictor for future type 2 diabetes, independent of the glucose tolerance status, and a 155 mg/dl cutoff point divides individuals with NGT, IFG, and IGT into low-, intermediate-, and high-risk groups.

-h plasma glucose concentration >155 mg/dl. Thus, the plasma glucose concentration at 1 h during the OGTT is a strong predictor for future type 2 diabetes, independent of the glucose tolerance status, and a 155 mg/dl cutoff point divides individuals with NGT, IFG, and IGT into low-, intermediate-, and high-risk groups. A predictive model based on the plasma glucose concentration at 1 h during the OGTT and the presence or absence of the metabolic syndrome, independent of the 2-h plasma glucose concentration, performs equally well in stratifying subjects for future risk of type 2 diabetes compared with the model that includes the 2-h plasma glucose concentration. The earlier model had 0.82 sensitivity and 0.63 specificity compared with 0.82 and 0.67 sensitivity and specificity, respectively, for the model based on 1-h plasma glucose concentration. Moreover, the later model (individuals with 1-h plasma glucose >155 mg/dl plus the metabolic syndrome or FPG >95 mg/dl) reduces the number of subjects in the very-high-risk group (>6.5% incidence per year), who are candidates for pharmacological intervention, from 18% (based on the model that includes the 2-h plasma glucose concentration) to 14% of the total study population. Furthermore, the model with the 1-h plasma glucose concentration plus the metabolic syndrome performs better in predicting future diabetes than does the American Diabetes Association criteria of IGT or IFG. Most importantly, ∼17% of normal glucose-tolerant subjects, who have intermediate and high risk for future type 2 diabetes and who were identified with the 1-h plasma glucose plus metabolic syndrome, would have been missed with the American Diabetes Association criteria alone. These observations underscore the importance of obtaining the plasma glucose concentration at 1 h during the OGTT.

ermediate and high risk for future type 2 diabetes and who were identified with the 1-h plasma glucose plus metabolic syndrome, would have been missed with the American Diabetes Association criteria alone. These observations underscore the importance of obtaining the plasma glucose concentration at 1 h during the OGTT. Substituting the metabolic syndrome with the triglyceride–to–HDL cholesterol ratio in the second model slightly reduces its predictability. However, the second model with the triglyceride–to–HDL cholesterol ratio is a good predictor for future risk of type 2 diabetes and classifies subjects into three risk groups. Because measurement of triglyceride and HDL cholesterol is part of the routine clinical practice, the second model could be used in routine clinical practice to assess the risk of nondiabetic subjects for future risk of type 2 diabetes.

for future risk of type 2 diabetes and classifies subjects into three risk groups. Because measurement of triglyceride and HDL cholesterol is part of the routine clinical practice, the second model could be used in routine clinical practice to assess the risk of nondiabetic subjects for future risk of type 2 diabetes. Why is the 1-h plasma glucose concentration a better predictor for future type 2 diabetes than the 2-h plasma glucose? It could be argued that the high predictability for 1-h plasma glucose is due to its high correlation with the 2-h plasma glucose (r = 0.58, P < 0.0001). However, the 1-h plasma glucose stratifies subjects with NGT, as well as subjects with IGT, into two risk groups, high and low. Thus, it is unlikely that its predictability is secondary to its correlation with the 2-h plasma glucose. Subjects who are destined to develop type 2 diabetes manifest two major defects: 1) insulin resistance in liver and skeletal muscle and 2) impaired β-cell function (19). Previous studies have demonstrated that subjects with hepatic insulin resistance have an increased FPG concentration and impaired suppression of hepatic glucose production during the OGTT, resulting in an excessive rise in plasma glucose concentration at 30 and 60 min (20). In nondiabetic subjects, the decline in plasma glucose concentration at 30–60 min during the OGTT is dependent on insulin sensitivity in skeletal muscle and β-cell function (21,22). Thus, insulin resistance in liver and skeletal muscle, as well as impaired β-cell function, would result in an increase in 1-h plasma glucose concentration. This renders the 1-h plasma glucose a good indicator for the major metabolic abnormalities that lead to the development of type 2 diabetes. Consistent with this, we previously demonstrated that the plasma glucose concentration at 1 h during the OGTT has a stronger correlation with surrogate measures of hepatic and muscle insulin resistance and β-cell dysfunction compared with the 2-h plasma glucose value (9).

In humans, uric acid is the end product of purine metabolism, and hyperuricemia might be caused by an overproduction of or by disturbances in the elimination of uric acid. Although some experimental evidence supports the idea of a beneficial role of uric acid because of its strong antioxidative properties (1), a number of epidemiological studies reported an association between high uric acid levels and cardiovascular disease, hypertension, kidney disease, metabolic syndrome, and even total mortality (2,3). Uric acid levels are, furthermore, positively associated with serum glucose in healthy subjects (4), and subjects with higher uric acid levels have a higher risk of developing type 2 diabetes or metabolic syndrome (5). However, the mechanisms underlying this association are still unclear. The regulation of uric acid levels is under strong genetic control, with heritability estimates ranging from 25 to70% (6,7). The elucidation of the genetic contributors to uric acid levels as an intermediate phenotype for various diseases might shed some light on the pathogenesis of these complex phenotypes and might help identify new targets for treating undesirably high uric acid levels.

heritability estimates ranging from 25 to70% (6,7). The elucidation of the genetic contributors to uric acid levels as an intermediate phenotype for various diseases might shed some light on the pathogenesis of these complex phenotypes and might help identify new targets for treating undesirably high uric acid levels. Recent genome-wide association studies identified a strong association of uric acid levels with genetic variants within SLC2A9 (8–11), a gene located in the chromosomal region 4p16.1 encoding a putative fructose transporter. There is strong evidence from both animal models and human studies supporting fructose as a highly lipogenic nutrient that contributes to tissue insulin insensitivity, metabolic disturbances, and the development of a prediabetic state when consumed in high quantities (12). In the present study we aimed to replicate the recently discovered association between genetic variation within the SLC2A9 gene and uric acid concentrations in the population-based Bruneck Study. This analysis was extended by a large case-control study of severely obese individuals from Utah who showed increased uric acid levels compared with those in control subjects to explore whether the genetic association is modified by obesity.

LC2A9 gene and uric acid concentrations in the population-based Bruneck Study. This analysis was extended by a large case-control study of severely obese individuals from Utah who showed increased uric acid levels compared with those in control subjects to explore whether the genetic association is modified by obesity. RESEARCH DESIGN AND METHODS Bruneck Study The Bruneck Study is a prospective population-based survey designed to investigate the epidemiology and pathogenesis of atherosclerosis (13). Briefly, the study population was recruited as a sex- and age-stratified random sample of all inhabitants of Bruneck, Italy (125 women and 125 men in each decade of age from the fifth to the eighth, n = 1,000). At the 1990 baseline, 93.6% of recruited subjects participated, with data assessment completed in 919 subjects. Follow-up examinations were performed in 1995, 2000, and 2005. Detailed information on prevalent and incident metabolic syndrome components, diabetes, and cardiovascular events is available from all examinations. The present analysis focuses on the 1995 reexamination and on the follow-up period for clinical events between 1995 and 2005. In 1995, the study population consisted of 826 subjects (96.5% of those alive). Sufficient DNA was available for 800 participants.

, and cardiovascular events is available from all examinations. The present analysis focuses on the 1995 reexamination and on the follow-up period for clinical events between 1995 and 2005. In 1995, the study population consisted of 826 subjects (96.5% of those alive). Sufficient DNA was available for 800 participants. Obesity case-control study from Utah The study included 1,869 individuals from two groups of subjects gathered in the state of Utah. The study population was composed of 1,038 subjects recruited for severe obesity (“severe obesity group” with a BMI between 33 and 92 kg/m2) and a general population sample of 831 individuals of the same ethnicity (“control subjects”). The two groups of subjects were described in detail elsewhere (14). In brief, the 1,038 subjects with severe obesity were either seeking gastric bypass surgery or were randomly chosen from a population-based sample of severely obese participants. The examination of patients undergoing gastric bypass surgery was done before the intervention. The control group consisted of 831 individuals from the same geographical region and was found to be representative of the Utah population, spanning the entire BMI range.

pulation-based sample of severely obese participants. The examination of patients undergoing gastric bypass surgery was done before the intervention. The control group consisted of 831 individuals from the same geographical region and was found to be representative of the Utah population, spanning the entire BMI range. Metabolic syndrome and type 2 diabetes Prevalent type 2 diabetes was considered to be present if a prior physician diagnosis had been made, if the fasting blood glucose upon screening was ≥126 mg/dl, or if insulin-sensitizing agents or diabetes medications were being taken by the individual. Metabolic syndrome was defined according to the scientific statement from the American Heart Association and the National Heart, Lung, and Blood Institute (15). The incidence of type 2 diabetes and metabolic syndrome was assessed in the Bruneck Study from the 1995 examination to the 2005 examination. Laboratory methods In both study populations, blood samples were collected after an overnight fasting period. Uric acid levels were measured using enzymatic-colorimetric methods (Bruneck Study: Merck, Vienna, Austria; Utah study: Roche, Indianapolis, IN) with intra-assay and interassay coefficients of variation <2%. Other clinical chemical parameters were measured as described recently (14).

after an overnight fasting period. Uric acid levels were measured using enzymatic-colorimetric methods (Bruneck Study: Merck, Vienna, Austria; Utah study: Roche, Indianapolis, IN) with intra-assay and interassay coefficients of variation <2%. Other clinical chemical parameters were measured as described recently (14). Four single nucleotide polymorphisms (SNPs) with genome-wide P values <10−7 in the data from Li et al. (8) and Döring et al. (10) were selected for genotyping using a 5′ nuclease allelic discrimination (TaqMan) assay (Applied Biosystems, Foster City, CA): rs6855911, rs7442295, rs6449213, and rs12510549. Genotyping was done within the Genotyping Unit of the Gene Discovery Core Facility at Innsbruck Medical University (Innsbruck, Austria).

al. (10) were selected for genotyping using a 5′ nuclease allelic discrimination (TaqMan) assay (Applied Biosystems, Foster City, CA): rs6855911, rs7442295, rs6449213, and rs12510549. Genotyping was done within the Genotyping Unit of the Gene Discovery Core Facility at Innsbruck Medical University (Innsbruck, Austria). Statistical and bioinformatic analysis To compare characteristics between individual groups, we applied t tests, Wilcoxon tests, and Pearson χ2 tests. Spearman correlation coefficients were used to describe the correlation between uric acid levels and components of the metabolic syndrome. A χ2 test for violation of the Hardy-Weinberg equilibrium was performed. General linear regression models were used to estimate the association of any of the four SNPs with uric acid levels adjusted for various covariates, assuming an additive model. The additive model was applied because of a priori evidence from previous studies in various populations (8,10,11) and from the inspection of the genotype-specific means of uric acid levels. Interactions between the SNPs investigated and BMI on uric acid levels were tested by adding an interaction term (SNP × BMI) and the main covariates (sex, SNP, and BMI) to the model. The effect modification was graphically illustrated for three BMI classes (<30, 30–40, and >40 kg/m2) and the three genotype levels using interaction plots applying PROC GENMOD. All statistical analyses were performed with SPSS (version 15.0; SPSS, Chicago, IL) or SAS (version 9.1; SAS Institute, Cary, NC).

the model. The effect modification was graphically illustrated for three BMI classes (<30, 30–40, and >40 kg/m2) and the three genotype levels using interaction plots applying PROC GENMOD. All statistical analyses were performed with SPSS (version 15.0; SPSS, Chicago, IL) or SAS (version 9.1; SAS Institute, Cary, NC). Because all polymorphisms that we investigated were located in noncoding regions, their possible effect on transcription factor binding sites was evaluated in silico using different components of the Genomatix Software Suite (Genomatix, Munich, Germany). Analyses included a search for the presence of known functional genetic elements (e.g., promoter sequences or microRNAs) using Eldorado (release 4.5), a search for unknown promoter-specific sequences using Promoter Inspector (release 4.6), and an investigation of possible effects of the polymorphisms on single transcription factor binding sites using SNPInspector (release 4.6). RESULTS Baseline clinical characteristics and laboratory data of the study groups from Bruneck and Utah are reported in Table 1. Analysis of the Utah group was also stratified by case-control status (severe obesity versus control). All four SNPs were found to be in Hardy-Weinberg equilibrium in all groups analyzed (P > 0.1), and the genotyping efficiency ranged between 97 and 98.5%. We observed no difference in genotype frequencies between the Bruneck and Utah groups (P > 0.1).

RESULTS Baseline clinical characteristics and laboratory data of the study groups from Bruneck and Utah are reported in Table 1. Analysis of the Utah group was also stratified by case-control status (severe obesity versus control). All four SNPs were found to be in Hardy-Weinberg equilibrium in all groups analyzed (P > 0.1), and the genotyping efficiency ranged between 97 and 98.5%. We observed no difference in genotype frequencies between the Bruneck and Utah groups (P > 0.1). Association of genetic variation within SLC2A9 and uric acid levels Age- and sex-adjusted linear regression models revealed a strong and highly significant association of each of the four SNPs with uric acid levels (Table 2) in all study groups, which was strongest in the population-based Bruneck Study. This association can clearly be described by an additive model: each copy of the minor allele lowered uric acid levels by 0.30–0.35 mg/dl on average, representing a relative decrease of 5–6%, with P values ranging from 10−9 to 10−11 in the combined analysis. Sex-specific associations between SLC2A9 and uric acid levels A sex-stratified analysis revealed much stronger associations between the four SNPs and uric acid levels in women than in men (Table 2). This observation was most pronounced in the population-based Bruneck Study, in which the effect estimates for each minor allele were up to twice as high in women than in men. Considering the fact that women have on average 20% lower uric acid levels than men, the effect differences get even more meaningful.

(Table 2). This observation was most pronounced in the population-based Bruneck Study, in which the effect estimates for each minor allele were up to twice as high in women than in men. Considering the fact that women have on average 20% lower uric acid levels than men, the effect differences get even more meaningful. Interaction with BMI The associations of the four SNPs with uric acid levels became stronger after additional adjustments for BMI, creatinine, gout medication, and alcohol intake, with P values ranging from 10−14 to 10−20 (Table 2). BMI was the variable that most contributed to the change in P values. Thus, we investigated whether BMI was an effect modifier of the SNP–uric acid association by introducing an interaction term BMI × SNP in addition to BMI and SNP to the model. This analysis was done by combining all three groups, but with adjustments for age, sex, and population (Bruneck versus Utah). The interaction was significant for three of the four SNPs (P = 0.023–0.035) and borderline significant for rs12510549 (P = 0.053). Figure 1 graphically illustrates these interactions for BMI groups <30, 30–40, and >40 kg/m2.

ing all three groups, but with adjustments for age, sex, and population (Bruneck versus Utah). The interaction was significant for three of the four SNPs (P = 0.023–0.035) and borderline significant for rs12510549 (P = 0.053). Figure 1 graphically illustrates these interactions for BMI groups <30, 30–40, and >40 kg/m2. Association with metabolic syndrome and type 2 diabetes We observed significant correlations between uric acid levels and each singular metabolic syndrome component, as well as with the sum of the components (Table 3). These correlations were seen in all three groups of subjects. Age- and sex-adjusted uric acid levels were significantly higher in subjects with metabolic syndrome compared with those without. With the exception of the severe obesity group, uric acid levels were also significantly higher in subjects with type 2 diabetes compared with those without (Table 3). Uric acid levels were significantly associated with the development of either metabolic syndrome and/or type 2 diabetes in the Bruneck Study from the 1995 examination to the 2005 examination. However, no association between the four SNPs and prevalent or incident metabolic syndrome or type 2 diabetes could be established.

c acid levels were significantly associated with the development of either metabolic syndrome and/or type 2 diabetes in the Bruneck Study from the 1995 examination to the 2005 examination. However, no association between the four SNPs and prevalent or incident metabolic syndrome or type 2 diabetes could be established. Bioinformatics Putative transcription factor binding sites were predicted for three of the four polymorphisms (five for rs12510549, two for rs6449213, and one for rs7442295), although no biologically obvious candidates could be found. The predicted candidates belonged mainly to pathways not directly connected with the phenotypes investigated, such as cellular growth (E2F) or immune response (PAX5, BCL6, NFAT, and NR2F) (data not shown). For the polymorphism rs12510549, the generation or disruption of five different putative binding sites was predicted. Among these, a putative binding site for NFAT factors was recognized. NFAT factors have been implicated in various roles during development and adaptation of several mammalian cell types outside the immune system (16). CONCLUSIONS We observed a strong sex-specific association between genetic variation within the SLC2A9 gene and uric acid concentrations. The finding was most pronounced in the population-based Bruneck Study and was replicated in severely obese and control individuals from Utah. This association was modified by BMI such that increasing BMI amplified effects of genetic variants on uric acid levels.

tion within the SLC2A9 gene and uric acid concentrations. The finding was most pronounced in the population-based Bruneck Study and was replicated in severely obese and control individuals from Utah. This association was modified by BMI such that increasing BMI amplified effects of genetic variants on uric acid levels. SLC2A9 was recently identified by four independent genome-wide association studies to be strongly associated with uric acid levels (8–11). SLC2A9 encodes a putative hexose transporter whose probable substrate is fructose (17). Fructose intake has been described as an important contributor to uric acid levels and gout (18,19), as the ADP generated during the phosphorylation of fructose is used for rapid production of uric acid (20). Epidemiological data showed that increased total fructose intake correlated with increasing incidence of obesity, metabolic syndrome (21), and gout (19). Over the past decades, a general increase in uric acid levels was observed, and it was hypothesized that fructose-induced hyperuricemia might be in part responsible for the rise in metabolic syndrome (12,22,23). The detection of genes that determine uric acid levels by influencing fructose metabolism would therefore be of interest. The actual mechanism of how genetic variation within SLC2A9 modulates uric acid levels has not been fully elucidated. One possibility would be an influence on the hepatic uptake of fructose and production of uric acid. On the other hand, SLC2A9 variants were associated with low fractional excretion of uric acid in various population samples, and experiments in Xenopus laevis oocytes showed that SLC2A9 has not only fructose but also strong uric acid transport activity (11).

n the hepatic uptake of fructose and production of uric acid. On the other hand, SLC2A9 variants were associated with low fractional excretion of uric acid in various population samples, and experiments in Xenopus laevis oocytes showed that SLC2A9 has not only fructose but also strong uric acid transport activity (11). It is important to note that we did not find an association between the genetic variants within the SLC2A9 gene and prevalent or incident metabolic syndrome or type 2 diabetes. This finding is intriguing given the pronounced association between the genetic variants and uric acid levels and the strong association of uric acid levels and these diseases in our study and in earlier studies. This may by explained, on the one hand, by lack of power: the variance in serum uric acid levels related to the genotypes investigated in population-based studies was about 1.2% in men and 6% in women (10), and the fraction of these two diseases explained by uric acid levels was also small. On the other hand, the possibility that uric acid level is a surrogate marker of the disease without being in the causal pathway cannot be ruled out. However, recent studies in rats showed that fructose-induced metabolic syndrome is partially prevented by lowering uric acid levels and that the reduction in endothelial nitric oxide bioavailability caused by uric acid may be a mechanism for insulin resistance and hypertension (23). A proof of a causal association of uric acid with disease end points might be possible by the application of a Mendelian randomization approach. However, this proof will probably require examination of several thousand individuals. Homozygotes of the wild type and of the rare allele differ in uric acid levels by 0.64–0.81 mg/dl, which corresponds to 11–13% of the mean levels. Based on the findings in earlier studies, such a difference in uric acid levels would change the rate of cardiovascular events by 1%.

nation of several thousand individuals. Homozygotes of the wild type and of the rare allele differ in uric acid levels by 0.64–0.81 mg/dl, which corresponds to 11–13% of the mean levels. Based on the findings in earlier studies, such a difference in uric acid levels would change the rate of cardiovascular events by 1%. A recent study with a systematic investigation of sex-specific differences of literature-reported genetic effects on various phenotypes documented that only one of 432 sex difference claims was consistently replicated in at least two other studies (24). The association of genetic variants within SCL2A9 with uric acid levels clearly adds to this list, as much stronger associations were found in women than in men in five population samples in previous studies (10,11) and in three population samples in this study. How obesity modulates the association between SLC2A9 variants and uric acid levels remains to be determined. It could be related to the higher fructose intake in obese subjects (22) and a different saturation capacity of fructose transport depending on the genotype. In summary, our study shows a strong association of genetic variants within the SLC2A9 gene and uric acid levels that is modified by sex and BMI. This research was funded by grants from the “Genomics of Lipid-associated Disorders–GOLD” of the Austrian Genome Research Programme GEN-AU to F.K., by a grant from the German National Genome Research Net to the GSF-Institute of Epidemiology, and by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (DK-55006) to S.C.H.

nts from the “Genomics of Lipid-associated Disorders–GOLD” of the Austrian Genome Research Programme GEN-AU to F.K., by a grant from the German National Genome Research Net to the GSF-Institute of Epidemiology, and by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (DK-55006) to S.C.H. We thank Anke Gehringer and Markus Haak for excellent laboratory work.

Gestational diabetes mellitus (GDM) is a common condition affecting 2–4% of pregnant women (1) and is associated with adverse outcomes for both the fetus and the mother. Previous GDM is a major risk factor for type 2 diabetes, which occurs in 20–60% of affected women within 5 years of the pregnancy (2). Women with a history of GDM are also at increased risk of other cardiovascular risk factors, such as obesity, hypertension, dyslipidemia, and the metabolic syndrome (3–5), as well as subclinical atherosclerosis (6). Taken together, these findings suggest that GDM identifies a population of young women at increased risk for cardiovascular disease (CVD). We used population-based administrative data to determine whether women with GDM have a heightened risk for CVD compared with women without GDM and whether any increase in risk is independent of subsequent type 2 diabetes. RESEARCH DESIGN AND METHODS We conducted a population-based retrospective cohort study using administrative databases from Ontario, Canada, that included hospital discharge abstracts, physician service claims, and demographic data. The Ontario Diabetes Database is a validated registry of physician-diagnosed nongestational diabetes that is identified using these administrative data (7). Individuals are linked between all data sources via a unique health card number, which is reproducibly encrypted in all of these data sources.

nd demographic data. The Ontario Diabetes Database is a validated registry of physician-diagnosed nongestational diabetes that is identified using these administrative data (7). Individuals are linked between all data sources via a unique health card number, which is reproducibly encrypted in all of these data sources. Women aged 20–49 years who had a hospitalization record indicating a live birth between April 1994 and March 1997 were selected. For women who had more than one birth during this period, one birth was selected at random. Those who had pregestational diabetes or a CVD event (as defined below) in the prior 3 years were excluded. Baseline characteristics were age at delivery, region of residence, and socioeconomic status (measured as the neighborhood income quintile). Subjects with missing data were excluded. Women were defined as having GDM using an algorithm analogous to that used by the validated registry to exclude GDM: one hospitalization record or two ambulatory physician claims bearing the diagnosis of diabetes or GDM between 120 days before and 180 days after delivery.

ile). Subjects with missing data were excluded. Women were defined as having GDM using an algorithm analogous to that used by the validated registry to exclude GDM: one hospitalization record or two ambulatory physician claims bearing the diagnosis of diabetes or GDM between 120 days before and 180 days after delivery. The primary outcome (CVD events) was defined as a hospitalization for acute myocardial infarction, stroke, coronary artery bypass, coronary angioplasty, or carotid endarterectomy. The prespecified secondary outcome (coronary artery disease [CAD] events) was hospitalization for acute myocardial infarction, coronary artery bypass, or coronary angioplasty. Subsequent diagnosis with diabetes was identified if the woman entered the diabetes registry postpartum. Although the registry does not distinguish between types, the majority of women developing diabetes in this group would have type 2 diabetes. All women were followed until March 2007, with censoring on death.

lasty. Subsequent diagnosis with diabetes was identified if the woman entered the diabetes registry postpartum. Although the registry does not distinguish between types, the majority of women developing diabetes in this group would have type 2 diabetes. All women were followed until March 2007, with censoring on death. Subjects with GDM were matched with 10 subjects without GDM based on baseline characteristics. Kaplan-Meier survival curves were constructed for both outcomes. Cox proportional hazards regression was used to model the association of GDM with each outcome, accounting for the matched design of the study. For each outcome, an unadjusted model and a model adjusting for subsequent diagnosis of diabetes as a time-dependent covariate were built. The assumption of proportionality was verified by plotting log(–log[survival]) versus log(time) to assess parallelism. The study was approved by the institutional review board of Sunnybrook Health Sciences Centre. RESULTS There were 356,891 potentially eligible women with live births during the study period. However, 3,127 were excluded because of preexisting diabetes, and 43 were excluded because of previous CVD. A further 2,036 were excluded due to missing data, mostly socioeconomic status. Of the remaining 351,685 subjects, 8,194 (2.3%) had GDM during the index pregnancy. The matched cohorts included 8,191 women with GDM and 81,262 without GDM. The mean age of both cohorts was 31 years.

were excluded because of previous CVD. A further 2,036 were excluded due to missing data, mostly socioeconomic status. Of the remaining 351,685 subjects, 8,194 (2.3%) had GDM during the index pregnancy. The matched cohorts included 8,191 women with GDM and 81,262 without GDM. The mean age of both cohorts was 31 years. The median follow-up time was 11.5 years. Diabetes developed during follow-up in 2,214 (27.0%) of the women with GDM and 2,596 (3.2%) of the women without GDM. Event-free survival for both CVD and CAD events are plotted in Fig. 1. Significant associations were found between GDM and both outcomes, but these associations were attenuated following adjustment for subsequent diabetes. CONCLUSIONS Our study is the first of its kind to show that young women with GDM have a substantially increased risk for CVD relative to women without GDM. The subsequent development of type 2 diabetes accounts for much of this increased risk, which reinforces the vital need for diabetes prevention strategies in this high-risk population. Our findings are consistent with a cross-sectional study conducted by Carr et al. (5), which reported that women with a history of GDM had odds ratios for CVD and CAD similar to those reported here (1.85 and 1.58, respectively). However, this study was cross-sectional and relied on retrospective self-report to ascertain exposures and outcomes. In contrast, our cohort study used a more rigorous end point assessment and followed a much larger population of women over many years.

d CAD similar to those reported here (1.85 and 1.58, respectively). However, this study was cross-sectional and relied on retrospective self-report to ascertain exposures and outcomes. In contrast, our cohort study used a more rigorous end point assessment and followed a much larger population of women over many years. Our study used administrative data where clinical information, such as cardiovascular risk factors, was unavailable. Women with GDM exhibit chronic insulin resistance (8), which is associated with a clustering of risk factors that are in the causal pathway to CVD. Therefore, women with GDM likely have very different risk factor profiles than those without GDM, and adjusting for these differences might obscure a clinically important association between GDM and CVD. In summary, women with GDM are at increased risk for CVD events compared with women without GDM, and much of this risk is attributable to the subsequent development of type 2 diabetes. As diabetes prevention interventions in women with a history of GDM have also been shown to slow progression of atherosclerosis (9), this study highlights the importance of diabetes prevention for this high-risk population. Supplementary Material Figure 1B B.R.S. and R.R are supported by the Canadian Institutes of Health Research and the Canadian Diabetes Association. G.L.B. is supported by the Canadian Institutes of Health Research, the Ontario Women's Health Council, and the Banting and Best Diabetes Centre at the University of Toronto. We thank Ellen Chan and Ping Li for assistance with data acquisition.

The Archimedes model is a large-scale simulation model of human physiology and health care systems (1). It has been extensively validated by its ability to quite closely replicate a wide variety of aggregate health outcomes in populations (1). The ability of Archimedes to make accurate predictions for individuals, however, has thus far not been validated. Using data from the San Antonio Heart Study (SAHS), we attempted such a validation. We also compared the area under the receiver operating characteristic curves (aROCs) derived from Archimedes with those derived from two other diabetes predicting models, namely, the SAHS predicting model (2) and the Atherosclerosis Risk in Communities (ARIC) predicting model (3). RESEARCH DESIGN AND METHODS The SAHS is a prospective cohort study consisting of 3,682 individuals (62% Mexican American and 38% non-Hispanic white) followed for 7–8 years (4). The SAHS predicting model is a multiple logistic regression model with incident diabetes as the dependent variable and a panel of baseline characteristics that are ordinarily available in a routine clinical setting as independent variables (2). The ARIC predicting model is a similarly constructed logistic regression model (3).

predicting model is a multiple logistic regression model with incident diabetes as the dependent variable and a panel of baseline characteristics that are ordinarily available in a routine clinical setting as independent variables (2). The ARIC predicting model is a similarly constructed logistic regression model (3). The Archimedes model is built from underlying anatomy and physiology and uses scores of ordinary and differential equations to represent metabolic pathways, occurrence and progression of diseases, signs and symptoms, treatments, and outcomes. A practical, free, readily available tool derived from the Archimedes model is the American Diabetes Association's Diabetes PHD (Personal Health Decisions; available at http://diabetes.org/diabetesPHD). Diabetes PHD can simultaneously predict the risk of diabetes and numerous other outcomes, including the effects of a wide variety of treatments in many different populations (e.g., those with diabetes). It was used here to provide external validation of its prediction of the incidence of diabetes.

tes.org/diabetesPHD). Diabetes PHD can simultaneously predict the risk of diabetes and numerous other outcomes, including the effects of a wide variety of treatments in many different populations (e.g., those with diabetes). It was used here to provide external validation of its prediction of the incidence of diabetes. Among the 3,228 individuals in the SAHS who were nondiabetic at baseline, 295 developed diabetes over the 7–8 years of follow-up. All the required elements for the Archimedes risk estimation were available in the subjects selected for the present analyses. The present analyses were restricted to the recent cohort 2 of SAHS, which included 1,734 nondiabetic individuals, 195 of whom were diabetic at follow-up. Within the SAHS database, we selected 100 individuals at random, 50 of whom were diabetic at follow-up and 50 who remained free of diabetes at follow-up. This sample size would provide 80% power to detect an aROC significantly (P < 0.05) greater than 0.70 (the low end of acceptable discrimination [5]) if the true aROC was >0.80 and 90% power if the true aROC was 0.83 (benchmark values near that of other established models) (2,3).

remained free of diabetes at follow-up. This sample size would provide 80% power to detect an aROC significantly (P < 0.05) greater than 0.70 (the low end of acceptable discrimination [5]) if the true aROC was >0.80 and 90% power if the true aROC was 0.83 (benchmark values near that of other established models) (2,3). The risk of developing diabetes for each individual was determined according to the years of follow-up for that individual (rounded to the nearest year), which ranged from 6–9 with a mean of 7.5. Data from each individual were entered into Diabetes PHD and the results obtained from the graphical output displayed on the computer screen. A second person confirmed the accuracy of the input and, in a random sample of 20 forms, also confirmed the output from Diabetes PHD. We also estimated the risk of diabetes for the same 100 individuals using both the SAHS diabetes predicting model and the ARIC predicting model. The aROC's and CIs for all three models were computed and compared (6). Finally, we computed the Spearman correlation coefficients between the risk estimates obtained from each pair of predicting models. RESULTS The aROC for Diabetes PHD was 0.818 (95% CI 0.739–0.899) and was not statistically different than the aROC of the SAHS model (0.869 [95% CI 0.801- 0.936]) or the ARIC model (0.870 [0.802–0.937]) (Fig. 1). The risk estimates from the SAHS model and ARIC model were highly correlated (r = 0.962), and both correlated well with Diabetes PHD (r = 0.834 and 0.842, respectively).

9) and was not statistically different than the aROC of the SAHS model (0.869 [95% CI 0.801- 0.936]) or the ARIC model (0.870 [0.802–0.937]) (Fig. 1). The risk estimates from the SAHS model and ARIC model were highly correlated (r = 0.962), and both correlated well with Diabetes PHD (r = 0.834 and 0.842, respectively). CONCLUSIONS With an aROC of 0.818, it is evident that the accuracy of Diabetes PHD (i.e., Archimedes) to predict an individual's risk of diabetes is excellent—almost as high as models specifically designed and used only for that purpose. The SAHS model may have had an unfair advantage over Archimedes because it was designed and optimized using the SAHS database and could be overfitted to the subset of SAHS cases selected for this analysis. It was for that reason that we used the ARIC predicting model: the latter was developed in an entirely independent dataset and performed as well as the SAHS model. Both the SAHS and ARIC models were built from person-specific data and optimized specifically for predicting incident diabetes. In contrast, Archimedes was designed to be used for a very wide range of purposes, calculates many different outcomes, was not built from person-specific data, and was not calibrated to determine the incidence of diabetes. Also, several of the variables Archimedes uses that may have enhanced its predictive capability were not included in this analysis.

ed to be used for a very wide range of purposes, calculates many different outcomes, was not built from person-specific data, and was not calibrated to determine the incidence of diabetes. Also, several of the variables Archimedes uses that may have enhanced its predictive capability were not included in this analysis. This report extends the validation of Archimedes and demonstrates its excellent ability to discriminate between individuals who will or will not develop diabetes. Its utility is comparable with models developed solely for that purpose. Because Diabetes PHD, derived from Archimedes, is freely available on the internet and calculates many additional outcomes, it is a powerful tool that can be reliably used for comprehensive risk assessment and decision making. Diabetes PHD is now widely accessed (∼80,000 users per year) for use in comprehensive risk assessment of cardiometabolic disease over a 30-year period and that helps diabetic patients better appreciate the likely benefits of risk factor reduction. Although the tool currently uses complex distributive computing, which limits its speed and capacity, a much more rapid version will soon become available with unlimited capacity. This will allow for widespread promotion. This study was supported by grants from the National Heart, Lung, and Blood Institute (R01 HL24799 and R01 HL36820).

Type 2 diabetes is associated with increased risk of all-cause mortality and cardiovascular disease (CVD). However, clinical trials to date have not demonstrated that achieving normal glucose levels can reduce the risk for cardiovascular events. In the UK Prospective Diabetes Study (UKPDS), intensive blood glucose reduction was achieved using metformin therapy in diet-treated overweight patients, resulting in a decreased risk of myocardial infarction and all-cause mortality. However, when a combination of metformin and sulfonylurea was prescribed in the same trial for glycemic control, there was a significant increased risk of diabetes-related death and all-cause mortality rather than a beneficial effect, a finding attributed by the investigators to be due to chance (1). In the UKPDS, sulfonylureas themselves were not associated with the risk of diabetes-related death or myocardial infarction (2), but in previous studies such as the University Group Diabetes Program (UGDP) some increased risk was seen (3), and a warning about increased risk of CVD is included in the Federal Drug Administration–approved label for this class of drugs.

associated with the risk of diabetes-related death or myocardial infarction (2), but in previous studies such as the University Group Diabetes Program (UGDP) some increased risk was seen (3), and a warning about increased risk of CVD is included in the Federal Drug Administration–approved label for this class of drugs. A recent systematic review of clinical trials of diabetes therapies noted that data on long-term outcomes were not available in most clinical trials (4). Observational studies investigating the association between combination therapy of metformin and sulfonylureas and risk of CVD and mortality have reported conflicting results. Some studies have reported that the use of this combination therapy increases the risk of all-cause and CVD mortality (5), while others have reported no association (6,7) or a decreased risk of mortality from all causes and CVD (8). Since these are likely the most commonly prescribed medications for type 2 diabetes, the possible increase in risk of all-cause mortality and cardiovascular events is troubling (1). Given these inconsistencies in the literature and the lack of clinical trials assessing the long-term effects of combination therapy of sulfonylureas and metformin, we conducted a meta-analysis of observational studies to examine the association between combination therapy of sulfonylureas and metformin and risk of CVD and all-cause mortality.

es in the literature and the lack of clinical trials assessing the long-term effects of combination therapy of sulfonylureas and metformin, we conducted a meta-analysis of observational studies to examine the association between combination therapy of sulfonylureas and metformin and risk of CVD and all-cause mortality. RESEARCH DESIGN AND METHODS A literature search of the MEDLINE database (from January 1966 through July 2007) was conducted using the medical subject headings “diabetes mellitus, type 2;” “drug therapy, combination;” “drug combinations;” “sulfonylurea compounds;” “acetohexamide;” “chlorpropamide;” “tolbutamide;” “tolazamide;” “glyburide;” “glipizide;” “biguanides;” and “metformin” and keyword “glimepiride.” The search was restricted to include studies conducted only in human subjects. Studies were also identified through a search of references cited in the original published studies and relevant review articles.

“tolbutamide;” “tolazamide;” “glyburide;” “glipizide;” “biguanides;” and “metformin” and keyword “glimepiride.” The search was restricted to include studies conducted only in human subjects. Studies were also identified through a search of references cited in the original published studies and relevant review articles. The contents of 299 abstracts or full-text manuscripts identified during the literature search were reviewed independently by two investigators in duplicate to determine whether they met the criteria for inclusion. When there were discrepancies between investigators for inclusion or exclusion, a third investigator conducted additional evaluation of the study and the discrepancies were resolved in conference. The following inclusion criteria were used for study selection: 1) observational study that investigated the relationship between combination therapy with metformin (biguanides) plus sulfonylureas and risk of CVD and/or mortality, 2) adjusted relative risk (RR) or equivalent (i.e., hazard ratio, odds ratio) and corresponding variance or equivalent reported, and 3) diagnosis of type 2 diabetes established using the standard criteria for the time of the study.

therapy with metformin (biguanides) plus sulfonylureas and risk of CVD and/or mortality, 2) adjusted relative risk (RR) or equivalent (i.e., hazard ratio, odds ratio) and corresponding variance or equivalent reported, and 3) diagnosis of type 2 diabetes established using the standard criteria for the time of the study. All data were independently abstracted in duplicate. Differences in data extraction were resolved in conference and by referencing the original publication. No authors were contacted to request additional information. A standardized abstraction form was used to record the following information: study title, first author's name, year of publication, study country, study years, name of cohort, study design (prospective or retrospective cohort study or case-control study), duration of follow-up, characteristics of the study population (sample size, distribution of age, race, and sex, mean diabetes duration, mean A1C), type of reference group, and confounding factors controlled for. The RR of cardiovascular mortality/morbidity and/or all-cause or cause-specific mortality associated with combination therapy and their corresponding CIs or SEs were abstracted. The number of events for all-cause mortality and cardiovascular mortality/morbidity were abstracted.

, and confounding factors controlled for. The RR of cardiovascular mortality/morbidity and/or all-cause or cause-specific mortality associated with combination therapy and their corresponding CIs or SEs were abstracted. The number of events for all-cause mortality and cardiovascular mortality/morbidity were abstracted. Statistical analysis RRs were used as the measure of association between combination therapy of metformin and sulfonylurea and CVD and all-cause mortality. The RRs of each study were weighted by the inverse of their variance. To stabilize the variances and to normalize the distributions, the RRs and corresponding SEs from each of the individual studies were transformed to their natural logarithms. When necessary, SEs were derived from the CIs provided in each original study. The primary data for time to event analyses were not available for the combined cohort. Therefore, for the overall analysis, RR estimates and 95% CIs for all-cause mortality and CVD associated with combination therapy were pooled irrespective of the reference group used. Subgroup analyses were conducted by reference group (diet, sulfonylurea monotherapy, or metformin monotherapy). Both fixed-effects and DerSimonian and Laird random-effects models were used to calculate the pooled RR of CVD and all-cause mortality associated with combination therapy (9). Although both models yielded similar findings, results from the random-effects model are presented herein owing to significant heterogeneity among the studies.

cts and DerSimonian and Laird random-effects models were used to calculate the pooled RR of CVD and all-cause mortality associated with combination therapy (9). Although both models yielded similar findings, results from the random-effects model are presented herein owing to significant heterogeneity among the studies. CVD was defined by each of the individual studies. We used cardiovascular mortality and all-cause mortality, as well as a composite end point of CVD hospitalizations (the first cardiovascular event either fatal or nonfatal event), or mortality as our study outcomes. One study reported RRs separately for coronary heart disease and stroke (10). For this study, we first weighted both of the RRs by the inverse of their variance and then pooled the RRs by using a fixed-effects model to obtain an overall estimate for the study. Begg's rank correlation test was used to examine the association between effect estimates and their variances, and Egger's linear regression test, which regresses Z statistics on the reciprocal of the SE for each study, was used to detect publication bias (11,12). Additionally, each study was omitted one at a time to evaluate the influence of that study on the pooled estimate. All analyses were performed using STATA version 8.2 (STATA, College Station, TX).

test, which regresses Z statistics on the reciprocal of the SE for each study, was used to detect publication bias (11,12). Additionally, each study was omitted one at a time to evaluate the influence of that study on the pooled estimate. All analyses were performed using STATA version 8.2 (STATA, College Station, TX). RESULTS Online appendix Figure A1 (available at http://dx.doi.org/10.2337/dc08-0167) depicts the flow of studies in the meta-analysis. Among 25 studies that met the inclusion criteria, 16 were excluded from the meta-analysis. Eleven studies did not report CVD or mortality as an outcome, three studies were duplicated, and two involved multiple drug combinations. Two studies examined the association between combination therapy of metformin and sulfonylurea in different groups of individuals according to which drug was given first, and these groups were treated as separate studies in the meta-analysis.

, three studies were duplicated, and two involved multiple drug combinations. Two studies examined the association between combination therapy of metformin and sulfonylurea in different groups of individuals according to which drug was given first, and these groups were treated as separate studies in the meta-analysis. The characteristics of the study participants and the design of the nine observational studies included in the meta-analysis are presented in Table 1 (5–8,10,13–16). Six of the studies were retrospective cohort studies, two were prospective cohort studies, and one was a nested case-control study. Of the nine studies, one was conducted in the U.S., two in Canada, one in Israel, and five in European countries. The number of participants in these studies ranged from 910 in the study by Olsson et al. (10) to 39,721 in the study by Kahler et al. (7). Mean age ranged from 58.9 to 71.3 years. The mean follow-up time ranged from 2.1 to 7.7 years. Among the nine studies, seven reported all-cause mortality, four reported cardiovascular mortality, and three reported cardiovascular hospitalizations. Of the 101,733 participants included in these studies, 25,091 participants received a combination therapy of metformin and sulfonylurea. Bruno et al. (13) and Koro et al. (16) did not specify the number of participants receiving combination therapy.

ascular mortality, and three reported cardiovascular hospitalizations. Of the 101,733 participants included in these studies, 25,091 participants received a combination therapy of metformin and sulfonylurea. Bruno et al. (13) and Koro et al. (16) did not specify the number of participants receiving combination therapy. Figure 1 depicts the results from the random-effects models pooling the adjusted RRs for all-cause mortality, CVD mortality, and CVD hospitalizations or mortality, respectively, associated with combination therapy of metformin and sulfonylurea. In addition, it shows the number of events associated with combination therapy in comparison with the control group for all-cause mortality, CVD mortality, and CVD hospitalizations or mortality. Pooled RR estimates were not statistically significant for all-cause mortality or CVD mortality, while the use of combination therapy was significantly associated with an increased risk of cardiovascular hospitalizations or mortality.

group for all-cause mortality, CVD mortality, and CVD hospitalizations or mortality. Pooled RR estimates were not statistically significant for all-cause mortality or CVD mortality, while the use of combination therapy was significantly associated with an increased risk of cardiovascular hospitalizations or mortality. In sensitivity analyses, significant heterogeneity was present for studies reporting all-cause mortality (P < 0.001). However, exclusion of any study did not change the pooled estimate. For studies reporting CVD mortality, significant heterogeneity was present (P < 0.001), and exclusion of the study by Johnson et al. (15) led to a significant increased risk of CVD mortality associated with combination therapy of metformin and sulfonylureas (RR 1.63 [95% CI 1.11–2.39]). Significant heterogeneity was also present for studies that reported cardiovascular hospitalizations or mortality (P = 0.001), and the exclusion of any study did not alter the pooled estimate. There was no evidence of publication bias by rank correlation or regression testing (P > 0.10 for all). In the study by Evans et al. (5), participants of the reference group were used more than once in computing the pooled estimate. Analyses were repeated omitting various combinations of this study, and no substantive changes in results were noted. Furthermore, we conducted a sensitivity analysis in which those studies that did not adjust for duration of diabetes or previous CVD were excluded (6,8,13,14,17). This information is included in Table 2.

nalyses were repeated omitting various combinations of this study, and no substantive changes in results were noted. Furthermore, we conducted a sensitivity analysis in which those studies that did not adjust for duration of diabetes or previous CVD were excluded (6,8,13,14,17). This information is included in Table 2. Subgroup analysis RR estimates of all-cause mortality, CVD mortality, and CVD hospitalizations or mortality associated with combination therapy of metformin and sulfonylurea for subgroups defined according to the comparator treatment are presented in online appendix Table A1. The estimated RRs were >1.0 in all subgroups except for the association between all-cause mortality and combination therapy compared with sulfonylurea. Compared with diet therapy, combination therapy significantly increased the RR of all-cause mortality, and combination therapy compared with metformin monotherapy significantly increased the RR of CVD hospitalizations or mortality.

Subgroup analysis RR estimates of all-cause mortality, CVD mortality, and CVD hospitalizations or mortality associated with combination therapy of metformin and sulfonylurea for subgroups defined according to the comparator treatment are presented in online appendix Table A1. The estimated RRs were >1.0 in all subgroups except for the association between all-cause mortality and combination therapy compared with sulfonylurea. Compared with diet therapy, combination therapy significantly increased the RR of all-cause mortality, and combination therapy compared with metformin monotherapy significantly increased the RR of CVD hospitalizations or mortality. CONCLUSIONS In the current meta-analysis, combination therapy of metformin and sulfonylurea significantly increased the RR of cardiovascular hospitalization or mortality (fatal and nonfatal events) irrespective of the reference group (diet therapy, metformin monotherapy, or sulfonylurea monotherapy) used. However, there were no statistically significant effects of combination therapy of sulfonylurea and metformin on CVD mortality or all-cause mortality. These results may help clarify the conflicting findings of several large observational studies that examined the effect of combination therapy with metformin and sulfonylureas on the risk of CVD events among patients with type 2 diabetes, while the association of this combination with all-cause and cardiovascular mortality remains obscure.

p clarify the conflicting findings of several large observational studies that examined the effect of combination therapy with metformin and sulfonylureas on the risk of CVD events among patients with type 2 diabetes, while the association of this combination with all-cause and cardiovascular mortality remains obscure. Due to the progressive nature of type 2 diabetes, many patients are put on combinations of oral antihyperglycemic agents in order to meet glycemic goals. For instance, in the recommended algorithm, the combination of sulfonylurea and metformin is the second step in the management of patients with type 2 diabetes (18). It is likely that patients on combination therapy are likely to have either a more rapidly progressive form of the disease or a longer duration of diabetes, perhaps both. The reduction of blood glucose in high-risk obese patients with type 2 diabetes on metformin therapy alone in the UKPDS was associated with a decrease in adverse cardiovascular events (2). However, when a combination of metformin and sulfonylurea was prescribed, there was an increased risk, which is in contrast with some of the observational studies. This discrepancy may be due to differences in the population between these studies.

PDS was associated with a decrease in adverse cardiovascular events (2). However, when a combination of metformin and sulfonylurea was prescribed, there was an increased risk, which is in contrast with some of the observational studies. This discrepancy may be due to differences in the population between these studies. It may not only be important to reduce blood glucose, but also to consider the choice of agent used to make such a reduction. A recent meta-analysis has created much controversy about some of the newer medications used to reduce blood glucose by suggesting that rosiglitazone may be associated with an increased risk of myocardial infarction and possibly death (19). It is noteworthy that much of this increased risk with rosiglitazone was seen in combination therapies (20). However, the interim analysis of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial has shown inconclusive results (21). Our meta-analysis is important in the context of that study, as the combination of metformin and sulfonylurea is the comparator group to the rosiglitazone combinations.

nalysis of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial has shown inconclusive results (21). Our meta-analysis is important in the context of that study, as the combination of metformin and sulfonylurea is the comparator group to the rosiglitazone combinations. Several observational studies have examined the association between combination therapy and risk of CVD and all-cause mortality. Evans et al. (5) carried out an analysis of a database of 400,000 people in Scotland and identified 5,730 patients who were prescribed oral hypoglycemia agents between 1994 and 2001. Patients treated with sulfonylureas alone or in combination with metformin appeared to have an increased RR of adverse cardiovascular outcomes compared with those treated with metformin alone. It was particularly disturbing to note that the combination of sulfonylurea with metformin seemed to abrogate the potential benefit of metformin on CVD outcome, as seen in the UKPDS (2). A study by Fisman et al. (14) was carried out among 2,275 patients with type 2 diabetes and coronary artery disease, as part of the Bezafibrate Infarction Prevention Study. The patients were followed for over 7 years, and the authors demonstrated that cardiovascular events and mortality were the same whether glyburide, a sulfonylurea, or metformin was used for treatment. However, there was a significant time-related increased mortality when the combination therapy was used. Olsson et al. (10) analyzed mortality in a small cohort of patients taking sulfonylureas alone or in combination with metformin and demonstrated a higher cardiovascular mortality in patients taking the combination than those taking sulfonylurea alone.

e-related increased mortality when the combination therapy was used. Olsson et al. (10) analyzed mortality in a small cohort of patients taking sulfonylureas alone or in combination with metformin and demonstrated a higher cardiovascular mortality in patients taking the combination than those taking sulfonylurea alone. In our meta-analysis, exclusion of the study by Johnson et al. (15) led to a significant increased risk of CVD mortality associated with combination therapy of metformin and sulfonylurea. The study by Johnson et al. (15) reported a reduced risk of CVD mortality associated with combination therapy of metformin and sulfonylurea when compared with sulfonylurea monotherapy, but the study had many limitations. A large number of patients were excluded because of short-term insulin use. Patients prescribed the combination therapy were 2.3 years younger than those prescribed metformin monotherapy and 5.8 years younger than those prescribed sulfonylurea monotherapy, a discrepancy that is difficult to explain. Patients with more severe disease or intercurrent illnesses including hospitalization for cardiovascular events may have required insulin use and were therefore excluded from the study.

etformin monotherapy and 5.8 years younger than those prescribed sulfonylurea monotherapy, a discrepancy that is difficult to explain. Patients with more severe disease or intercurrent illnesses including hospitalization for cardiovascular events may have required insulin use and were therefore excluded from the study. In our analysis, we found a relatively greater association with fatal and nonfatal CVD events than in fatal events alone, suggesting that the incidence of CVD events may be increased with combination therapy, but there may have been a lower case-fatality rate. This contrasts with the recent data from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study (22) in which intensive treatment with multiple combinations of diabetes therapies was associated with decreased nonfatal CVD events but increased fatal events. It is impossible to determine the reason for this discrepancy, although it is possible that patients in the observational studies included in our analysis did not have a level of glycemia as low as that attempted in the ACCORD trial.

herapies was associated with decreased nonfatal CVD events but increased fatal events. It is impossible to determine the reason for this discrepancy, although it is possible that patients in the observational studies included in our analysis did not have a level of glycemia as low as that attempted in the ACCORD trial. Several hypothetical considerations may explain the increased risk associated with such a combination. First, it is possible that patients needing such a combination have a more aggressive form of the disease and therefore more rapid deterioration in glycemic control over time. Second, sulfonylureas are associated with weight gain, whereas metformin is associated with weight loss, as well as some improvement in a variety of cardiovascular risk factors. Any weight gain induced by the combination may negate some of these beneficial effects and increase risks. Other possible explanations include the known propensity of sulfonylureas to cause hypoglycemia. When used in combination with a drug like metformin, which may decrease hepatic glucose production, recovery from hypoglycemia may be impaired. Hypoglycemia may increase the risk of cardiovascular abnormalities, including ischemia and a propensity to cause arrhythmias (23,24). There is also considerable controversy about the impact of sulfonylureas on ischemic preconditioning (25), but nothing is known about the effects of combination therapy.

may be impaired. Hypoglycemia may increase the risk of cardiovascular abnormalities, including ischemia and a propensity to cause arrhythmias (23,24). There is also considerable controversy about the impact of sulfonylureas on ischemic preconditioning (25), but nothing is known about the effects of combination therapy. Although a meta-analysis is not the best way to test the efficacy and safety of such a combination of treatments, it is highly unlikely that a large-scale clinical trial to test this hypothesis will be carried out. Thus, we must rely on data from observational studies to arrive at conclusions and make appropriate recommendations. It is also unclear to what extent certain biases and methodological limitations, such as residual confounding, might exist in the studies included in this meta-analysis, since the majority of these studies were retrospective database analyses. In addition, the reference group varied among the studies. For instance, some studies used diet as the reference group, while others used sulfonylureas or metformin monotherapy as the reference group. Finally, we observed substantial quantitative heterogeneity across the studies, but the small number of studies limited our ability to explore possible sources of this variability. Additionally, findings from the subgroup analyses should be interpreted cautiously, as the number of studies examined was small.

nce group. Finally, we observed substantial quantitative heterogeneity across the studies, but the small number of studies limited our ability to explore possible sources of this variability. Additionally, findings from the subgroup analyses should be interpreted cautiously, as the number of studies examined was small. Overall, our results provide a mix of reassurance and concern to prescribers of diabetes medications who use combination therapies to achieve good glycemic control. Since sulfonylurea and metformin are likely the most widely used combination, it is possible that such use leads to early improvement in glycemic control, which, in itself, may lead to better microvascular outcomes. Although diet alone is associated with lower mortality risk, in the UKPDS, diet alone was associated with increased microvascular complications (2). Therefore, one must balance the risks and benefits of medications used while making treatment decisions. We emphasize that this meta-analysis has limitations and serves to examine published data to generate hypotheses. Such analysis should not be used as a basis for clinical decisions. We hope that our analysis will prompt the planning of future clinical trials to determine not only the value of good glycemic control, but also the safest and most cost effective way to achieve glycemic goals. Clearly, we need further studies to assess the association of combination therapy of metformin and sulfonylurea with all-cause and/or cardiovascular mortality as well as to understand the potential mechanism of its deleterious effects.

control, but also the safest and most cost effective way to achieve glycemic goals. Clearly, we need further studies to assess the association of combination therapy of metformin and sulfonylurea with all-cause and/or cardiovascular mortality as well as to understand the potential mechanism of its deleterious effects. This study was not funded. K.R. was partially supported by grant P20-RR17659 from the National Center for Research Resources (National Institutes of Health [NIH]). Diabetes research and education at Tulane University Health Sciences Center is supported in part by the Tullis-Tulane Alumni Chair in Diabetes and the Earl Madison Ellis fund. V.F. is supported in part by the American Diabetes Association (ADA) and the NIH (ACCORD and TINSAL T2D trials). V.F. has also received research support (to Tulane) from Glaxo Smith Kline, Novartis, Takeda, Astra Zeneca, Pfizer, sanofi-aventis, Eli Lilly, NIH, and ADA; and honoraria from Glaxo Smith Kline, Novartis, Takeda, Pfizer, sanofi-aventis, and Eli Lilly.

It is now 10 years since the last technical review on preventative foot care was published (1), which was followed by an American Diabetes Association (ADA) position statement on preventive foot care in diabetes (2). Many studies have been published proposing a range of tests that might usefully identify patients at risk of foot ulceration, creating confusion among practitioners as to which screening tests should be adopted in clinical practice. A task force was therefore assembled by the ADA to address and concisely summarize recent literature in this area and then recommend what should be included in the comprehensive foot exam for adult patients with diabetes. The committee was cochaired by the immediate past and current chairs of the ADA Foot Care Interest Group (A.J.M.B. and D.G.A.), with other panel members representing primary care, orthopedic and vascular surgery, physical therapy, podiatric medicine and surgery, and the American Association of Clinical Endocrinologists. THE PATHWAY TO FOOT ULCERATION The lifetime risk of a person with diabetes developing a foot ulcer may be as high as 25%, whereas the annual incidence of foot ulcers is ∼2% (3–7). Up to 50% of older patients with type 2 diabetes have one or more risk factors for foot ulceration (3,6). A number of component causes, most importantly peripheral neuropathy, interact to complete the causal pathway to foot ulceration (1,3–5). A list of the principal contributory factors that might result in foot ulcer development is provided in Table 1.

e 2 diabetes have one or more risk factors for foot ulceration (3,6). A number of component causes, most importantly peripheral neuropathy, interact to complete the causal pathway to foot ulceration (1,3–5). A list of the principal contributory factors that might result in foot ulcer development is provided in Table 1. The most common triad of causes that interact and ultimately result in ulceration has been identified as neuropathy, deformity, and trauma (5). As identification of those patients at risk of foot problems is the first step in preventing such complications, this report will focus on key components of the foot exam. COMPONENTS OF THE FOOT EXAM History While history is a pivotal component of risk assessment, a patient cannot be fully assessed for risk factors for foot ulceration based on history alone; a careful foot exam remains the key component of this process. Key components of the history include previous foot ulceration or amputation. Other important assessments in the history (Table 2) include neuropathic or peripheral vascular symptoms (7,8), impaired vision, or renal replacement therapy. Lastly, tobacco use should be recorded, since cigarette smoking is a risk factor not only for vascular disease but also for neuropathy.

ulceration or amputation. Other important assessments in the history (Table 2) include neuropathic or peripheral vascular symptoms (7,8), impaired vision, or renal replacement therapy. Lastly, tobacco use should be recorded, since cigarette smoking is a risk factor not only for vascular disease but also for neuropathy. General inspection A careful inspection of the feet in a well-lit room should always be carried out after the patient has removed shoes and socks. Because inappropriate footwear and foot deformities are common contributory factors in the development of foot ulceration (1,5), the shoes should be inspected and the question “Are these shoes appropriate for these feet?” should be asked. Examples of inappropriate shoes include those that are excessively worn or are too small for the person's feet (too narrow, too short, toe box too low), resulting in rubbing, erythema, blister, or callus. Features that should be assessed during foot inspection are outlined in Table 3 and are discussed below. Dermatological assessment. The dermatological assessment should initially include a global inspection, including interdigitally, for the presence of ulceration or areas of abnormal erythema. The presence of callus (particularly with hemorrhage), nail dystrophy, or paronychia should be recorded (9), with any of these findings prompting referral to a specialist or specialty clinic. Focal or global skin temperature differences between one foot and the other may be predictive of either vascular disease or ulceration and could also prompt referral for specialty foot care (10–13).

or paronychia should be recorded (9), with any of these findings prompting referral to a specialist or specialty clinic. Focal or global skin temperature differences between one foot and the other may be predictive of either vascular disease or ulceration and could also prompt referral for specialty foot care (10–13). Musculoskeletal assessment. The musculoskeletal assessment should include evaluation for any gross deformity (14). Rigid deformities are defined as any contractures that cannot easily be manually reduced and are most frequently found in the digits. Common forefoot deformities that are known to increase plantar pressures and are associated with skin breakdown include metatarsal phalangeal joint hyperextension with interphalangeal flexion (claw toe) or distal phalangeal extension (hammer toe) (15–17). (Examples of these deformities are shown in Fig. 1.) An important and often overlooked or misdiagnosed condition is Charcot arthropathy. This occurs in the neuropathic foot and most often affects the midfoot. This may present as a unilateral red, hot, swollen, flat foot with profound deformity (18–20). A patient with suspected Charcot arthropathy should be immediately referred to a specialist for further assessment and care.

tion is Charcot arthropathy. This occurs in the neuropathic foot and most often affects the midfoot. This may present as a unilateral red, hot, swollen, flat foot with profound deformity (18–20). A patient with suspected Charcot arthropathy should be immediately referred to a specialist for further assessment and care. Neurological assessment Peripheral neuropathy is the most common component cause in the pathway to diabetic foot ulceration (1,4,5,7). The clinical exam recommended, however, is designed to identify loss of protective sensation (LOPS) rather than early neuropathy. The diagnosis and management of the latter were covered in a 2004 ADA technical review (7). The clinical examination to identify LOPS is simple and requires no expensive equipment. Five simple clinical tests (Table 3), each with evidence from well-conducted prospective clinical cohort studies, are considered useful in the diagnosis of LOPS in the diabetic foot (1–7). The task force agrees that any of the five tests listed could be used by clinicians to identify LOPS, although ideally two of these should be regularly performed during the screening exam—normally the 10-g monofilament and one other test. One or more abnormal tests would suggest LOPS, while at least two normal tests (and no abnormal test) would rule out LOPS. The last test listed, vibration assessment using a biothesiometer or similar instrument, is widely used in the U.S.; however, identification of the patient with LOPS can easily be carried out without this or other expensive equipment.

est LOPS, while at least two normal tests (and no abnormal test) would rule out LOPS. The last test listed, vibration assessment using a biothesiometer or similar instrument, is widely used in the U.S.; however, identification of the patient with LOPS can easily be carried out without this or other expensive equipment. 10-g monofilaments. Monofilaments, sometimes known as Semmes-Weinstein monofilaments, were originally used to diagnose sensory loss in leprosy (21). Many prospective studies have confirmed that loss of pressure sensation using the 10-g monofilament is highly predictive of subsequent ulceration (3,21,22). Screening for sensory loss with the 10-g monofilament is in widespread use across the world, and its efficacy in this regard has been confirmed in a number of trials, including the recent Seattle Diabetic Foot Study (4,21,23,24). Nylon monofilaments are constructed to buckle when a 10-g force is applied; loss of the ability to detect this pressure at one or more anatomic sites on the plantar surface of the foot has been associated with loss of large-fiber nerve function. It is recommended that four sites (1st, 3rd, and 5th metatarsal heads and plantar surface of distal hallux) be tested on each foot.

force is applied; loss of the ability to detect this pressure at one or more anatomic sites on the plantar surface of the foot has been associated with loss of large-fiber nerve function. It is recommended that four sites (1st, 3rd, and 5th metatarsal heads and plantar surface of distal hallux) be tested on each foot. The technique for testing pressure perception with the 10-g monofilament is illustrated in Fig. 2; patients should close their eyes while being tested. Caution is necessary when selecting the brand of monofilament to use, as many commercially available monofilaments have been shown to be inaccurate. Single-use disposable monofilaments or those shown to be accurate by the Booth and Young (23) study are recommended. The sensation of pressure using the buckling 10-g monofilament should first be demonstrated to the patient on a proximal site (e.g., upper arm). The sites of the foot may then be examined by asking the patient to respond “yes” or “no” when asked whether the monofilament is being applied to the particular site; the patient should recognize the perception of pressure as well as identify the correct site. Areas of callus should always be avoided when testing for pressure perception.

ot may then be examined by asking the patient to respond “yes” or “no” when asked whether the monofilament is being applied to the particular site; the patient should recognize the perception of pressure as well as identify the correct site. Areas of callus should always be avoided when testing for pressure perception. 128-Hz tuning forks. The tuning fork is widely used in clinical practice and provides an easy and inexpensive test of vibratory sensation. Vibratory sensation should be tested over the tip of the great toe bilaterally. An abnormal response can be defined as when the patient loses vibratory sensation and the examiner still perceives it while holding the fork on the tip of the toe (3,4). Pinprick sensation. Similarly, the inability of a subject to perceive pinprick sensation has been associated with an increased risk of ulceration (4). A disposable pin should be applied just proximal to the toenail on the dorsal surface of the hallux, with just enough pressure to deform the skin. Inability to perceive pinprick over either hallux would be regarded as an abnormal test result.

prick sensation has been associated with an increased risk of ulceration (4). A disposable pin should be applied just proximal to the toenail on the dorsal surface of the hallux, with just enough pressure to deform the skin. Inability to perceive pinprick over either hallux would be regarded as an abnormal test result. Ankle reflexes. Absence of ankle reflexes has also been associated with increased risk of foot ulceration (4). Ankle reflexes can be tested with the patient either kneeling or resting on a couch/table. The Achilles tendon should be stretched until the ankle is in a neutral position before striking it with the tendon hammer. If a response is initially absent, the patient can be asked to hook fingers together and pull, with the ankle reflexes then retested with reinforcement. Total absence of ankle reflex either at rest or upon reinforcement is regarded as an abnormal result.

in a neutral position before striking it with the tendon hammer. If a response is initially absent, the patient can be asked to hook fingers together and pull, with the ankle reflexes then retested with reinforcement. Total absence of ankle reflex either at rest or upon reinforcement is regarded as an abnormal result. Vibration perception threshold testing. The biothesiometer (or neurothesiometer) is a simple handheld device that gives semiquantitative assessment of vibration perception threshold (VPT). As for vibration using the 128-Hz tuning fork, vibration perception using the biothesiometer is also tested over the pulp of the hallux. With the patient lying supine, the stylus of the instrument is placed over the dorsal hallux and the amplitude is increased until the patient can detect the vibration; the resulting number is known as the VPT. This process should initially be demonstrated on a proximal site, and then the mean of three readings is taken over each hallux. A VPT >25 V is regarded as abnormal and has been shown to be strongly predictive of subsequent foot ulceration (15,22). Vascular assessment Peripheral arterial disease (PAD) is a component cause in approximately one-third of foot ulcers and is often a significant risk factor associated with recurrent wounds (5,25). Therefore, the assessment of PAD is important in defining overall lower-extremity risk status. Vascular examination should include palpation of the posterior tibial and dorsalis pedis pulses (10,26), which should be characterized as either “present” or “absent” (26).

risk factor associated with recurrent wounds (5,25). Therefore, the assessment of PAD is important in defining overall lower-extremity risk status. Vascular examination should include palpation of the posterior tibial and dorsalis pedis pulses (10,26), which should be characterized as either “present” or “absent” (26). Diabetic patients with signs or symptoms of vascular disease (Table 2) or absent pulses on screening foot examination should undergo ankle brachial pressure index (ABI) pressure testing and be considered for a possible referral to a vascular specialist. The ABI is a simple and easily reproducible method of diagnosing vascular insufficiency in the lower limbs. Blood pressure at the ankle (dorsalis pedis or posterior tibial arteries) is measured using a standard Doppler ultrasonic probe. This technique is outlined in Fig. 3. The ABI is obtained by dividing the ankle systolic pressure by the higher of the two brachial systolic pressures (8). An ABI >0.9 is normal, <0.8 is associated with claudication, and <0.4 is commonly associated with ischemic rest pain and tissue necrosis.

Doppler ultrasonic probe. This technique is outlined in Fig. 3. The ABI is obtained by dividing the ankle systolic pressure by the higher of the two brachial systolic pressures (8). An ABI >0.9 is normal, <0.8 is associated with claudication, and <0.4 is commonly associated with ischemic rest pain and tissue necrosis. The ADA Consensus Panel on PAD recommended measurement of ABI in diabetic patients over 50 years of age and consideration of ABI measurement in younger patients with multiple PAD risk factors, repeating normal tests every 5 years (8). ABI may therefore be part of the annual comprehensive foot exam in these patient subgroups. ABI measurements may be misleading in diabetes because the presence of medial calcinosis renders the arteries incompressible and results in falsely elevated or supra-systolic ankle pressures. In the presence of incompressible calf or ankle arteries (ABI >1.3), measurements of digital arterial systolic pressure (toe pressure) or transcutaneous oxygen tension may be performed.

e presence of medial calcinosis renders the arteries incompressible and results in falsely elevated or supra-systolic ankle pressures. In the presence of incompressible calf or ankle arteries (ABI >1.3), measurements of digital arterial systolic pressure (toe pressure) or transcutaneous oxygen tension may be performed. Risk classification and referral/follow-up Once the patient has been thoroughly assessed as described above, he or she should be assigned to a foot risk category (Table 4). These categories are designed to direct referral and subsequent therapy by the specialty clinician or team (17,20) and frequency of follow-up by the generalist or specialist. Increased category is associated with an increased risk for ulceration, hospitalization, and amputation (17). Patients in risk category 0 generally do not need referral and should receive general foot care education and undergo comprehensive foot examination annually. Patients in foot risk category 1 may be managed by a generalist or specialist every 3–6 months. Consideration should be given to an initial specialist referral to assess the need for specialized treatment and follow-up. Those in categories 2 and 3 should be referred to a foot care specialist or specialty clinic and seen every 1–3 months.

sk category 1 may be managed by a generalist or specialist every 3–6 months. Consideration should be given to an initial specialist referral to assess the need for specialized treatment and follow-up. Those in categories 2 and 3 should be referred to a foot care specialist or specialty clinic and seen every 1–3 months. CONCLUSIONS It cannot be overstated that the complications of the diabetic foot are common, complex, and costly, mandating aggressive and proactive preventative assessments by generalists and specialists. All patients with diabetes must have their feet evaluated at least at yearly intervals for the presence of the predisposing factors for ulceration and amputation (neuropathy, vascular disease, and deformities). This report summarizes a simple protocol for doing so. If abnormalities are present, more frequent evaluation of the diabetic foot is recommended depending on risk category, as described above and in Table 4. It is through systematic examination and risk assessment, patient education, and timely referral that we may further reduce the unnecessarily high prevalence of lower-extremity morbidity in this population. The meeting of the Task Force was supported by an unrestricted educational grant from KCI, San Antonio, TX.

CONCLUSIONS It cannot be overstated that the complications of the diabetic foot are common, complex, and costly, mandating aggressive and proactive preventative assessments by generalists and specialists. All patients with diabetes must have their feet evaluated at least at yearly intervals for the presence of the predisposing factors for ulceration and amputation (neuropathy, vascular disease, and deformities). This report summarizes a simple protocol for doing so. If abnormalities are present, more frequent evaluation of the diabetic foot is recommended depending on risk category, as described above and in Table 4. It is through systematic examination and risk assessment, patient education, and timely referral that we may further reduce the unnecessarily high prevalence of lower-extremity morbidity in this population. The meeting of the Task Force was supported by an unrestricted educational grant from KCI, San Antonio, TX. A.J.M.B. has received honoraria/consulting fees from Pfizer and Eli Lilly. R.G.F. has served on the speakers’ bureaus of KCI, Oculus, Pfizer, and Organogenesis and has received research support from Regenesis Biomedical and Derma Sciences. L.A.L. is a stockholder and on the board of directors of Diabetica Solutions and Pathways Disease Management; a stockholder of XL Health; on the scientific advisory board and speakers’ bureau of and has received research support from KCI; and a stockholder and on the scientific advisory boards of Cytomedics and Pegasus. P.S. is on the scientific advisory boards of Advanced Biohealing and Greystone; a consultant for Calretex, Cardiun, Heal Or, Taisho, and Hypermed; a speaker for Fox Hollow, Bristol- Meyers Squibb, sanofi-aventis, Merck, and Organogenesis; and has received research grants from Tissue Repair Company, Baxter, and PamLab. D.K.W. has received honoraria from Small Bones Innovation, Diabetic Global Foot Conference, and New Horizons in Cardiovascular Medicine.

Three chronic diseases—cancer, cardiovascular disease (CVD), and diabetes—are responsible for a majority of the morbidity, mortality, and health care costs in the U.S (1–8). To help reduce the toll of these diseases, the American Cancer Society, American Diabetes Association, and American Heart Association have recommended a variety of prevention activities (8). Each is supported by good evidence of effectiveness (8–16) and widely accepted. However, despite this support, there are large gaps in how well they are applied, and a high proportion of the U.S. population is not receiving prevention activities from which they would benefit (17–21).

riety of prevention activities (8). Each is supported by good evidence of effectiveness (8–16) and widely accepted. However, despite this support, there are large gaps in how well they are applied, and a high proportion of the U.S. population is not receiving prevention activities from which they would benefit (17–21). To stimulate greater attention to prevention and to help physicians and health care delivery organizations implement prevention activities, it is important to know the answers to several questions. First, how many people alive today are candidates for at least one prevention activity? Second, how much of the morbidity, mortality, and cost of these diseases is potentially preventable? Stated another way, by how much could the burden of chronic diseases be reduced if prevention activities were applied with 100% performance, compliance, and effectiveness? Third, what could realistically be accomplished if patients, physicians, and health plans throughout the country pursued prevention at levels of performance and compliance achieved by the most successful organizations? Fourth, how do the various prevention activities compare? Which are the most important in terms of their potential effects on health outcomes, costs, and cost-effectiveness? Fifth, what does prevention cost? If pursued at maximum feasibility levels, would the costs be offset by the savings? Finally, what are the main factors that determine the cost-effectiveness of a prevention activity, and what are the best ways to make prevention more attractive financially? This report offers answers to these questions for the prevention of CVD.

at maximum feasibility levels, would the costs be offset by the savings? Finally, what are the main factors that determine the cost-effectiveness of a prevention activity, and what are the best ways to make prevention more attractive financially? This report offers answers to these questions for the prevention of CVD. RESEARCH DESIGN AND METHODS Overview Ideally, the answers to the above questions would be obtained by examining the results of clinical trials. While there are studies that document that each of the prevention activities is effective, none of the existing studies addresses a representative sample of the U.S. population, addresses specific treatment goals that are being recommended, or includes representative U.S. costs. Furthermore, it is not possible to conduct the needed trials because of the large number of activities, long time horizons, large numbers of subjects required, and high cost of such research.

le of the U.S. population, addresses specific treatment goals that are being recommended, or includes representative U.S. costs. Furthermore, it is not possible to conduct the needed trials because of the large number of activities, long time horizons, large numbers of subjects required, and high cost of such research. Lacking clinical trials, the only alternative is to use a mathematical model. For this analysis, we selected the Archimedes model from other available mathematical models because of its ability to simulate the U.S. population at a person-specific level, its ability to simulate current patterns of care, its inclusion of all the relevant diseases and prevention activities in a single integrated model, its ability to analyze the prevention activities precisely as they are recommended, its ability to address all the questions of interest using a consistent methodology, and its demonstrated accuracy in reproducing the trials that document the effectiveness of each of the recommended interventions.

gle integrated model, its ability to analyze the prevention activities precisely as they are recommended, its ability to address all the questions of interest using a consistent methodology, and its demonstrated accuracy in reproducing the trials that document the effectiveness of each of the recommended interventions. Archimedes model The Archimedes model is a person-by-person, object-by-object, large-scale simulation model of physiology, disease, and health care systems written at a high level of detail using object-oriented programming and run on a distributed computing network (22–26). The core of the model is a set of ordinary and differential equations that represent the physiological pathways pertinent to diseases and their complications. Currently, the model includes coronary artery disease (CAD), stroke, diabetes and its complications, congestive heart failure, obesity, smoking, asthma, and the metabolic syndrome in a single integrated model. The model also includes aspects of diseases and health care systems needed to analyze downstream clinical events, utilization, and costs including signs and symptoms; patient encounters with the health care system (e.g., emergency room visits, office visits, and admissions); protocols and guidelines; tests and treatments; patient adherence to treatment recommendations; and clinical events that affect logistics, utilization, and financial costs.

nd costs including signs and symptoms; patient encounters with the health care system (e.g., emergency room visits, office visits, and admissions); protocols and guidelines; tests and treatments; patient adherence to treatment recommendations; and clinical events that affect logistics, utilization, and financial costs. Physiological variables that are continuous in reality are continuous in the model (e.g., blood pressure and glucose levels), time is continuous, symptoms are driven by underlying variables, tests measure underlying variables, treatments affect underlying variables, and outcomes are determined by the progression of the variables. Costs related to the conditions that are in the model are calculated by tracking all the pertinent cost-generating events using micro-costing methods (32). Costs of other conditions that are not currently calculated in the model, such as cancer or osteoporosis (“unrelated costs”) (32), are added separately as a function of variables that are in the model (e.g., age, sex, weight, and disease states). The model uses person-specific data from real populations (e.g., the National Health and Nutrition Education Survey [NHANES]) to create simulated populations that match the real populations, person by person. Each individual can be matched to variables such as demographics, risk factors, biological variables, current and past medical histories, and current treatments. The methods for creating the copies of real people preserve the distributions and correlations of all the important risk factors and biological variables.

on. Each individual can be matched to variables such as demographics, risk factors, biological variables, current and past medical histories, and current treatments. The methods for creating the copies of real people preserve the distributions and correlations of all the important risk factors and biological variables. The model's accuracy is checked by using it to simulate clinical trials that have been conducted in the real world and comparing the predicted results with the real results. This has been done successfully for several hundred treatments and outcomes in 48 randomized controlled trials thus far. Methods and results for the first 74 validation exercises involving 18 trials have been published (24). More than half of those (10 of 18 trials) were independent validations (33) in which no results in the trial were used to build or modify the model. More information about the Archimedes model, including additional details about the equations and sources, is available elsewhere (26). The current study For this study, we analyzed 11 prevention activities relating to CVD and combinations of these activities (Table 1). We conducted the analysis in three steps. First, we used person-specific data from the most current NHANES (1998–2004) to determine the characteristics (including sex and ethnicity), risk factors, and current levels of prevention in the U.S. population (34). We also used the NHANES data to create simulated populations that matched the real U.S. population.

ps. First, we used person-specific data from the most current NHANES (1998–2004) to determine the characteristics (including sex and ethnicity), risk factors, and current levels of prevention in the U.S. population (34). We also used the NHANES data to create simulated populations that matched the real U.S. population. Second, we created a care delivery setting that could serve as a representation of how health care is currently delivered in the U.S. We modified different aspects of the care setting through sensitivity analysis. For the representative setting, we based the use of prevention activities and degree of control of risk factors on the practices and success rates in the NHANES population. We based the treatment of symptoms and complications (e.g., management of diabetes and CVD) on national guidelines. We based the costs of drugs on information provided by drugstore.com and the cost of general medical care (e.g., emergency visits, office visits and admissions, and procedures) on costs experienced by Kaiser Permanente Southern California or from the literature (35). The costs of the prevention activities assumed for the reference case are given in Table 2. For the reference case, the costs of unrelated care and extra costs for the last year of life (beyond the costs related to the diseases calculated explicitly in the model) were set to zero. Different assumptions about the costs of prevention activities, general medical costs, and unrelated medical costs were all studied through sensitivity analysis.

ts of unrelated care and extra costs for the last year of life (beyond the costs related to the diseases calculated explicitly in the model) were set to zero. Different assumptions about the costs of prevention activities, general medical costs, and unrelated medical costs were all studied through sensitivity analysis. The third step was to use the simulated populations and simulated care delivery setting to conduct 13 simulated clinical trials. Eleven of the trials addressed prevention activities, one by one (Table 1). The other two trials addressed the combination of all 11 activities, given either with 100% performance and success in reaching the treatment targets or at more feasible levels. To the extent possible, the reference assumptions about feasible levels of performance (Table 1) were based on the levels of success that have been achieved in various clinical settings (36–42). Uncertainty about feasible performance levels was studied through sensitivity analysis. Additional simulated trials were conducted to study the sensitivity of the results to bundling of prevention services.

mance (Table 1) were based on the levels of success that have been achieved in various clinical settings (36–42). Uncertainty about feasible performance levels was studied through sensitivity analysis. Additional simulated trials were conducted to study the sensitivity of the results to bundling of prevention services. Analogous to the treatment arms of a clinical trial, the simulated population created for each trial was subjected to two management protocols. One management protocol represented “current care”: for each individual, we determined that person's current level of adoption of prevention (e.g., smoking habits, weight, and blood pressure) and assumed that the level of care responsible for that level of prevention would continue. Behaviors and physiological variables would be allowed to progress naturally as occurs with age but with no changes in any aspects of their care relevant to the prevention activities listed in Table 1. In these current care treatment arms, individuals were given additional treatments (beyond their current levels of prevention care) only if they developed symptoms, in which case the model assumed they sought care, or if clinical events such as heart attacks occurred.

ant to the prevention activities listed in Table 1. In these current care treatment arms, individuals were given additional treatments (beyond their current levels of prevention care) only if they developed symptoms, in which case the model assumed they sought care, or if clinical events such as heart attacks occurred. Each of the first 11 “one-by-one” trials also had a “prevention” arm in which people who were candidates for the applicable prevention activity were identified and treated to a level slightly (∼2–3%) below whatever target was specified for the applicable prevention activity. For these treatment arms, each individual in the simulated population was examined at the initiation of the trial and annually thereafter to determine whether he or she was a candidate for treatment according to whatever prevention activity was the subject of the trial. If a person met the criteria for the applicable prevention activity, then he or she would be treated to slightly below the corresponding target of that prevention activity. For example, if the trial was to estimate the effect of controlling A1C in people with currently uncontrolled diabetes, then each individual in the simulated population was given a simulated examination at the start of the trial to determine whether he or she had a diagnosis of diabetes and an A1C level >7%. If so, that person was treated to reduce their A1C level to 6.8%. Everyone with a condition was then reexamined at annual intervals to determine whether their A1C levels had increased to >7% and treated as needed to maintain A1C levels <7%. People who were not candidates for the applicable prevention activity at the start of the simulation were followed annually (screened) to see if they developed the condition in the interval following the previous examination, and if so, they were treated accordingly. The cost of screening (i.e., office visits and tests) was not considered.

andidates for the applicable prevention activity at the start of the simulation were followed annually (screened) to see if they developed the condition in the interval following the previous examination, and if so, they were treated accordingly. The cost of screening (i.e., office visits and tests) was not considered. For each of these simulated trials, we calculated the outcomes under two sets of assumptions about performance and compliance. In the first case, we analyzed the outcomes that would occur if 100% performance and compliance levels were achieved. This trial was done to estimate the maximum potential of prevention achievable by the recommended activities. In the second case, we applied more realistic, albeit aggressive, assumptions about what might constitute levels of performance that were feasible. In addition to the one-by-one trials, we created two simulated trials to estimate the overall proportion of U.S. adults who are candidates for any intervention and the overall effect of providing all of the prevention activities to anyone who was a candidate for them. In one of these trials, all people who were candidates for any of the prevention activities were treated with 100% performance and effectiveness. In the other, treatments were delivered at the more feasible levels of performance.

fect of providing all of the prevention activities to anyone who was a candidate for them. In one of these trials, all people who were candidates for any of the prevention activities were treated with 100% performance and effectiveness. In the other, treatments were delivered at the more feasible levels of performance. The sample size for each simulated trial was 50,000. The results were then scaled to the U.S. adult population, which in 2005 was ∼200 million individuals. Each trial was run for 30 years. All outcomes were calculated continuously and reported at annual intervals. For each trial, we calculated a wide range of health and economic outcomes. Here we report the total number of myocardial infarctions (MIs) (including repeat MIs), deaths from coronary heart disease (CHD), stroke, life-years, quality-adjusted life-years (QALYs), cost of prevention activities, cost of care other than the prevention activities, total medical costs, and cost per QALY. Quality-of-life weights for the various clinical states and outcomes were based on a survey by Sullivan and Ghushchyan (43) and varied in the sensitivity analysis. For calculating cost per QALY, both costs and QALYs were discounted 3%, with different discount rates studied through sensitivity analysis.

d cost per QALY. Quality-of-life weights for the various clinical states and outcomes were based on a survey by Sullivan and Ghushchyan (43) and varied in the sensitivity analysis. For calculating cost per QALY, both costs and QALYs were discounted 3%, with different discount rates studied through sensitivity analysis. RESULTS Of the 200 million people in the U.S. today between the ages of 20 and 80 years, ∼156 million (78%) meet the indications for at least one of the prevention activities listed in Table 1. Table 1 shows the numbers of people who are candidates for each particular activity; they vary widely from ∼3.2 million individuals with CAD and LDL cholesterol >100 mg/dl (1.8% of adults) to ∼60 million who have BMI >30 kg/m2 (30.1% of adults). Table 3 shows the outcomes that can be expected to occur in today's adults (independent of sex or ethnicity) over the next 30 years in the reference health care setting if the use of prevention activities continues at current levels (top row) and the differences in outcomes that could theoretically be achieved if prevention activities were adopted with 100% performance, compliance, and effectiveness. These entries show the maximum potential of prevention in reducing clinical outcomes of CVD. For example, if prevention continues at its current level, today's adults in the U.S. can expect to have ∼43 million MIs. If everyone adopted the prevention activities for which they are indicated, ∼27.4 million (63%) of those MIs could be prevented. Other columns show the effects on stroke, life-years, and QALYs.

CVD. For example, if prevention continues at its current level, today's adults in the U.S. can expect to have ∼43 million MIs. If everyone adopted the prevention activities for which they are indicated, ∼27.4 million (63%) of those MIs could be prevented. Other columns show the effects on stroke, life-years, and QALYs. Table 3 also shows the effects on health care costs. The cost of caring for CVD, diabetes, and CHD over the coming 30 years will be in the order of $9.5 trillion. If all the recommended prevention activities were applied with 100% success, those costs would be reduced by ∼$904 billion, or almost 10%. However, assuming the costs shown in Table 2, the prevention activities themselves would cost ∼$8.5 trillion, offsetting the savings by a factor of almost 10 and increasing total medical costs by ∼$7.6 trillion (162%). The far right column of Table 3 shows the cost per QALY for each activity, assuming the reference costs in Table 2. Smoking cessation is the only prevention activity that can be expected to save money, with the reductions in costs of events more than offsetting the cost of the smoking cessation programs. Next in cost-effectiveness is the use of aspirin in high-risk individuals. The effects on the same outcomes using the maximum feasible levels of prevention activities are shown in Table 4.

be expected to save money, with the reductions in costs of events more than offsetting the cost of the smoking cessation programs. Next in cost-effectiveness is the use of aspirin in high-risk individuals. The effects on the same outcomes using the maximum feasible levels of prevention activities are shown in Table 4. Tables 3 and 4 show the effect of the prevention activities on the U.S. population as a whole. The effects take into account two factors, the number of people who are candidates for a particular activity and the effect of the activity on those who are candidates (i.e., effect/person × number of people). Table 5 shows the benefits of prevention from the perspective of the individuals who have particular risk factors. Each row shows the absolute risk reduction or magnitude of the outcome over 30 years and, where applicable (i.e., for MI and stroke), the number needed to treat (NNT) to prevent one event (30-year NNT). The table also shows the increase in life expectancy, with and without adjustment for quality of life, for those who are candidates for each activity. In some cases, the prevention activity increases a person's length of life by an amount sufficient to then increase their risk of an adverse outcome (e.g., A1C control on strokes).

lso shows the increase in life expectancy, with and without adjustment for quality of life, for those who are candidates for each activity. In some cases, the prevention activity increases a person's length of life by an amount sufficient to then increase their risk of an adverse outcome (e.g., A1C control on strokes). Sensitivity analysis Table 6 summarizes the results of the sensitivity analysis on cost per QALY for a range of assumptions about the cost of the prevention activities (±20%), quality-of- life weights (±20%), unrelated medical costs ($0 to $10,000/person/year), the cost of dying ($0 to $40,000), the cost of general medical care (±20%), and discount rates (0 to 6%). The most important determinants of the costs and cost per QALY are the costs of the prevention activities themselves (Table 7).

quality-of- life weights (±20%), unrelated medical costs ($0 to $10,000/person/year), the cost of dying ($0 to $40,000), the cost of general medical care (±20%), and discount rates (0 to 6%). The most important determinants of the costs and cost per QALY are the costs of the prevention activities themselves (Table 7). CONCLUSIONS Our results lead to seven main conclusions. First, there are large gaps in the application of prevention, and thus large opportunities to reduce the morbidity and mortality of CVD. Even after taking into account current use of prevention activities, the great majority of adults in the U.S. today (78%) still meet the indications for at least 1 of the 11 prevention activities we studied (Table 1). If every person could receive the prevention activities for which he or she is a candidate, MIs could be reduced >60% (from ∼43 million over 30 years to ∼16 million), strokes could be reduced ∼30% (from ∼33 million over 30 years to ∼23 million), and everyone's life expectancies could be increased an average of 1.3 years and at a higher quality of life than currently experienced. Second, even if the full potential of prevention cannot be achieved because of incomplete performance, compliance, and effectiveness, the benefits of aggressive but feasible levels of performance are still large. If performance levels could be uniformly raised to those achieved by the best health care delivery systems (Table 1), 36% of heart attacks and 20% of strokes would be prevented, and life expectancies would be increased an average of 0.7 years.

e benefits of aggressive but feasible levels of performance are still large. If performance levels could be uniformly raised to those achieved by the best health care delivery systems (Table 1), 36% of heart attacks and 20% of strokes would be prevented, and life expectancies would be increased an average of 0.7 years. Third, the 11 prevention activities vary widely in their effectiveness. Viewed from the perspective of the U.S. population as a whole (Table 3), the effects on MIs range from prevention of ∼7.1 million with weight control (BMI <30 kg/m2) to <1 million for cholesterol treatment in low-risk people (LDL cholesterol <160 mg/dl). From the perspectives of individuals who are candidates for particular prevention activities (Table 5), the benefits range from an absolute reduction of MI by 39% (30-year NNT = 3) by control of LDL cholesterol <100 mg/dl in people with established CAD to a decrease in the chance of an MI by an absolute 5% (30-year NNT = 21) by control of LDL cholesterol in people who are at low risk.

lar prevention activities (Table 5), the benefits range from an absolute reduction of MI by 39% (30-year NNT = 3) by control of LDL cholesterol <100 mg/dl in people with established CAD to a decrease in the chance of an MI by an absolute 5% (30-year NNT = 21) by control of LDL cholesterol in people who are at low risk. Fourth, as they are currently delivered, almost all of the prevention activities are expensive. If applied fully, using current protocols and the reference assumptions about costs (Table 2), they would increase health care costs by ∼$8.5 trillion over 30 years (Table 3), or ∼$283 billion per year, or ∼$1,700 per person per year (data not shown). The only cost-saving activity is smoking cessation. Even if $600 is spent annually (versus $350, as shown in Table 2) helping a smoker quit, the savings from preventing downstream CVD events more than offset those costs, yielding a net savings. Aspirin use is relatively inexpensive even if delivered with annual visits; net costs are ∼$50 billion over 30 years, or ∼$90 per candidate per year (Table 3). The other 11 activities increase costs from $0.4 trillion to $1.8 trillion over 30 years (Table 4).

events more than offset those costs, yielding a net savings. Aspirin use is relatively inexpensive even if delivered with annual visits; net costs are ∼$50 billion over 30 years, or ∼$90 per candidate per year (Table 3). The other 11 activities increase costs from $0.4 trillion to $1.8 trillion over 30 years (Table 4). Fifth, the activities vary widely in the value they provide, as measured by cost per QALY (Table 3). Only smoking cessation can be expected to save money over the 30-year follow-up period, and even that does not begin to save money until after 8 years (data not shown). Aspirin for high-risk people has a low cost per QALY (<$3,000). Weight control and control of pre-diabetes (fasting plasma glucose <110 mg/dl) have costs per QALY of ∼$18,000. The next five—blood pressure control in diabetic and nondiabetic people and LDL cholesterol control in high-risk people and people with CAD or diabetes—have cost per QALY between $20,000 and the often-cited but arbitrary threshold of $50,000. The lowest value is provided by LDL cholesterol control in low-risk people, ∼$270,000/QALY. The latter has important policy and clinical implications, as it is currently one of the most heavily promoted of all the prevention activities. If the objective is to prevent CVD, then smoking cessation, aspirin, and control of pre-diabetes and weight would be better uses of resources.

low-risk people, ∼$270,000/QALY. The latter has important policy and clinical implications, as it is currently one of the most heavily promoted of all the prevention activities. If the objective is to prevent CVD, then smoking cessation, aspirin, and control of pre-diabetes and weight would be better uses of resources. Sixth, the “importance” of the prevention activities, in terms of MI and stroke reduction, varies depending on whether the benefits are viewed from the perspective of the population as a whole (Tables 3 and 4) or the individuals who are candidates (Table 5). The former takes into account the number of people who are candidates for an activity, as well as the amount of benefit per candidate. The latter measures only the amount of benefit per candidate. A case in point is LDL cholesterol control in people with established CAD. The benefits of treatment of individuals with CAD who have LDL cholesterol >100 mg/dl are the largest of all the prevention activities (an absolute reduction of MI risk of 40%). However, for the population as a whole, this activity ranks 7th in terms of the number of MIs prevented. Although the per-person benefits are large, only a small proportion (∼1.6%) of the population is a candidate for this activity.

the largest of all the prevention activities (an absolute reduction of MI risk of 40%). However, for the population as a whole, this activity ranks 7th in terms of the number of MIs prevented. Although the per-person benefits are large, only a small proportion (∼1.6%) of the population is a candidate for this activity. Seventh, for the purposes of reducing the costs of the prevention activities, the most important component is the cost of the interventions themselves: the drugs, weight loss programs, and smoking cessation programs. If ways could be found to reduce the costs of the interventions, overall costs could be reduced and value could be increased to reach more acceptable levels (Table 7).

n activities, the most important component is the cost of the interventions themselves: the drugs, weight loss programs, and smoking cessation programs. If ways could be found to reduce the costs of the interventions, overall costs could be reduced and value could be increased to reach more acceptable levels (Table 7). All of these conclusions are very robust to a wide range of assumptions (Table 6). However, as with any cost-effectiveness analysis or clinical trial, the specific results in the tables should be considered only approximate, for several reasons. First, because risk factors, behaviors, practice protocols, performance levels, and costs vary widely across the country, there is no single set of results that will be accurate in every setting. Second, behaviors, tests, treatments, and other factors will inevitably change in ways that cannot be predicted today. Third, actual practices will deviate from the scenarios we have analyzed. For example, while we analyzed the effect of treating a variable to the goal specified in national guidelines, some people will be treated to lower levels, while others will not reach the specified goals. Fourth, some prevention activities have effects that go beyond the boundaries of our analysis. For example, we did not include nonmedical costs such as lost productivity and absenteeism, nor do our estimates of savings and effectiveness include the effects of the prevention activities on non-CVD and nondiabetes outcomes, such as the effects of smoking on cancer. Fifth, there is some degree of uncertainty when risk factors are modified in either a real or simulated clinical trial. There is further uncertainty when one carries them out for 30 years. However, we have based the effects of modifying risk factors on the data available in the literature on both natural history and from therapeutic trials. We would hope that our ability to have more cost-effective therapies will improve in the future.

There is further uncertainty when one carries them out for 30 years. However, we have based the effects of modifying risk factors on the data available in the literature on both natural history and from therapeutic trials. We would hope that our ability to have more cost-effective therapies will improve in the future. Last, we did not consider the costs associated with screening to detect individuals with abnormal values. However, for some of the prevention services studied (e.g., those in people with diabetes), monitoring is routine and there is no need for additional testing. For the others, screening adds costs, but since such testing occurs infrequently (i.e., every 3–5 years), the associated costs are not likely to change the relative value of prevention services. Moreover, if screening is bundled at a single office visit (e.g., lipid profile, blood pressure measurement, weight, and smoking status), the overall impact of screening is likely to be negligible.

equently (i.e., every 3–5 years), the associated costs are not likely to change the relative value of prevention services. Moreover, if screening is bundled at a single office visit (e.g., lipid profile, blood pressure measurement, weight, and smoking status), the overall impact of screening is likely to be negligible. To our knowledge, only one other study, conducted by the National Commission on Prevention Priorities (NCCP), has tried to analyze a broad range of prevention activities (44). In that study, each activity was assigned 1–5 points on each of two measures—clinically preventable burden of disease and cost effectiveness—for a total score ranging from 2 to 10. The study also found that for CVD prevention, smoking cessation and aspirin received high scores. However, our analysis differs in many ways: we report the actual number of people who are candidates for each activity, the effects of each activity one by one and in combination, and the numbers of CVD events, costs, and cost per QALY. Other differences are that our analysis is based on a single integrated model and consistent methodology that includes a representative sample of the U.S. population, current use of prevention activities, representative costs, the recommended treatment goals for prevention activities, and a comprehensive sensitivity analysis. The NCCP's analysis was based on the results of cost-effectiveness analyses done separately for each of the prevention activities. Each of the analyses was done by different investigators, using different models, different sets of assumptions, and different populations. None of the populations was a representative sample of the U.S. population, and none of the treatments in the analyses precisely matched the recommended prevention activities.

ies. Each of the analyses was done by different investigators, using different models, different sets of assumptions, and different populations. None of the populations was a representative sample of the U.S. population, and none of the treatments in the analyses precisely matched the recommended prevention activities. In summary, approximately three-fourths of U.S. adults would benefit from at least one recommended prevention activity to reduce the incidence of CVD. Full deployment of these interventions could potentially prevent approximately two-thirds of MIs and one-third of strokes. However, as they are currently delivered, most of the interventions will substantially increase costs. If our health care system were able to reduce the cost of prevention activities, then the full potential for reducing the burden of CVD could be realized. Author contributions: All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: All authors. Acquisition of data: D.E. Analysis and interpretation of data: All authors. Drafting of manuscript: All authors. Obtained funding: R.K., R.M.R., and R.S. Financial disclosures: None reported. This study was funded by the American Cancer Society, the American Diabetes Association, and the American Heart Association.

Author contributions: All authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: All authors. Acquisition of data: D.E. Analysis and interpretation of data: All authors. Drafting of manuscript: All authors. Obtained funding: R.K., R.M.R., and R.S. Financial disclosures: None reported. This study was funded by the American Cancer Society, the American Diabetes Association, and the American Heart Association. We are grateful for the helpful review of the manuscript by the Science Advisory and Coordinating Committee of the American Heart Association and the health professional members of the Executive Committee of the American Diabetes Association.

Perspectives on the News commentaries are part of a free monthly CME activity. The Mount Sinai School of Medicine, New York, New York, designates this activity for 2.0 AMA PRA Category 1 credits. If you wish to participate, review this article and visit www.diabetes.procampus.net to complete a posttest and receive a certificate. The Mount Sinai School of Medicine is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians. This article is based on presentations at the Metropolitan Diabetes Society on 11 December 2007 in New York, New York, and at the American Diabetes Association's 55th Annual Advanced Postgraduate Course, held 1–3 February 2008 in San Francisco, California (these lectures are available online at http://professional.diabetes.org), summarizing a number of somewhat divergent views recently expressed by different speakers on aspects type 2 diabetes treatment.

Diabetes Association's 55th Annual Advanced Postgraduate Course, held 1–3 February 2008 in San Francisco, California (these lectures are available online at http://professional.diabetes.org), summarizing a number of somewhat divergent views recently expressed by different speakers on aspects type 2 diabetes treatment. Mechanistically based treatment considerations At the American Diabetes Association (ADA) Postgraduate Course, Ralph DeFronzo (San Antonio, TX) reviewed the mechanisms of action and utility of various antidiabetic drugs, suggesting that sulfonylureas “are very unlikely to create a durable decline in A1C,” based on understanding of the physiology. Studies with glimepiride (1) and glipizide (2) show falls in fasting glucose of 40–50 mg/dl and in A1C by 1.5%—with monotherapy controlling 25–30% of patients—which he characterized as “a very good effect, initially.” However, DeFronzo said that “after the first 6–12 months the A1C starts to rise progressively.” Sulfonylurea-induced insulin secretion increases portal insulin levels, suppressing hepatic glucose production and lowering fasting glucose to a greater extent than postprandial glucose. In the UK Prospective Diabetes Study (UKPDS), sulfonylureas and insulin reduced microvascular risk by 37%, but myocardial infarction, stroke, and congestive health failure decreased by 14, 12, and 16% (none of the latter decreases reaching statistical significance) (3), leading DeFronzo to contend that “there is no evidence that treatment with insulin-based therapy” reduces macrovascular disease.

microvascular risk by 37%, but myocardial infarction, stroke, and congestive health failure decreased by 14, 12, and 16% (none of the latter decreases reaching statistical significance) (3), leading DeFronzo to contend that “there is no evidence that treatment with insulin-based therapy” reduces macrovascular disease. Insulin resistance is basic to type 2 diabetes, and β-cell failure begins prior to actual development of diabetes with imbalance between insulin resistance and insulin secretion. DeFronzo asserted that β-cell function decreases by approximately 20% by the time glucose intolerance is present, so appropriate treatment approaches must both reverse insulin resistance and improve β-cell function. The ideal antidiabetic agent would correct hyperglycemia, prevent microvascular complications, improve known cardiovascular disease risk factors, prevent macrovascular complications, and correct the pathophysiological disturbances responsible for type 2 diabetes.

reverse insulin resistance and improve β-cell function. The ideal antidiabetic agent would correct hyperglycemia, prevent microvascular complications, improve known cardiovascular disease risk factors, prevent macrovascular complications, and correct the pathophysiological disturbances responsible for type 2 diabetes. At the level of the liver, metformin and thiazolidinediones (TZDs) are similarly effective in improving insulin action, although TZDs are considerably more potent in their peripheral action. DeFronzo stated that TZDs “unequivocally” are β-cell protective, citing the findings of the Diabetes Prevention Program (4) and TRoglitazone In the Prevention Of Diabetes (TRIPOD) (5) studies with troglitazone, the Diabetes REduction Assessment with ramipril and rosiglitazone Medication (DREAM) study findings with rosiglitazone (6), and the Pioglitazone In the Prevention Of Diabetes (PIPOD) (7) and the Actos Now for Prevention of Diabetes (ACT NOW) studies (clinicaltrials.gov, reg. no. NCT00220961) with prioglitazone. During the first 6 months of the A Diabetes Outcome Progression Trial (ADOPT) of individuals with newly diagnosed diabetes comparing glyburide, metformin, and rosiglitazone, glyburide led to particular improvement, but over time “the best drug in this study was rosiglitazone” (8). DeFronzo commented that in addition to the liver, muscle, and the β-cell, “the fourth bad actor is the fat cell,” which is also insulin resistant, leading to overproduction of fatty acids, which further worsen insulin resistance in liver and in muscle and impair β-cell function.

ug in this study was rosiglitazone” (8). DeFronzo commented that in addition to the liver, muscle, and the β-cell, “the fourth bad actor is the fat cell,” which is also insulin resistant, leading to overproduction of fatty acids, which further worsen insulin resistance in liver and in muscle and impair β-cell function. DeFronzo characterized TZDs as the only agents effective in inhibiting lipolysis and reducing levels of inflammatory cytokines. He noted their potential benefits in nonalcoholic steatohepatitis (9). TZDs lower fasting glucose by 40–50 mg/dl, reduce A1C by ∼1.5%, and control diabetes in 25–30% of patients in clinical trials. In studies of both drug-naive and sulfonylurea-treated diabetic patients receiving placebo or one of the TZDs, A1C decreased from 8.5 to 7%, leptin decreased, and adiponectin increased. Although the TZDs are associated with weight gain, given the improved metabolic outcome, DeFronzo described this as a merely “cosmetic” consequence. The ratio of change in insulin divided by change in glucose (a measure of insulin secretion) and measures of insulin resistance both improved with TZD treatment, which DeFronzo considered “definitive” evidence of improvement in β-cell function. Both pioglitazone (10) and rosiglitazone (11) improve nonoxidative glucose disposal, and these drugs reduce multiple components of the insulin receptor substrate. Although rosiglitazone tends to raise LDL and apolipoprotein (apo)B levels while pioglitazone is LDL neutral and decreases apoB and triglycerides, other than the lipid-lowering effect there is little difference. In the PROspective pioglitAzone Clinical Trial In macroVascular Events (PROactive) study of 5,238 high-risk type 2 diabetic individuals, pioglitazone nonsignificantly decreased total events by 10% (12). DeFronzo opined that leg revascularization was an unfortunate component (“a major mistake”) added to the composite primary end point, as it occurred more often with pioglitazone. The “principal secondary end point” of death, myocardial infarction, or stroke did show significant decrease. He suggested, then, that the TZDs “have a particular benefit” (13) and that “if fat stays in fat cells it cannot hurt you,” while elevated levels in intramuscular, intrahepatic, visceral, arterial, and β-cell deposition of fat all have adverse consequences. TZDs increase fat oxidation, perhaps a major explanation of this therapeutic effect.

at the TZDs “have a particular benefit” (13) and that “if fat stays in fat cells it cannot hurt you,” while elevated levels in intramuscular, intrahepatic, visceral, arterial, and β-cell deposition of fat all have adverse consequences. TZDs increase fat oxidation, perhaps a major explanation of this therapeutic effect. Metformin appears to act, at least in part, by activating AMP kinase in a fashion similar to its activation by exercise. The agent decreases hepatic acetyl CoA carboxylase and sterol response element–binding protein 1c expression, both effects reducing hepatic gluconeogenesis. Metformin also exhibits a weak stimulatory effect on muscle glucose uptake, possibly involving AMP kinase and potentially further contributing to the glycemic effect of metformin. In the UKPDS, there was a 29% reduction in microvascular disease, and there were 39, 41, and 42% decreases in myocardial infarction, stroke, and death, respectively, leading DeFronzo to suggest that metformin is preferable to sulfonylureas as initial therapy. He did not discuss the troublesome increase in diabetes-related mortality seen in the UKPDS with the combination of sulfonylureas plus metformin vs. sulfonylureas alone (14). The progressive rise in A1C in the UKPDS also occurred with metformin, leading DeFronzo to conclude that the drug does not stabilize β-cell function.

id not discuss the troublesome increase in diabetes-related mortality seen in the UKPDS with the combination of sulfonylureas plus metformin vs. sulfonylureas alone (14). The progressive rise in A1C in the UKPDS also occurred with metformin, leading DeFronzo to conclude that the drug does not stabilize β-cell function. Exenatide and liraglutide are incretin analogs, representing the use of “a very, very old concept,” described nearly 80 years ago by La Barre (15), that oral glucose elicits a greater insulin response than intravenous glucose in response to an equivalent hyperglycemic stimulus. The effect is mediated by glucagon-like peptide (GLP)-1 and glucose-dependent insulinotropic polypeptide, produced by the L-cells of the ileum and the K-cells of the duodenum, respectively, in response to neuronal signals to the presence of carbohydrate in the gastrointestinal tract, with GLP-1 also having effects on appetite and gastric emptying. Both GLP-1 and glucose-dependent insulinotropic polypeptide are rapidly degraded by dipeptidyl peptidase (DPP)-4, so GLP-1–based therapy can involve either prolongation of half-life by DPP-4 inhibition or administration of a GLP-1 receptor agonist. DeFronzo suggested that GLP-1 receptor analogs also may preserve β-cell function, though he expressed reservations about whether DPP-4 inhibitors will be demonstrated to produce this effect. In initial studies, exenatide increased insulin secretion in type 2 diabetic patients in a dose-related and glucose-sensitive fashion (16). Metformin-treated type 2 diabetic patients receiving 5 and 10 μg exenatide twice daily showed a reduction in A1C by 1.0 and 1.2%, respectively, from baseline levels of ∼8.3%, with evidence of persistence of the effect over 3.5 years in an open-label extension study (although one must realize that this fails to reach the level of evidence of a randomized controlled trial such as ADOPT and the UKPDS). Even in the absence of weight loss, a 0.7–0.8% reduction in A1C was seen, while patients also exhibiting weight loss showed a 1.7% decrease in A1C at 82 weeks in the open-label study.

(although one must realize that this fails to reach the level of evidence of a randomized controlled trial such as ADOPT and the UKPDS). Even in the absence of weight loss, a 0.7–0.8% reduction in A1C was seen, while patients also exhibiting weight loss showed a 1.7% decrease in A1C at 82 weeks in the open-label study. Liraglutide, DeFronzo said, “works in a different way,” primarily lowering fasting glucose, with improvement in A1C similar to that seen with exenatide. Analysis of response to the DPP-4 inhibitor sitagliptin showed 0.6, 0.7, and 0.9% reductions in A1C in monotherapy and in combination with metformin and pioglitazone, respectively, with better effect in newly diagnosed patients (17–20). Sitagliptin does not delay gastric emptying or increase splanchnic glucose uptake and is weight neutral. A meta-analysis of studies of GLP-1 receptor agonists and of the DPP-4 inhibitors reported a 0.2% greater A1C response with the former, which were also associated with weight loss, leading DeFronzo to suggest that these benefits outweigh the patient preference issue of pill vs. injection; however, the majority of the exenatide studies in the meta-analysis had baseline A1C 8.5%, while most of the DPP-4 inhibitor studies had baseline 8%, potentially explaining, in part, the greater reduction in A1C with the former agent. DeFronzo concluded by recommending that type 2 diabetic patients receive “triple agent therapy from the beginning” with pioglitazone, metformin, and exenatide, speculating that it might even be reasonable to begin pharmacologic treatment when patients develop impaired glucose tolerance or, perhaps, even at the time of development of insulin resistance, to prevent the progressive loss of β-cells that has typically occurred by the time of presentation of type 2 diabetes.

atide, speculating that it might even be reasonable to begin pharmacologic treatment when patients develop impaired glucose tolerance or, perhaps, even at the time of development of insulin resistance, to prevent the progressive loss of β-cells that has typically occurred by the time of presentation of type 2 diabetes. Clinically based treatment considerations Mary Ann Banerji (New York, NY) discussed clinical benefits and side effects of glucose-lowering medications at the ADA Postgraduate Course. Diabetes is one of the most common noncommunicable diseases worldwide, with prevalence predicted to increase to 370 million by the year 2030, driven in part by the increasing prevalence of obesity. Epidemiologic evidence does not suggest a threshold A1C for adverse macro- and microvascular outcomes (21). The ADA recommendations are, then, to target “the lowest A1C possible without unacceptable hypoglycemia, with action recommended for A1C 7%.” Given these concepts, the intensiveness of pharmacologic treatment of diabetes in the U.S. has increased, but it is not clear that glycemia is improving. Rather, with conventional approaches, A1C typically remains elevated (22). The current recommendation is that metformin be given to all patients (23) and that addition of basal insulin, a sulfonylurea, or a TZD be considered, although Banerji recommended, “just be careful about glitazones.” The Agency for Healthcare Research and Quality (www.ahrq.gov), has posed the following question: “Do oral diabetes medications for the treatment of adults with type 2 diabetes differ in their ability to affect the following proximal clinical outcomes: A1C, blood pressure, lipids, w[eigh]t, [and] 2 hour postprandial glucose?” Banerji reviewed some of the available information that can be used to address these basic points.

diabetes medications for the treatment of adults with type 2 diabetes differ in their ability to affect the following proximal clinical outcomes: A1C, blood pressure, lipids, w[eigh]t, [and] 2 hour postprandial glucose?” Banerji reviewed some of the available information that can be used to address these basic points. The UKPDS showed that use of sulfonylureas, metformin, or insulin did not maintain patients at goal (24,25). In the ADOPT study, after 2 years, as DeFronzo discussed, rosiglitazone was best at maintaining glycemia. Meta-analysis of a large number of placebo-controlled studies showed mean A1C lowering of 1% with pioglitazone, 1.2% with rosiglitazone, 1.1% with metformin, 1.5% with sulfonylureas, 0.5% with nateglinide, and 0.8% with acarbose (26), leading to the suggestion that the newer agents are not as potent, although this analysis does not control for baseline levels, with Banerji noting that studies beginning at higher baseline A1C levels report greater falls (27), making it likely that the seeming differences between agents are largely explicable by studies carried out from different starting points. “Combination therapy does work,” she noted, in particular citing benefits of administration of metformin with sulfonylureas and with TZDs. There is evidence that triple therapy is similarly effective when either rosiglitazone or insulin glargine is added to a metformin-sulfonylurea combination, although with greater benefit of insulin seen at high basline A1C levels (28). There is a 2–3 mmHg drop in blood pressure with TZDs, a further potential benefit of these agents. The meta-analysis showed, however, that both TZDs increased LDL cholesterol, although they also increased HDL cholesterol; pioglitazone decreased while rosiglitazone increased triglyceride levels. In comparison with metformin, weight increased both with sulfonylureas and with TZDs, without a significant difference between the effect of these two classes in the meta-analysis. Acarbose was weight neutral, and this has been Banerji's clinical experience with metformin as well. In the ADOPT study, weight decreased with metformin and at 5 years was 6.9 and 2.5 kg more with rosiglitazone and with glyburide, respectively. The meglitinides are also useful agents.

two classes in the meta-analysis. Acarbose was weight neutral, and this has been Banerji's clinical experience with metformin as well. In the ADOPT study, weight decreased with metformin and at 5 years was 6.9 and 2.5 kg more with rosiglitazone and with glyburide, respectively. The meglitinides are also useful agents. Continuous glucose monitoring of type 2 diabetic patients shows glycemic variation to lead to increased oxidative stress (29), which may be related to a greater regression of carotid intima-media thickness reported in association with repaglinide than with glyburide (30). Similarly, nateglinide's particular effect on postprandial glycemia leads it to cause less hypoglycemia than glyburide (31).

ation to lead to increased oxidative stress (29), which may be related to a greater regression of carotid intima-media thickness reported in association with repaglinide than with glyburide (30). Similarly, nateglinide's particular effect on postprandial glycemia leads it to cause less hypoglycemia than glyburide (31). Banerji characterized congestive heart failure with TZDs as “a real problem,” such that patients with strong risk factors for heart failure, including having previously had heart failure, should not be considered good candidates for these agents (32). In contrast, stable heart failure is no longer considered a contraindication to use of metformin. If a patient with edema is receiving drugs associated with fluid retention such as nonsteroidal anti-inflammatory agents or has a local cause such as venous insufficiency, the excess fluid retention caused by the TZD, although not representing heart failure, may still be an issue. Fracture and macular edema are additional concerns with TZDs. Lactic acidosis in patients treated with metformin and low cardiac output and gastrointestinal symptoms in patients treated with metformin, as well as acarbose and exenatide, are additional drug-related adverse effects that may be relevant to the choice of treatment for a given individual. Other patient-specific factors for deciding on a treatment approach include the individual's risk of hypoglycemia and of weight gain, their degree of hyperglycemia, and whether there is evidence of renal or hepatic disease. Given these considerations, Banerji suggested that all the oral agents may be appropriate, in different patients, for “first-line” use.

ding on a treatment approach include the individual's risk of hypoglycemia and of weight gain, their degree of hyperglycemia, and whether there is evidence of renal or hepatic disease. Given these considerations, Banerji suggested that all the oral agents may be appropriate, in different patients, for “first-line” use. For an individual with diabetes, treatment with insulin requires a complex set of behaviors. Banerji cited a survey reporting that patients consider the requirement for insulin to be as disadvantageous as having a major complication (33). Although this is likely to depend on the skill of the health care provider in encouraging insulin use, one should certainly be cognizant of quality-of-life factors when considering whether to recommend insulin. Banerji also pointed out that decision making strictly on the basis of A1C fails to take into account the variability of its relationship to glycemia, such that a person with a mean glucose of 150 mg/dl might have an A1C level ranging from 6.5 to 7.4%—a concept recently addressed in some detail elsewhere (34).

insulin. Banerji also pointed out that decision making strictly on the basis of A1C fails to take into account the variability of its relationship to glycemia, such that a person with a mean glucose of 150 mg/dl might have an A1C level ranging from 6.5 to 7.4%—a concept recently addressed in some detail elsewhere (34). The second question asked by the Agency for Healthcare Research and Quality is whether treatment of type 2 diabetes decreased micro- and macrovascular complications. In the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications, there was decreased cardiovascular risk after many years of follow-up (35). Macrovascular outcome benefit was not shown with insulin and sulfonylureas in the UKPDS, and although diabetes-related mortality was lower with metformin monotherapy, it was significantly increased by metformin in combination with sulfonylureas in this study. Banerji extended DeFronzo's ideas on PROactive, pointing out that in addition to there being no significant benefit of pioglitazone in primary outcome, the significant reduction in death, myocardial infarction (other than silent), and stroke appeared to be accounted for by decreased A1C, triglyceride, and blood pressure and increased HDL levels and could well be said to have been offset by increased heart failure and peripheral arterial disease events. Overall, she concluded, evidence of differences in outcome between different oral antidiabetic agents is weak.

red to be accounted for by decreased A1C, triglyceride, and blood pressure and increased HDL levels and could well be said to have been offset by increased heart failure and peripheral arterial disease events. Overall, she concluded, evidence of differences in outcome between different oral antidiabetic agents is weak. Considerations related to ADA treatment guidelines Robert Ratner (Washington, DC) spoke at the Metropolitan Diabetes Society meeting on the ADA/European Association for the Study of Diabetes guidelines, giving his alternative approach to the treatment of type 2 diabetes. He noted that the guidelines recommend initiation of treatment with lifestyle interventions plus metformin upon diagnosis of diabetes, while the strategy adopted in the UKPDS involved a 3-month lifestyle intervention, during which A1C fell from 9% to 7%. Might lifestyle intervention alone, then, be a useful strategy for some diabetic patients? The newest version of the guidelines suggests that TZDs may not be as safe as other approaches to second-line treatment because of issues with heart failure and bone loss; Ratner asked whether this constitutes an appropriate rejection of the use of these agents. Furthermore, he questioned the use of 7% as the A1C target for therapeutic decision making advocated by the guidelines, pointing out that only a minority of diabetic patients achieve A1C <7%. In an analysis of some 300,000 A1C tests performed at a clinical laboratory, just 45% of levels were <7%, and many of these were <6%, suggesting that these tests might have been performed for diagnosis rather than part of following treatment. In the ADA physician-recognition program describing the “best patients” in the “best centers,” Ratner noted, only 25% of patients had A1C <7% in 1997, 37% in 2001, and 46% in 2003. An algorithm-driven treatment protocol implemented in Boston led to only half of patients achieving A1C <7% (36). Ratner concluded, “There's got to be something more to get the A1C down.” The problem, he pointed out, is the progressive loss of glycemic control over time characteristic of type 2 diabetes. In the UKPDS, as discussed by Banerji, at 3, 6, and 9 years, ∼45, 30, and 20% of treated patients, respectively, maintained A1C <7%. In ADOPT, one-third of individuals receiving glyburide, one-quarter of those receiving metformin, and one-fifth of those receiving rosiglitazone failed at 5 years with regard to the much more readily achieved goal of reaching fasting blood glucose ≤180 mg/dl.

d 20% of treated patients, respectively, maintained A1C <7%. In ADOPT, one-third of individuals receiving glyburide, one-quarter of those receiving metformin, and one-fifth of those receiving rosiglitazone failed at 5 years with regard to the much more readily achieved goal of reaching fasting blood glucose ≤180 mg/dl. “Is the problem with our patients and our doctors,” Ratner asked, “or is it with our interventions?”

d 20% of treated patients, respectively, maintained A1C <7%. In ADOPT, one-third of individuals receiving glyburide, one-quarter of those receiving metformin, and one-fifth of those receiving rosiglitazone failed at 5 years with regard to the much more readily achieved goal of reaching fasting blood glucose ≤180 mg/dl. “Is the problem with our patients and our doctors,” Ratner asked, “or is it with our interventions?” Ratner suggested that although all diabetic individuals can in principle achieve A1C <7% with insulin, large doses are required, and the patient and physician must accept high levels of hypoglycemia—and weight gain. In the Treating to Target in Type 2 Diabetes study of 708 individuals with sulfonylurea and metformin failure, randomized to aspart 70/30 twice daily, detemir daily, or aspart three times daily and using a carefully followed algorithm, 9, 18, and 4%, respectively, required the addition of another insulin dose, with final treatment dose 0.52, 0.49, and 0.61 units · kg−1 · day−1 and with A1C decreasing to 7.3, 7.6, and 7.2% from a baseline level of 8.4% (37). Why, he asked, did not every patient achieve an A1C level <7%? This goal was in fact attained in the study in only 42, 28, and 49% of treated individuals. A conclusion may be that “we do not have the systems … to deliver optimum diabetes care to the masses” and that we need newer and better treatment approaches. Further, Ratner commented, “we are being naive if we believe that … the only goal” is to achieve an A1C level <7% or, according to the more aggressive American Association of Clinical Endocrinologists guidelines, <6.5%. Weight-neutral, or, better, weight loss–inducing treatment approaches with lower risk of significant hypoglycemia are required to maximize patient adherence, as well as to reduce cardiovascular disease risk and cardiovascular disease outcomes. New drugs must be safe and must also be effective, with physicians recognizing that “diabetes is a serious disease” with a complex risk-benefit equation. Weight gain is seen with TZDs, sulfonylureas, and insulin, and hypoglycemia is seen with sulfonylureas and particularly with insulin, with 20% of prandial insulin patients in the Treating to Target in Type 2 Diabetes study experiencing it.

izing that “diabetes is a serious disease” with a complex risk-benefit equation. Weight gain is seen with TZDs, sulfonylureas, and insulin, and hypoglycemia is seen with sulfonylureas and particularly with insulin, with 20% of prandial insulin patients in the Treating to Target in Type 2 Diabetes study experiencing it. New agents such as the DPP-4 inhibitors and incretin mimetics fulfill some of these needs, although neither allows the majority of patients to achieve goal, with Ratner suggesting that “these drugs seem to work better the earlier you give them.” Initial sitagliptin plus metformin does appear to potentially be a very effective approach, while only approximately 45% of persons taking sitagliptin as add on to metformin or to pioglitazone attain A1C <7%. It will be crucial to develop durability data, such as the 5-year data from ADOPT and the 10-year data from UKPDS. “We need new therapies,” Ratner said, “because we're not doing well with what we've got…. Even in our clinical trials,” he continued, “[only] 45-50% are achieving target.”

r to pioglitazone attain A1C <7%. It will be crucial to develop durability data, such as the 5-year data from ADOPT and the 10-year data from UKPDS. “We need new therapies,” Ratner said, “because we're not doing well with what we've got…. Even in our clinical trials,” he continued, “[only] 45-50% are achieving target.” An unmet need is to alter the natural history of insulin secretory failure, recognizing that at present we have no real way of measuring β-cell preservation. Even more, we need to develop agents such as AGE blockers and protein kinase C inhibitors, which reduce the adverse effect of hyperglycemia. Ratner acknowledged the need for long-term safety and efficacy data but suggested that there is a “need to go beyond the cost of the drugs,” as pharmaceutical treatments comprise only 11% of health care costs. Cost, he pointed out, is much more strongly associated with complications than with medications (38). Furthermore, although blood pressure treatment is highly cost-effective, other interventions considered reasonable, such as mammography in older women and lipid-lowering treatment, have costs per quality-adjusted life-year roughly comparable with those of glycemic control (39).

complications than with medications (38). Furthermore, although blood pressure treatment is highly cost-effective, other interventions considered reasonable, such as mammography in older women and lipid-lowering treatment, have costs per quality-adjusted life-year roughly comparable with those of glycemic control (39). Perspectives on TZDs and cardiovascular disease Ratner also spoke at the ADA Postgraduate Course, discussing the risk-to-benefit ratio of the TZDs, using material given at the Food and Drug Administration (FDA) hearing in July 2007, available from www.FDA.gov/ohrms/dockets/ac/07/slides/2007–4308s1–00-index.htm. He gave a set of disclosures of his research support, advisory boards, and stock ownership, as well as his “intellectual disclosure” that he is examining the effect of rosiglitazone on coronary atherosclerosis, as measured by intravascular ultrasound. The TZDs, he said, target insulin resistance, improve glycemic control, do not cause hypoglycemia, improve lipids (in different ways with different agents), and appear to benefit β-cell function. The troglitazone experience, with the idiosyncratic side effect of liver damage considered along with the multiple potential benefits of this class, is important to note “because we've come full circle” to realizing the TZD class effect of reducing liver fat, with preliminary evidence of improvement in nonalcoholic steatohepatitis. Ratner suggested that, similarly, we should not rush to decide that there is cardiac toxicity associated with any currently used TZD. The TZDs do have side effects. Weight gain is typical, although correlating with the degree of improvement in A1C. The greatest weight gain is seen in patients receiving sulfonylureas and with insulin treatment, and this can be attenuated with caloric restriction. TZDs are associated with edema, particularly when used in combination with insulin, although with no evidence of decrease in cardiac performance. Left ventricular contractility, stroke volume, cardiac index, systemic vascular resistance, and blood pressure all improve, as documented initially with troglitazone (40). The edema and heart failure associated with TZD use, then, are related to volume overload and preload rather than to intrinsic adverse cardiac effect; this was addressed in the ADA/American Heart Association consensus statement on TZD use and fluid retention (32).

mprove, as documented initially with troglitazone (40). The edema and heart failure associated with TZD use, then, are related to volume overload and preload rather than to intrinsic adverse cardiac effect; this was addressed in the ADA/American Heart Association consensus statement on TZD use and fluid retention (32). As discussed by DeFronzo, both the Diabetes Prevention Program and TRIPOD studies suggested that administration of the troglitazone led to β-cell rest, reducing progression of pre-diabetes to diabetes by 75% during 1.5 years and by 55% over 2.5 years, respectively, with similar reductions in development of diabetes by approximately 60% in the ACT NOW and DREAM studies. The ADOPT trial showed that in newly diagnosed type 2 diabetic patients, glycemic control was more durable with rosiglitazone than with metformin or glyburide. Ratner pointed out that TZDs, then, reduce development of diabetes and lower A1C, while improving insulin sensitivity, lowering free fatty acids, improving blood pressure, decreasing albuminuria, lowering C-reactive protein, and increasing adiponectin. Reduction of carotid intima-media thickness has been shown with troglitazone (41), pioglitazone (42–44), and rosiglitazone (45). TZDs have also been shown to markedly reduce rates of restenosis following coronary angioplasty and stent procedures (46), to decrease hepatic steatosis and improve steatitis (47), and to reduce waist-to-hip ratio. Lipid effects differ somewhat between rosiglitazone and pioglitazone, with both lowering free fatty acids but with the latter having greater effect in reducing triglycerides and not changing LDL, which increases with rosiglitazone, whereas HDL cholesterol increases 2.4 and 5.2 mg/dl, respectively, with the two agents (48). Although TZDs increase heart failure rates, following hospitalization, individuals who have received TZDs have better clinical outcome (49), further suggesting a benefit of the approach. In the PROactive trial, there was a 10% reduction in the primary outcome (from 23.5 to 21%) among individuals receiving placebo vs. pioglitazone over 36 months (12). As event rates only began to separate around 24 months, Ratner commented, it is entirely possible that a longer trial would have shown stronger evidence of benefit.

active trial, there was a 10% reduction in the primary outcome (from 23.5 to 21%) among individuals receiving placebo vs. pioglitazone over 36 months (12). As event rates only began to separate around 24 months, Ratner commented, it is entirely possible that a longer trial would have shown stronger evidence of benefit. Reports of the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI-2D), the Action to Control CV Risks in diabetes (ACCORD), and the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trials will soon be available, although Ratner cautioned that power calculations suggest that patients already receiving aspirin, converting enzyme inhibitors, and statins have relatively low event rates, such that larger numbers are needed than was recognized when these trials began. The DREAM trial suggested that cardiovascular outcomes increased with rosiglitazone, although only the increase in heart failure was significant. In a controversial meta-analysis purporting to show an adverse cardiovascular effect of rosiglitazone (50), myocardial ischemia events were post hoc, nonadjudicated end points; there was no access to the actual data; and of the 42 studies used, only 11 were peer reviewed, with 26 never published. There were small numbers of events, and the trials were of short duration. It has been said, Ratner commented, that “meta-analysis is to analysis as metaphysics is to physics” (51), with the reported increase in risk belied by the identical incidence of myocardial ischemia events in the rosiglitazone and control groups: 0.6 and 0.62%, respectively. The meta-analysis must, then, give greater weight to some than to other studies. A reanalysis of these data by the FDA using patient-level data found no increase in what was termed “serious ischemia” and found a nonsignificant difference between the 0.73 and 0.67% respective risks of the combination of diagnosed myocardial infarction, cardiovascular disease, and stroke, although the combination of serious plus non-serious ischemia risk was 2 vs. 1.5%, a significant increase. The increased risk of myocardial ischemia was particularly seen when rosiglitazone was administered to individuals taking insulin or nitrates, findings which were incorporated into revised product labeling.

troke, although the combination of serious plus non-serious ischemia risk was 2 vs. 1.5%, a significant increase. The increased risk of myocardial ischemia was particularly seen when rosiglitazone was administered to individuals taking insulin or nitrates, findings which were incorporated into revised product labeling. Further criticism of the original meta-analysis includes its failure to perform a continuity correction, with such analyses demonstrating no significant adverse effect of rosiglitazone (52). Ratner acknowledged that “the trends [with rosiglitazone vs. pioglitazone] are in opposite directions” but questioned the suggestion that the former agent has caused more than 100,000 deaths. The Veterans Health Administration study of the relationship between all-cause mortality and oral antidiabetic drugs showed that with adjustment for age, diabetes duration, A1C, creatinine, cardiovascular history, lipid and blood pressure treatment, and diabetes-related physician visits, there was no significant difference in mortality among 39,721 diabetic patients treated with sulfonylureas, metformin, TZDs, combinations, or no drugs (53). Analysis of a managed-care medication dataset from WellPoint, Inc., in the FDA presentation compared 22,050 individuals receiving rosiglitazone, 23,768 receiving pioglitazone, and 120,771 receiving other agents. Those receiving TZDs were older and had more hospitalizations and higher rates of complications than individuals using other oral hypoglycemic agents, but the comparison failed to reveal differences between the two with regard to myocardial infarction or other complications. The Data Safety Monitoring Boards of the BARI 2D and ACCORD trials, both using rosiglitazone, failed to show an adverse cardiovascular effect of the agent.

s using other oral hypoglycemic agents, but the comparison failed to reveal differences between the two with regard to myocardial infarction or other complications. The Data Safety Monitoring Boards of the BARI 2D and ACCORD trials, both using rosiglitazone, failed to show an adverse cardiovascular effect of the agent. The risk-benefit calculation for TZDs must, then, take into account their glycemic benefit in monotherapy and as adjuvants. They have added benefit in preventing diabetes development and in maintaining glycemic control; may have beneficial β-cell effects, reduce liver fat, and reduce progression of nonalcoholic steatohepatitis; and have pleiotropic that may decrease cardiovascular risk. They do increase weight and cause fluid retention and, now, have been shown to increase fractures. Furthermore, pioglitazone and rosiglitazone have different characteristics. “The available evidence,” Ratner concluded, “is insufficient to definitively determine if TZDs increase, reduce, or have a neutral effect on ischemic CVD or death.” NEWS FROM THE FOOD AND DRUG ADMINISTRATION From time to time, new announcements by the FDA pertaining to aspects of diabetes treatment will be highlighted in this section.

The risk-benefit calculation for TZDs must, then, take into account their glycemic benefit in monotherapy and as adjuvants. They have added benefit in preventing diabetes development and in maintaining glycemic control; may have beneficial β-cell effects, reduce liver fat, and reduce progression of nonalcoholic steatohepatitis; and have pleiotropic that may decrease cardiovascular risk. They do increase weight and cause fluid retention and, now, have been shown to increase fractures. Furthermore, pioglitazone and rosiglitazone have different characteristics. “The available evidence,” Ratner concluded, “is insufficient to definitively determine if TZDs increase, reduce, or have a neutral effect on ischemic CVD or death.” NEWS FROM THE FOOD AND DRUG ADMINISTRATION From time to time, new announcements by the FDA pertaining to aspects of diabetes treatment will be highlighted in this section. A number of agents used or having potential to be used in diabetes treatment have come under scrutiny by the FDA for potential adverse effects related to malignancy. A concern about Regranex (becaplermin) gel, used for the treatment of lower-extremity ulcers, was recently updated with a boxed warning addition to the prescribing information for the agent, based on a study suggesting increased risk of death from cancer in patients treated with three or more tubes of Regranex compared with those who did not use the product. The FDA makes a particular point of recommending that the potential risks of using this agent be discussed with patients and only be used when benefits can be expected to outweigh the risks. This recommendation is in certain ways similar to the boxed warning made for erythropoiesis-stimulating agents several months ago, although there the caution was that people with existing malignancy may have increased mortality and more rapid tumor progression. There has been interest in the role of tumor necrosis factor (TNF)-α in mediating aspects of insulin resistance, with animal models of decreased TNF-α showing improvement in aspects of pre-diabetes and diabetes. It is therefore noteworthy that a number of TNF-α blockers (marketed as Remicade, Enbrel, Humira, and Cimzia), used in conditions such as juvenile idiopathic arthritis and Crohn's disease, are now the subject of an FDA safety review regarding the possibility that these agents may be causally related to development of lymphoma and other cancers in children and young adults. Potential applications to treatment of insulin-resistant states will undoubtedly need to be considered possibly dangerous.

s disease, are now the subject of an FDA safety review regarding the possibility that these agents may be causally related to development of lymphoma and other cancers in children and young adults. Potential applications to treatment of insulin-resistant states will undoubtedly need to be considered possibly dangerous. Antiepileptic drugs, which are extensively used in the treatment of painful diabetic neuropathy, have been shown to have approximately twice the risk of suicidal behavior or ideation (0.43%) as seen in patients receiving placebo (0.22%). Such symptoms have been observed from 1 to 24 weeks after starting the antiepileptic drugs, including carbamazepine (marketed as Carbatrol, Equetro, Tegretol, and Tegretol XR), felbamate (marketed as Felbatol), gabapentin (marketed as Neurontin), lamotrigine (marketed as Lamictal), levetiracetam (marketed as Keppra) oxcarbazepine (marketed as Trileptal), pregabalin (marketed as Lyrica), tiagabine (marketed as Gabitril), topiramate (marketed as Topamax), valproate (marketed as Depakote, Depakote ER, Depakene, Depacon), and zonisamide (marketed as Zonegran).

In the clinical management of diabetes, the A1C assay has become indispensable. Used worldwide to monitor chronic glycemia, the assay is an essential tool to determine whether a patient has achieved the core goal of therapy for diabetes: a marked and sustained reduction in plasma glucose to achieve as close to a normal level as can be safely attained. With the publication of the A1c-Derived Average Glucose (ADAG) study in this issue of Diabetes Care (1), the evolution of the A1C assay continues and an important milestone has been reached. To better appreciate this recent report, a brief and admittedly incomplete historical perspective may be useful. It was 60 years ago that Allen et al. (2) showed that hemoglobin A (which makes up about 97% of total hemoglobin) contains three minor components, designated HbA1a, HbA1b, and HbA1c (A1C). In the decades that followed, we learned that a hexose molecule is attached to these components (3) and that hemoglobin A actually has two more minor glycated derivatives. The five altogether comprise ∼5–7% of the HbA molecule (4).

l hemoglobin) contains three minor components, designated HbA1a, HbA1b, and HbA1c (A1C). In the decades that followed, we learned that a hexose molecule is attached to these components (3) and that hemoglobin A actually has two more minor glycated derivatives. The five altogether comprise ∼5–7% of the HbA molecule (4). In the early course of the biochemical dissection of hemoglobin, Huisman and Dozy (5) noted, virtually in passing, that the level of glycated hemoglobin components was increased in a few individuals they studied who happened to have diabetes. It took 4 more years, however, for Rahbar and colleagues (6,7) to document that diabetes is clearly associated with an elevation in glycated hemoglobin. The Rahbar reports stimulated other investigators to confirm these initial findings and to seek an explanation for how glucose binds to hemoglobin. It was not for another few years, in 1972, that Bunn et al. (8) elegantly showed that the cause of the increased glycated hemoglobin in diabetes, which was predominantly the A1C component, was a result of excess nonenzymatic glycation that occurred throughout the lifespan of red cells and in an essentially irreversible process.

t for another few years, in 1972, that Bunn et al. (8) elegantly showed that the cause of the increased glycated hemoglobin in diabetes, which was predominantly the A1C component, was a result of excess nonenzymatic glycation that occurred throughout the lifespan of red cells and in an essentially irreversible process. The A1C-diabetes story then shifted from clinical chemistry to clinical medicine. Koenig et al. (9) were the first to show that A1C levels correlated well with fasting blood glucose, and they concluded that A1C levels “probably reflect … the mean daily blood glucose concentration … and may provide a better index of control of the diabetic patient.” Indeed, soon after their report, many other investigators confirmed a strong association between A1C and glycemic control and that the measurement had clinical utility (10–15), clearly surpassing in utility what was then the conventional assessment of metabolic control over time (e.g., signs, symptoms, urine, and blood glucose levels) (15). The thorough biochemical experiments performed in the 1970s and 1980s, most notably by Mortensen and Christophersen (16), demonstrated that the fraction of A1C in a sample depends on the glucose levels over a previous period, along with red cell turnover, reaching a steady state sometime between 4 and 12 weeks. Such kinetics were supported by many clinical studies in both type 1 and type 2 diabetic patients where the A1C level was found to correlate well with glucose regulation (17) or the mean blood glucose derived over time from multiple fingersticks (9,15,18–24).

er, reaching a steady state sometime between 4 and 12 weeks. Such kinetics were supported by many clinical studies in both type 1 and type 2 diabetic patients where the A1C level was found to correlate well with glucose regulation (17) or the mean blood glucose derived over time from multiple fingersticks (9,15,18–24). As the use of the A1C test gained traction, dozens of different analytical methods based on different assay principles (e.g., ion-exchange chromatography, affinity chromatography, immunoassay, and electrophoresis) were used to measure glycated hemoglobin. Without a common reference method and in the absence of a standardized assay, results varied considerably when the same sample was tested by different laboratories or methods or even when the same sample was tested repeatedly by one methodology. It was quite common, for example, to have values ranging from 4.0 to 8.1% on the same blood sample (25). In addition, the assays used then (and even now) in clinical medicine not only measured A1C itself but also more or lesser amounts of the other glycated hemoglobin components, and results were reported as A1C, HbA1, or total glycated hemoglobin. The results were also influenced by other interfering substances in the sample.

dition, the assays used then (and even now) in clinical medicine not only measured A1C itself but also more or lesser amounts of the other glycated hemoglobin components, and results were reported as A1C, HbA1, or total glycated hemoglobin. The results were also influenced by other interfering substances in the sample. The Diabetes Control and Complications Trial (DCCT) Study Group, recognizing these problems, centralized the measurement of A1C from the onset of the study so as to avoid confounding results if such a key analyte were to be measured at many sites (26). Also, in anticipation of the DCCT results, the American Association for Clinical Chemistry (AACC) established, in 1993, an A1C standardization workgroup to bring consistency to the measurement of A1C and to facilitate the traceability of results back to the DCCT such that these results could be directly related to the risk or progression of diabetes complications. After the standardization protocol was developed, the American Association for Clinical Chemistry group was dissolved and the National Glycohemoglobin Standardization Program (NGSP) began in 1996 (27). Briefly, in the NGSP, the reference method is the measurement of A1C by ion-exchange high-performance liquid chromatography, as was used in the DCCT. Manufacturers of testing equipment can receive NGSP certification if their instruments are calibrated to match the results obtained by the NGSP. Laboratories can also be certified by the same protocol and thereby document optimal performance in their setting.

-performance liquid chromatography, as was used in the DCCT. Manufacturers of testing equipment can receive NGSP certification if their instruments are calibrated to match the results obtained by the NGSP. Laboratories can also be certified by the same protocol and thereby document optimal performance in their setting. All this has led to a dramatic reduction in interlaboratory variability and a marked improvement in the precision and comparability of values (28). In 2007, ∼99% of all A1C test results in the U.S. were traceable to those obtained in the DCCT, with similar percentages in test results throughout the U.K. and in Canada (D. Sacks, personal communication). Although comparable data are not readily available from other countries, it appears that much of the world's A1C testing is traceable to the DCCT numbers. Still, issues remain. First, the high-performance liquid chromatography reference method used by the NGSP is somewhat nonspecific in that the methodology, like many others, measures more than just A1C in a sample. Although this problem is obviated by the consistent use of one reference method, in the world of clinical chemistry, this situation is “metrologically unsound.” Second, although most methods used worldwide are NGSP certified, there are other standardization programs, most notably in Japan (29) and in Sweden (30). Thus, there is no truly international standardization program.

f one reference method, in the world of clinical chemistry, this situation is “metrologically unsound.” Second, although most methods used worldwide are NGSP certified, there are other standardization programs, most notably in Japan (29) and in Sweden (30). Thus, there is no truly international standardization program. Both of these issues led the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) in 1995 to embark on the development of a reference method that would be very specific, i.e., only measures A1C, and that could lead to worldwide standardization based on a metrologically sound international measurement system (31). Not only did the IFCC succeed (32,33) in developing such an assay but the reference method has been approved by all of their member societies, and a global network of reference laboratories has been established (33). But progress often brings other difficulties and problems. First, the IFCC method is very complicated, requires costly equipment (a mass spectrometer), and is very expensive. Thus, as with many other reference methods, it cannot be used by a clinical laboratory to measure A1C in routine samples. That means it can only be used to calibrate laboratory instruments that measure A1C as before, i.e., by any one of a wide variety of methods. Although disappointing, this does not diminish the virtue of now having a much more robust standardization program.

ed by a clinical laboratory to measure A1C in routine samples. That means it can only be used to calibrate laboratory instruments that measure A1C as before, i.e., by any one of a wide variety of methods. Although disappointing, this does not diminish the virtue of now having a much more robust standardization program. Second, and much more important, since the new reference method measures A1C itself, and thus non-A1C components are no longer detected, the normal range for A1C is reduced—by about two percentage points lower than that currently reported. Moreover, the IFCC recommended (to be metrologically correct) that A1C be expressed in millimoles A1C per mole of total hemoglobin; this would result in a normal range of around 29–43 mmol A1C/mol hemoglobin (34). A shift to lower A1C percentages would no doubt be intolerably confusing and likely lead to a deterioration in glycemic control (35), but a wholesale shift to the IFCC units would surely create mayhem. Although one could clearly program a laboratory instrument to convert the new IFCC values to DCCT-derived values, the IFCC maintained that the expression of an analyte as a percentage is not metrologically sound and thus should not be used. In response to the direction proposed by the IFCC, an American Diabetes Association/European Association for the Study of Diabetes/International Diabetes Federation workgroup was formed (including some of the IFCC leadership) to make recommendations on how this impending crisis could be avoided (36).

hould not be used. In response to the direction proposed by the IFCC, an American Diabetes Association/European Association for the Study of Diabetes/International Diabetes Federation workgroup was formed (including some of the IFCC leadership) to make recommendations on how this impending crisis could be avoided (36). What emerged was not only the recommendation that DCCT-derived numbers should be maintained if possible but also that an international study should commence to look more closely at the relationship between A1C and mean blood glucose. If the study was “successful,” at least we could adopt an A1C-derived unit (e.g., “estimated average glucose” in milligrams per deciliter or millimoles per liter) that would obviate the IFCC objection to having laboratory results expressed as a percentage. This path forward was then ratified in an official consensus statement issued by all four organizations (37).

ld adopt an A1C-derived unit (e.g., “estimated average glucose” in milligrams per deciliter or millimoles per liter) that would obviate the IFCC objection to having laboratory results expressed as a percentage. This path forward was then ratified in an official consensus statement issued by all four organizations (37). The rationale for another study examining the relationship between mean blood glucose and A1C stemmed from the belief that the previously published reports used a variety of measures of glucose concentration, recruited only small numbers of subjects (and mostly those with type 1 diabetes), performed measurements over a relatively short time period, and, most notably, performed relatively infrequent sampling of blood glucose (mostly during the daytime). For example, the often-cited conversion table in the American Diabetes Association Standards of Medical Care in Diabetes—2007 (38) was based on very limited capillary glucose sampling in the DCCT, and the study was actually not intended to establish the relationship between average glucose and A1C. Thus, greater confidence was needed that A1C truly represents an average glucose.

Diabetes Association Standards of Medical Care in Diabetes—2007 (38) was based on very limited capillary glucose sampling in the DCCT, and the study was actually not intended to establish the relationship between average glucose and A1C. Thus, greater confidence was needed that A1C truly represents an average glucose. The results of the international study are now reported (1) and confirm and extend previous findings. The strengths of the study are that it examined the relationship between average glucose and A1C across a wide spectrum of A1C values—from ∼5% to as high as 13%—and in more people than ever before studied. Also, both normal subjects and subjects with type 1 and type 2 diabetes were enrolled in numbers sufficient to conclude that the relationship between the two variables was consistent between these subgroups and also in relation to other important variables (i.e., age, ethnicity, smoking). Finally, the study obtained ∼2,700 glucose measurements in each participant, which is far greater than the number obtained in nearly all previous studies. The results clearly support the hypothesis that there is a strong linear relationship between mean blood glucose and A1C, with a coefficient of correlation (R2) of 0.84.

lly, the study obtained ∼2,700 glucose measurements in each participant, which is far greater than the number obtained in nearly all previous studies. The results clearly support the hypothesis that there is a strong linear relationship between mean blood glucose and A1C, with a coefficient of correlation (R2) of 0.84. The data from ADAG indicate that at any mean glucose or A1C level, there is some scatter (see Fig. 1 in the ADAG report), thereby conveying a less than perfect correlation. Is that primarily due to measurement error, or does it suggest that an A1C level reflects processes beyond a straightforward time and glucose concentration–dependent glycation of hemoglobin? Addressing this uncertainty would require an even larger study conducted ideally at only one site, with more diverse subjects, uninterrupted continuous glucose monitoring for months at a time, and, most important, a measurement error much lower than currently seen. The report by Nathan et al. (39) in which 24,000 glucose measurements were done on each participant resulted in an R2 (0.81) and regression equation very similar to that reported in the ADAG study, suggesting that performing more measurements will not in itself improve the correlation. Thus, we have ∼16–19% of the variation unaccounted for, but given that there is a small measurement error in the determination of A1C (perhaps 2–5%), and a larger coefficient of variation in the measurement of glucose (10–20%), the constraints imposed by methodology can explain the residual variation.

The data from ADAG indicate that at any mean glucose or A1C level, there is some scatter (see Fig. 1 in the ADAG report), thereby conveying a less than perfect correlation. Is that primarily due to measurement error, or does it suggest that an A1C level reflects processes beyond a straightforward time and glucose concentration–dependent glycation of hemoglobin? Addressing this uncertainty would require an even larger study conducted ideally at only one site, with more diverse subjects, uninterrupted continuous glucose monitoring for months at a time, and, most important, a measurement error much lower than currently seen. The report by Nathan et al. (39) in which 24,000 glucose measurements were done on each participant resulted in an R2 (0.81) and regression equation very similar to that reported in the ADAG study, suggesting that performing more measurements will not in itself improve the correlation. Thus, we have ∼16–19% of the variation unaccounted for, but given that there is a small measurement error in the determination of A1C (perhaps 2–5%), and a larger coefficient of variation in the measurement of glucose (10–20%), the constraints imposed by methodology can explain the residual variation. We are unable, of course, to conclude from the study that the relationship holds for all populations. That is, many populations (e.g., Asians, Pacific Islanders, children) were not studied, and it is conceivable that the physiology of glycation differs in such groups, although there is no obvious reason why that would be so. A recent study (40) that showed a relatively poor correlation between average glucose and A1C in children should not raise doubts about the translation of the ADAG study to other populations. In that report (40), it is unclear whether the A1C values were stable throughout the study and how many glucose measurements were obtained in each participant, and there are doubts regarding the precision and accuracy of the continuous glucose-monitoring system device used and other issues (1).

dy to other populations. In that report (40), it is unclear whether the A1C values were stable throughout the study and how many glucose measurements were obtained in each participant, and there are doubts regarding the precision and accuracy of the continuous glucose-monitoring system device used and other issues (1). In the ADAG study, the differences between various ethnic groups were not statistically significant. However, the study was not adequately powered to detect such differences and, in one group, the differences came close to being significant. Although other reports have shown an association between ethnicity and A1C at similar levels of glycemia (41,42), in all the studies, glucose measurements were very infrequent, the populations studied were not controlled for hemoglobinopathy, and there were no measures of the rate of glycation as it relates to ethnicity. Clearly, this is an area that needs further investigation.

icity and A1C at similar levels of glycemia (41,42), in all the studies, glucose measurements were very infrequent, the populations studied were not controlled for hemoglobinopathy, and there were no measures of the rate of glycation as it relates to ethnicity. Clearly, this is an area that needs further investigation. It is important to note that the ADAG investigators attempted to study patients with “stable” glycemia—predefined as a change in A1C of <1% during the study—and all but 4% were stable as so defined. It is not surprising that some change would occur, particularly since patients were doing considerably more self-monitoring than in real life. However, a 1% change with a baseline of 11% has implications different from those associated with a similar change at 7%. However, the investigators quite rightly chose the end-of-study A1C to relate to estimated average glucose (eAG); therefore, clinicians can have confidence that the average glucose reflects antecedent glycemia over a 3-month period.

ne of 11% has implications different from those associated with a similar change at 7%. However, the investigators quite rightly chose the end-of-study A1C to relate to estimated average glucose (eAG); therefore, clinicians can have confidence that the average glucose reflects antecedent glycemia over a 3-month period. So what does this mean for clinical practice? At the simplest and most basic level, when clinicians explain to patients what A1C “means,” they should have greater confidence that the common explanation that has essentially been in effect for decades—“it's your average blood glucose over the last few months”—is true. In addition, knowing one's average glucose level should be beneficial to clinicians and patients in that the measure of long-term glucose control (A1C) can reliably be conveyed in the same units as those provided to patients at the time of diagnosis and the values obtained from patient self-monitoring.

months”—is true. In addition, knowing one's average glucose level should be beneficial to clinicians and patients in that the measure of long-term glucose control (A1C) can reliably be conveyed in the same units as those provided to patients at the time of diagnosis and the values obtained from patient self-monitoring. Finally, we have a new opportunity for (re)education on the importance of glycemic control and the seriousness of diabetes. Because the study results fulfilled the a priori criteria, the agreement forged in the consensus statement from the European Association for the Study of Diabetes, International Diabetes Federation, and IFCC (37) will take effect. Thus, we hope that clinicians who order an A1C test will receive a lab report containing the familiar A1C value, an eAG derived from that measurement, and a likely-to-be-ignored IFCC unit (in millimoles per mol). The American Diabetes Association and European Association for the Study of Diabetes are planning to begin a comprehensive educational effort on knowing one's average glucose and are publishing a new conversion table in guidelines that is based on the equation derived from the ADAG study.

unit (in millimoles per mol). The American Diabetes Association and European Association for the Study of Diabetes are planning to begin a comprehensive educational effort on knowing one's average glucose and are publishing a new conversion table in guidelines that is based on the equation derived from the ADAG study. Patients may still get confused that the “average” glucose on their own meters does not match the eAG. However, this is also an opportunity to educate patients about fluctuations in glucose that may occur at times different from their own testing schedules. In addition, the 95% CIs of eAG for any A1C value imply uncertainty of the “true” mean—and more so at very high A1C levels. But it should be remembered that every point estimate in medicine has uncertainty related to laboratory imprecision and inaccuracy and that this variation is almost always ignored. If necessary, however, these CIs give clinicians an opportunity to present patients a “range” in which their average glucose lies. Another potential limitation of the study is the specific and careful exclusion of individuals with conditions likely to affect A1C, e.g., hemoglobinopathy. Inadequate recognition of the latter in clinical practice remains a limitation to the interpretation of A1C and will thereby limit the utility of discussing eAG in such patients.

al limitation of the study is the specific and careful exclusion of individuals with conditions likely to affect A1C, e.g., hemoglobinopathy. Inadequate recognition of the latter in clinical practice remains a limitation to the interpretation of A1C and will thereby limit the utility of discussing eAG in such patients. Despite the limitations discussed above, the study by Nathan and colleagues (1) is likely to remain a key reference regarding the relationship between A1C and average glucose. To be sure, the term A1C, along with its current units and normal range, will not vanish or change. Also, whatever instrument and assay is used in a clinical laboratory will continue to remain the same, even though the reference method used for calibration will now be more precise. A provider wedded to conveying an A1C to his or her patients will certainly be able to continue doing so. But for those interested in adding another strategy to improve outcomes, we now have a new term that will likely be easier to explain to patients and to convey more meaning and importance to glucose control.

One of the enduring myths in health care is that prevention saves money. The intuition behind the idea appears unassailable: if we spend a little money on preventive measures, we can avoid the expensive complications of disease. The maxim “an ounce of prevention is worth a pound of cure,” coined by Benjamin Franklin and listed in the Dictionary of Cultural Literacy (1), seems self-evident such that it does not warrant serious challenge. The “prevention saves money” argument is popular with many politicians and health policy makers. For example, the recent Presidential primary season witnessed pledges from many candidates not only to improve health care but also to save money by spending more on prevention (2). Careful analysis has shown that while preventive care sometimes saves the health system money, usually it does not (3). Economic evaluations tend to conclude that although preventing the need for treatment can offset some costs from avoidable illness, the overall costs can be substantial because of the number of people who must undergo preventive measures (2). Moreover, prevention does not avert death; it only postpones it. As people age, their health costs increase. Therefore, activities and programs that achieve savings by preventing a fatal disease can lead to downstream health care costs as the people who would otherwise die grow older.

t undergo preventive measures (2). Moreover, prevention does not avert death; it only postpones it. As people age, their health costs increase. Therefore, activities and programs that achieve savings by preventing a fatal disease can lead to downstream health care costs as the people who would otherwise die grow older. Of course, this result does not mean that as a society we should not invest in prevention but, instead, that we should do so with careful analysis and without unrealistic expectations. Preventive measures costing more than they save can still represent good value for money, but whether they will depends entirely on the intervention and population at hand. The study by Kahn et al. is a welcome addition to this discussion (4). Their analysis, which examined the effects of 11 nationally recommended prevention activities for cardiovascular disease (CVD)-related morbidity, mortality, and costs in the U.S., is important on several levels.

intervention and population at hand. The study by Kahn et al. is a welcome addition to this discussion (4). Their analysis, which examined the effects of 11 nationally recommended prevention activities for cardiovascular disease (CVD)-related morbidity, mortality, and costs in the U.S., is important on several levels. First, it quantifies the enormous potential health gains from preventing CVD morbidity and mortality. Ensuring that individuals receive appropriate preventative measures would, at feasible levels of implementation, substantially reduce the number of myocardial infarctions and strokes, adding as many as 130 million life-years and nearly 150 million quality-adjusted life-years to the U.S. adult population over a 30-year period. Aggregating over the population, the biggest reductions in life-years lost would come from controlling blood pressure in the nondiabetic population (<140/90 mmHg), controlling blood pressure in people with diabetes (<130/80 mmHg), and controlling pre-diabetes (fasting plasma glucose <110 mg/dl). Controlling LDL cholesterol levels (<100 mg/dl) in coronary artery disease patients achieves the largest improvement in individual life expectancy, but the population impact is limited by the fact that only 1.5% of all individuals fall into this category.

nd controlling pre-diabetes (fasting plasma glucose <110 mg/dl). Controlling LDL cholesterol levels (<100 mg/dl) in coronary artery disease patients achieves the largest improvement in individual life expectancy, but the population impact is limited by the fact that only 1.5% of all individuals fall into this category. Second, the authors arrive at a conclusion consistent with previous studies but that will nonetheless come as a surprise to many readers. Despite the health gains, only one of the 11 evaluated preventative activities (smoking cessation) is projected to save money. Although 100% implementation of all 11 measures reviewed would decrease the $9.5 trillion cost of caring for CVD, diabetes, and congestive heart failure over the next 30 years by approximately $900 billion, the prevention activities themselves would cost $8.5 trillion, thus increasing total medical costs by some $7.6 trillion. Third, the paper shows the value of formal cost-effectiveness analysis to help guide priority setting and adds to the ever-growing number of cost-utility studies in the literature (5). This kind of formal analysis provides decision makers with a structured, rational approach for considering competing alternatives for improving the return on resources expended. Kahn et al. show that smoking cessation, aspirin, and control of pre-diabetes likely represent the most efficient way to reduce CVD-related morbidity and mortality.

alysis provides decision makers with a structured, rational approach for considering competing alternatives for improving the return on resources expended. Kahn et al. show that smoking cessation, aspirin, and control of pre-diabetes likely represent the most efficient way to reduce CVD-related morbidity and mortality. Fourth, the paper illustrates the use of sophisticated computer-simulation models to help inform policy decisions. The Archimedes model used by Kahn et al. draws on detailed information on U.S.-population disease patterns and health care utilization to analyze downstream clinical events and costs. As the authors note, this kind of model is particularly useful for research on preventative activities, where randomized controlled trials are often unavailable and would be prohibitively costly to conduct. Models do not replace randomized controlled trials but can extend them and enhance our knowledge by simulating clinical trials and asking “what if” questions: How much morbidity, mortality, and cost are potentially preventable? How do various prevention activities compare in terms of their cost-effectiveness?

conduct. Models do not replace randomized controlled trials but can extend them and enhance our knowledge by simulating clinical trials and asking “what if” questions: How much morbidity, mortality, and cost are potentially preventable? How do various prevention activities compare in terms of their cost-effectiveness? A common objection to the issue of simulation models is that their estimates come at the cost of introducing uncertain assumptions (6). However, the methodology does not introduce the uncertainty associated with these assumptions. Rather, it makes the assumptions contributing to this uncertainty explicit and quantifies the uncertainty's impact on the estimated results. For example, by repeating their analysis using different sets of plausible parameters, Kahn et al. determined that assumptions about the cost of the preventive measures and the proportion of eligible individuals receiving services contribute the most to the final result's uncertainty. To be sure, the programming and mathematics underlying the Archimedes model are somewhat complex, but the alternative to formal modeling is to make policy based on implicit assumptions about future consequences.

proportion of eligible individuals receiving services contribute the most to the final result's uncertainty. To be sure, the programming and mathematics underlying the Archimedes model are somewhat complex, but the alternative to formal modeling is to make policy based on implicit assumptions about future consequences. Finally, the study underscores the fact that the issue in prevention is largely not one with insufficient evidence regarding which clinical prevention activities work. Though there are undoubtedly opportunities to improve this evidence base, the activities analyzed by Kahn et al., as the authors highlight, are supported by good evidence and are widely accepted. Instead, the problem is how to increase their use in an efficient manner. Part of the answer lies in better health promotion activities and education and outreach efforts to deliver services and improve compliance. These activities cost money but may well be worth the needed resources. We must improve the evidence base on the cost-effectiveness of such efforts. Part of the answer also lies in systemic changes, such as infrastructure to link clinicians with community resources, and changes in the system to create incentives that encourage the appropriate delivery of efficient interventions (7).

urces. We must improve the evidence base on the cost-effectiveness of such efforts. Part of the answer also lies in systemic changes, such as infrastructure to link clinicians with community resources, and changes in the system to create incentives that encourage the appropriate delivery of efficient interventions (7). In an ideal world, critical health care–spending decisions would be informed by directly applicable randomized clinical trial results. Because logistical and ethical constraints make such evidence unavailable in many cases, information must be synthesized and extrapolated across time and to populations not directly studied. As the analysis by Kahn et al. demonstrates, the results are not always consistent with intuition; in this case, it turns out that prevention generally does not save money. The analysis does, however, identify interventions that efficiently produce health improvements and, hence, suggests programs that policy makers should target to increase population compliance. Given the substantial health implications and health care resources involved, using tools like the Archimedes model can help ensure that we are making the best use of available information to identify measures to improve public health.

Goldstein BJ, Feinglos MN, Lunceford JK, Johnson J, Williams-Herman DE, for the Sitagliptin 036 Study Group: Effect of initial combination therapy with sitagliptin, a dipeptidyl peptidase-4 inhibitor, and metformin on glycemic control in patients with type 2 diabetes. Diabetes Care 30:1979–1987, 2007 In Table 2 of the article above, an asterisk was missing denoting a significant difference between the 1,000 mg metformin b.i.d. group and placebo. The corrected table appears on the following page. The online version reflects these changes. Table 2 Efficacy end points following a meal tolerance test Parameter Placebo Sitagliptin 100 mg q.d. Metformin 500 mg b.i.d. Metformin 1,000 mg b.i.d. Sitagliptin 50 mg + metformin 500 mg b.i.d. Sitagliptin 50 mg + metformin 1,000 mg b.i.d.

In Table 2 of the article above, an asterisk was missing denoting a significant difference between the 1,000 mg metformin b.i.d. group and placebo. The corrected table appears on the following page. The online version reflects these changes. Table 2 Efficacy end points following a meal tolerance test Parameter Placebo Sitagliptin 100 mg q.d. Metformin 500 mg b.i.d. Metformin 1,000 mg b.i.d. Sitagliptin 50 mg + metformin 500 mg b.i.d. Sitagliptin 50 mg + metformin 1,000 mg b.i.d. 2-h PPG (mg/dl) n 129 136 141 138 147 152 Baseline 276.8 ± 66.7 285.4 ± 82.8 292.7 ± 74.6 283.4 ± 81.8 291.8 ± 84.6 286.9 ± 76.2 Week 24 281.3 ± 88.2 233.8 ± 89.5 236.3 ± 70.8 207.0 ± 69.4 196.3 ± 72.5 170.3 ± 58.6 Change from baseline 0.3 (−10.4 to 11.0) −51.9 (−62.3 to −41.5) −53.4 (−63.6 to −43.2) −78.0 (−88.3 to −67.6) −92.5 (−102.6 to −82.5) −116.6 (−126.4 to −106.7) Change from placebo — −52.2 (−67.1 to −37.3)* −53.7 (−68.5 to −38.9)* −78.3 (−93.1 to −63.4)* −92.8 (−107.5 to −78.1)*† −116.9 (−131.4 to −102.3)*† Glucose AUC (mg · h−1 · dl−1) n 127 133 144 137 149 146 Baseline 523.0 ± 109.4 530.2 ± 123.9 538.4 ± 116.0 532.1 ± 125.6 539.9 ± 136.7 530.0 ± 120.7 Week 24 533.3 ± 146.9 453.2 ± 135.3 451.3 ± 118.4 401.7 ± 120.4 383.9 ± 114.0 339.0 ± 100.4 Change from baseline 6.1 (−11.7 to 23.9) −78.0 (−95.4 to −60.6) −84.4 (−101.1 to −67.7) −130.8 (−147.9 to −113.6) −152.4 (−168.9 to −136.0) −192.3 (−209.0 to −175.7) Change from placebo — −84.2 (−109.1 to −59.3)* −90.5 (−115.5 to −66.1)* −136.9 (−161.6 to −112.2)* −158.6 (−182.8 to −134.3)*† −198.5 (−222.8 to −174.1)*† Insulin AUC (μIU · h−1 · ml−1) n 119 117 130 125 126 133 Baseline 76.5 ± 52.9 71.2 ± 50.2 80.1 ± 56.4 84.3 ± 53.6 79.0 ± 57.2 74.9 ± 47.2 Week 24 75.0 ± 50.3 78.1 ± 49.3 83.1 ± 54.6 82.2 ± 51.4 86.3 ± 55.7 74.5 ± 40.6 Change from baseline −1.8 (−7.5 to 3.9) 5.2 (−0.6 to 11.0) 3.6 (−1.9 to 9.1) −0.4 (−6.0 to 5.2) 7.5 (2.0–13.1) −1.1 (−6.5 to 4.4) Change from placebo — 7.0 (−1.2 to 15.1) 5.4 (−2.5 to 13.3) 1.4 (−6.6 to 9.4) 9.4 (1.4–17.3)‡ 0.7 (−7.1 to 8.6) C-peptide AUC (ng · h−1 · ml−1) 125 135 143 138 146 145 n Baseline 9.8 ± 4.3 9.8 ± 4.3 10.3 ± 4.1 10.9 ± 4.3 9.9 ± 4.3 10.1 ± 4.1 Week 24 9.8 ± 4.2 10.5 ± 4.0 10.2 ± 4.1 10.6 ± 4.1 10.6 ± 4.1 10.4 ± 4.1 Change from baseline −0.1 (−0.5 to 0.3) 0.6 (0.2–1.0) 0.0 (−0.4 to 0.4) −0.1 (−0.5 to 0.3) 0.6 (0.2–1.0) 0.3 (−0.1 to 0.7) Change from placebo — 0.7 (0.1–1.3)‡ 0.1 (−0.5 to 0.7) 0.0 (−0.6 to 0.6) 0.7 (0.1–1.3)‡§ 0.4 (−0.2 to 1.0) Insulin AUC/glucose AUC 117 115 128 121 125 130 n Baseline 0.16 ± 0.14 0.15 ± 0.12 0.16 ± 0.12 0.18 ± 0.15 0.16 ± 0.14 0.16 ± 0.12 Week 24 0.16 ± 0.14 0.19 ± 0.13 0.20 ± 0.15 0.23 ± 0.17 0.24 ± 0.17 0.23 ± 0.14 Change from

om placebo — 0.7 (0.1–1.3)‡ 0.1 (−0.5 to 0.7) 0.0 (−0.6 to 0.6) 0.7 (0.1–1.3)‡§ 0.4 (−0.2 to 1.0) Insulin AUC/glucose AUC 117 115 128 121 125 130 n Baseline 0.16 ± 0.14 0.15 ± 0.12 0.16 ± 0.12 0.18 ± 0.15 0.16 ± 0.14 0.16 ± 0.12 Week 24 0.16 ± 0.14 0.19 ± 0.13 0.20 ± 0.15 0.23 ± 0.17 0.24 ± 0.17 0.23 ± 0.14 Change from baseline 0.00 (−0.02 to 0.02) 0.04 (0.02–0.05) 0.04 (0.02–0.06) 0.05 (0.03–0.07) 0.08 (0.06–0.10) 0.08 (0.06–0.09) Change from placebo — 0.03 (0.01–0.06)‡ 0.04 (0.01–0.06)‡ 0.05 (0.02–0.07)* 0.08 (0.05–0.10)*† 0.07 (0.05–0.10)*‖ Data are means ± SD for baseline and week 24 and least-squares mean change (95% CI) for change from baseline or placebo. * P ≤ 0.001 for the between-group difference relative to placebo. † P ≤ 0.001 for the between-group difference comparing coadministration and its respective components. ‡ P ≤ 0.05 for the between-group difference relative to placebo. § P < 0.05 for the between-group difference comparing coadministration and metformin 500 mg b.i.d. ‖ P ≤ 0.05 for the between-group difference comparing coadministration and its respective components. AUC, area under the curve; PPG, postprandial plasma glucose.

We read with interest the recent article by Abràmoff et al.(1) but were disappointed in their conclusion that automated grading software could not be recommended for clinical practice. Our group's published work (2) shows that automated grading of diabetic retinopathy, based on image-quality assessment and microaneurysm detection, can safely reduce the burden of grading in diabetic retinopathy screening programs. Comparing manual and automated grading against a reference standard grading of 14,406 images (from 6,722 patients), we found that our automated system attained a higher sensitivity for detection of patients requiring “full disease” grading than the manual graders. The automated system detected 97.9% of patients having referable diabetic retinopathy. Although the specificity of the automated system was lower than for manual graders, the grading workload was reduced and offered useful financial savings (3).

tion of patients requiring “full disease” grading than the manual graders. The automated system detected 97.9% of patients having referable diabetic retinopathy. Although the specificity of the automated system was lower than for manual graders, the grading workload was reduced and offered useful financial savings (3). Screening is a means for reducing the risk of disease in the screened population, and, in practice, large-scale implementation means that there is a compromise between sensitivity and specificity. Hence a recommendation against using automated grading is only valid if it is shown that there is a higher performing and readily available alternate methodology. More specifically, it is important that an automated grading system is compared with what can be achieved by human experts who are routinely employed within a screening program. In the real world, such manual grading is imperfect. For example, we found that the full disease graders, whose job is to be highly specific, missed 18 of 330 cases of referable diabetic retinopathy (2). Hence, our main criticism of the study by Abràmoff et al. is that the lack of a common reference standard resulted in insufficient evidence to draw their main conclusion, namely, that the automated grading software could not be recommended for clinical practice.

Screening is a means for reducing the risk of disease in the screened population, and, in practice, large-scale implementation means that there is a compromise between sensitivity and specificity. Hence a recommendation against using automated grading is only valid if it is shown that there is a higher performing and readily available alternate methodology. More specifically, it is important that an automated grading system is compared with what can be achieved by human experts who are routinely employed within a screening program. In the real world, such manual grading is imperfect. For example, we found that the full disease graders, whose job is to be highly specific, missed 18 of 330 cases of referable diabetic retinopathy (2). Hence, our main criticism of the study by Abràmoff et al. is that the lack of a common reference standard resulted in insufficient evidence to draw their main conclusion, namely, that the automated grading software could not be recommended for clinical practice. We also note two other factors that may have influenced the results and made them difficult to generalize. First, selection bias may have been a factor. The data were selected on the basis that patients previously shown to have diabetic retinopathy are not rescreened. While this may be the policy of the EyeCheck program, the data may not be regarded as “unselected” outside the context of this particular program. Second, the authors note that there seemed to be a slight effect associated with increasing camera resolution. However, the results show a variation in specificity from 22 to 83% depending on camera resolution. This suggests that performance may be greatly improved by using the higher resolution images.

icular program. Second, the authors note that there seemed to be a slight effect associated with increasing camera resolution. However, the results show a variation in specificity from 22 to 83% depending on camera resolution. This suggests that performance may be greatly improved by using the higher resolution images. We congratulate Abràmoff et al. on this study. However, we believe that the conclusions are not universally applicable. Our work shows that the automated analysis of retinal images does have an important role to play in diabetic retinal screening programs.

We thank Olson et al. (1) for their close reading of our recent study (2), where we examined sensitivity and specificity of automated diabetic retinopathy detection and demonstrated an area under the receiver operating characteristic curve of 0.87. A limited, 500-patient sample of all 10,000 photographic exams was examined by multiple, masked experts. We felt uncomfortable recommending a system for clinical practice for which patient safety compared to an accepted (gold) standard could not be established, concluding that it should be tested against widely accepted clinical standards, if practical. We have recently presented studies of an improved algorithm on a new, larger dataset of 15,000 exams with an area under the curve of 0.90 (3). Most of these results support the work of Olson and colleagues (4), although the fact that their system only detects small hemorrhages and microaneurysms is a serious limitation in our view, and sensitivity and specificity based upon the photographic interpretation by a single reader is unlikely to become widely accepted. Failure to detect large, rare, and/or advanced lesions deserves disproportionate attention. If a patient with isolated neovascularization of the disc, <1:5,000 in our series, were to be missed by a system but would not have been missed by a person, that is a failure likely to lead to vision loss or blindness for that patient, potential litigation, and a backlash against implementation of automated detection.

If a patient with isolated neovascularization of the disc, <1:5,000 in our series, were to be missed by a system but would not have been missed by a person, that is a failure likely to lead to vision loss or blindness for that patient, potential litigation, and a backlash against implementation of automated detection. Groups translating automated diabetic retinopathy detection into clinical practice operate in environments that differ on regulatory, legal, budgetary, and reimbursement aspects, but we disagree that “a recommendation against automated grading is only valid if it is shown that there is a higher performing and readily alternative methodology” (1). The currently established practice is human expert reading, and the burden of proof is therefore on the new system to be introduced, which is automated reading. For automated reading to gain widespread acceptance, no shortcuts regarding safety concerns will likely be permitted by regulatory agencies, payers, and patients. One study's entry criteria may be perceived as another's selection bias. The target population of the EyeCheck project consists of patients who had not been previously identified to have diabetic retinopathy. In most settings, as patients are identified with diabetic retinopathy they are referred for evaluation or treatment, removing them from the screened population. To establish any other inclusion criteria would have constituted selection, affecting the potential application of this data to current clinical practice.

hy. In most settings, as patients are identified with diabetic retinopathy they are referred for evaluation or treatment, removing them from the screened population. To establish any other inclusion criteria would have constituted selection, affecting the potential application of this data to current clinical practice. The potential positive effect of camera resolution on algorithmic performance is intriguing, although with less costly cameras presently offering at least 1,024 × 1,024 pixels, this debate may be self-limiting. We believe that comparison of algorithms to standardized datasets (http://roc.healthcare.uiowa.edu) as well as to the gold standard are required and should include: 1) demonstration in a prospective multicenter study of similar or better detection; on populations with defined race and ethnicity distributions, 2) acceptable comparison of detection to standard multifield stereo photographs read according to the Early Treatment of Diabetic Retinopathy Study standard; and 3) sensitivity/specificity analysis with standard and severity-weighted receiver operating characteristic curves. In summary, we agree that automated detection of diabetic retinopathy can make the prevention of blindness and vision loss objective, more accessible, and more cost-effective, provided safety issues are not overlooked.

Endothelial dysfunction occurs early in the course of type 2 diabetes and contributes to the development of macrovascular complications of the disease (1,2). Consumption of saturated fatty acids (SAFAs) impairs endothelial function for up to 6 h postmeal (3), whereas data on the effect of monounsaturated fatty acids (MUFAs) on endothelial function in subjects with type 2 diabetes are limited. According to recent nutritional recommendations, individuals with diabetes should substitute SAFA for MUFA in their diet (4), and the predominant source of MUFA in many countries is oleic acid contained in olive oil. However, the effect of consumption of olive oil on endothelial function in subjects with type 2 diabetes is not known. We tested the hypothesis that consumption of MUFA in the form of olive oil exerts a better effect on endothelial function in subjects with type 2 diabetes than that associated with consumption of butter. Because endothelial function is affected by high blood glucose, lipid and insulin concentrations, and increased oxidative stress (2), we measured these parameters during the study. RESEARCH DESIGN AND METHODS We studied 21 men and 12 women with type 2 diabetes attending the outpatient diabetes clinic of Laiko General Hospital. Current smokers, subjects aged >70 years, and those with clinically apparent macrovascular disease, renal impairment or microalbuminuria, A1C >8.5%, and fasting triglycerides >300 mg/dl were excluded.

We studied 21 men and 12 women with type 2 diabetes attending the outpatient diabetes clinic of Laiko General Hospital. Current smokers, subjects aged >70 years, and those with clinically apparent macrovascular disease, renal impairment or microalbuminuria, A1C >8.5%, and fasting triglycerides >300 mg/dl were excluded. We designed a crossover study. Subjects consumed two different standard test meals on two separate mornings. The test meals were given in random order with an interval of ∼1 week in between. The SAFA-rich meal consisted of four pieces of toasted white bread and 40 g butter (total energy content 557.6 kcal, 50.1 g carbohydrates, 9.2 g protein, and 35.6 g fat; 62.9% SAFA, 0.3% polyunsaturated fatty acids [PUFAs], and 31.9% MUFA). The MUFA-rich meal consisted of four pieces of toasted white bread and 33 g extra-virgin olive oil (total energy content 559.4 kcal, 50.1 g carbohydrates, 9.2 g protein, and 35.8 g fat; 14.6% SAFA, 7.9% PUFA, and 77.0% MUFA). Endothelial function was assessed by determination of the change of the brachial artery diameter after removal of ischemic occlusion on the forearm (flow-mediated dilatation [FMD]), as previously described (5). Blood flow was measured at rest and within 15 s after the cuff release. Blood was collected after an overnight fast of 10–12 h for determination of A1C, glucose, lipids, insulin, and total plasma antioxidant capacity (TPAC). FMD, blood flow, and biochemical parameters were determined in the fasting state and 2, 4, and 6 h postprandially.

ured at rest and within 15 s after the cuff release. Blood was collected after an overnight fast of 10–12 h for determination of A1C, glucose, lipids, insulin, and total plasma antioxidant capacity (TPAC). FMD, blood flow, and biochemical parameters were determined in the fasting state and 2, 4, and 6 h postprandially. Two-way ANOVA for repeated measurements was performed to examine the effect of time (within-subject factor), the test meal (between-subjects factor), and their interaction on the studied parameters in the two phases of the study. The observed power of two-way ANOVA for the FMD at a 0.05 level was >90% for the aforementioned effects. RESULTS Mean ± SD age was 58.1 ± 9.2 years, duration of diabetes 3.8 ± 3.2 years, BMI 29.6 ± 4.3 kg/m2, waist circumference 102.8 ± 10.9 cm, and A1C 7.0 ± 1.3%. After consumption of the MUFA-rich meal, FMD did not change, whereas after consumption of the SAFA-rich meal, a significant reduction in FMD was observed (Table 1). The FMD values, expressed as incremental area under the curve, were increased by 5.2 ± 2.5% after the MUFA-rich meal and decreased by 16.7 ± 6.0% after the SAFA-rich meal (Δ = −11.5 ± 6.4% between the test meals, P = 0.008). Baseline brachial artery diameter, baseline and peak blood flow, and percent increase in blood flow in the brachial artery did not change during the study after consumption of either test meal. Additionally, no significant differences in these parameters were observed between the test meals (Table 1).

eals, P = 0.008). Baseline brachial artery diameter, baseline and peak blood flow, and percent increase in blood flow in the brachial artery did not change during the study after consumption of either test meal. Additionally, no significant differences in these parameters were observed between the test meals (Table 1). After consumption of either test meal, plasma glucose, insulin, and triglyceride levels increased during the study, while the concentrations of total and HDL cholesterol and TPAC did not change. No significant differences were found in these parameters between the two meals, and the time-by-meal interaction was not significant (data not shown). CONCLUSIONS The main finding of this study is that consumption of a single MUFA-rich meal in the form of extra-virgin olive oil does not impair endothelial function in subjects with type 2 diabetes. On the contrary, consumption of a SAFA-rich meal exerts a noxious effect on endothelial function that starts at 2 h and is maintained up to 6 h postprandially. Notably, the differential effects of MUFA- and SAFA-rich diets on endothelial function were observed for similar changes in plasma glucose, insulin, and lipid concentrations in TPAC and reactive hyperemia.

eal exerts a noxious effect on endothelial function that starts at 2 h and is maintained up to 6 h postprandially. Notably, the differential effects of MUFA- and SAFA-rich diets on endothelial function were observed for similar changes in plasma glucose, insulin, and lipid concentrations in TPAC and reactive hyperemia. Concerning the effect of MUFA on endothelial function in subjects with type 2 diabetes, one previous study showed that consumption of safflower and canola oil did not impair endothelial function 4 h postmeal (6), while another study demonstrated that substitution of PUFA for olive oil in a diet for 2 months resulted in improvement in FMD (7). Thus, our finding for a protective effect of MUFA on endothelium corroborates these reports. Consumption of olive oil attenuates endothelial cell activation in humans (8,9), and in vitro studies demonstrated that endothelial cells exposed to oleic acid reduce the expression of adhesion molecules (10). Furthermore, extra-virgin olive oil is rich in polyphenols that enhance the formation of nitric oxide by endothelial cells and protect endogenous antioxidant defenses postprandially (11–13). These data suggest that the protective effects of extra-virgin olive oil on endothelium could be due to the oleic acid per se, to the natural antioxidants contained in it, or to both.

nols that enhance the formation of nitric oxide by endothelial cells and protect endogenous antioxidant defenses postprandially (11–13). These data suggest that the protective effects of extra-virgin olive oil on endothelium could be due to the oleic acid per se, to the natural antioxidants contained in it, or to both. Studies examining the effect of diet on endothelial function are of clinical relevance for prevention strategies in subjects with type 2 diabetes, a population vulnerable to macrovascular complications. We studied type 2 diabetic subjects without complications and with short diabetes duration; therefore, our findings cannot be extrapolated to all patients with type 2 diabetes. Moreover, we examined the effect of a single meal on endothelial function; prospective studies are needed to clarify the long-term effects of olive oil consumption on endothelial function.

(see online appendix, available at http://dx.doi.org/10.2337/dc08-1035), 329 were randomized to double-blind treatment. Baseline characteristics were similar among treatment groups. Participants had a mean ± SD age of 53.4 ± 11.1 years and baseline A1C 7.9 ± 0.08% and were predominantly male (53.2%) and white (66.9%). Mean A1C decreased significantly more with 12.5 mg (−0.56%; P < 0.001) or 25 mg (−0.59%; P < 0.001) alogliptin than with placebo (−0.02%) by week 26. Significant A1C reductions were observed as early as week 4 (P < 0.001). FPG reductions were significantly greater with alogliptin than with placebo at week 26 (P < 0.001) and as early as week 1 (P ≤ 0.002). The percentage of patients who required hyperglycemic rescue was significantly less with alogliptin (12.5 mg, 9.8%; 25 mg, 7.6%; P ≤ 0.001) than with placebo (29.7%). Minor reductions in weight with alogliptin were neither clinically nor statistically significant relative to placebo. Results for secondary efficacy measures and exploratory assessments are summarized in Table 1.

nce level of 0.05. Continuous secondary efficacy analyses were performed as for the primary analysis, except that the baseline covariate corresponds with the tested end point. Incidence variables were compared using nonparametric, extended Mantel-Haenszel X2 tests; covariates were the same as for the primary end point. The safety population included patients who took at least one dose of the study drug (double blind). Safety assessments included adverse events, clinical laboratory findings, 12-lead electrocardiograms, physical examination findings, vital signs, and hypoglycemic events. Skin and digits were specifically examined because of lesions previously observed in monkeys given DPP-4 inhibitors other than alogliptin (4–6). Safety findings were summarized with descriptive statistics. RESULTS Of 420 patients enrolled (see online appendix, available at http://dx.doi.org/10.2337/dc08-1035), 329 were randomized to double-blind treatment. Baseline characteristics were similar among treatment groups. Participants had a mean ± SD age of 53.4 ± 11.1 years and baseline A1C 7.9 ± 0.08% and were predominantly male (53.2%) and white (66.9%).

Insulin resistance, metabolic syndrome, and diabetes are associated with inflammation (1–3). It is unclear, however, whether the inflammatory process causes insulin resistance and accelerates progression to diabetes or whether insulin resistance and diabetes increase inflammation. While Alaskan Eskimos have high rates of cardiovascular disease (4) and subclinical infection (5), diabetes, insulin resistance, and metabolic syndrome occur less often than in U.S. whites (6). Therefore, exploration of the relationships between subclinical infection, inflammatory markers, insulin resistance, and diabetes in this population may illuminate possible mechanisms. RESEARCH DESIGN AND METHODS The Genetics of Coronary Artery Disease in Alaska Natives (GOCADAN) study population includes Eskimos residing in the Norton Sound region of Alaska. Of this population, 1,214 family members, aged ≥18 years, were recruited (October 2000–April 2004) (7). Questionnaires provided demographic, health habits, and medical history data. Sitting blood pressure was measured, and the mean of the second and third measurements was used for analysis. Fasting blood measures included glucose, insulin, fibrinogen, high-sensitivity C-reactive protein (CRP), and homocysteine (HCY) (7). Dietary intake was assessed using a food-frequency questionnaire validated for Alaska Natives (8).

e was measured, and the mean of the second and third measurements was used for analysis. Fasting blood measures included glucose, insulin, fibrinogen, high-sensitivity C-reactive protein (CRP), and homocysteine (HCY) (7). Dietary intake was assessed using a food-frequency questionnaire validated for Alaska Natives (8). IgG, IgA, and IgM antibodies to C. pneumoniae and IgG antibodies to other pathogens were determined using commercially available enzyme-linked immunosorbent assay (ELISA) kits (5). Obesity was defined as BMI ≥30 kg/m2, metabolic syndrome was defined by Adult Treatment Panel III criteria, and insulin resistance was defined by homeostasis model assessment (HOMA) (9). Diabetes and impaired fasting glucose (IFG) were defined by American Diabetes Association criteria. Participants wore a Digiwalker pedometer for 7 days.

as BMI ≥30 kg/m2, metabolic syndrome was defined by Adult Treatment Panel III criteria, and insulin resistance was defined by homeostasis model assessment (HOMA) (9). Diabetes and impaired fasting glucose (IFG) were defined by American Diabetes Association criteria. Participants wore a Digiwalker pedometer for 7 days. Excluded from this study were 10 men and 30 women with diabetes. Participant characteristics were compared using the Kruskal-Wallis test, Pearson's χ2 test, Wilcoxon's rank-sum test, or Fisher's exact test. ANOVA was used to compare least-square means of insulin resistance or IFG, as estimated by HOMA of insulin resistance (HOMA-IR), for each quartile of CRP and HCY and number of subclinical infections. Logistic regression was used to compute adjusted odds ratios of metabolic syndrome across CRP or HCY quartiles (using the first quartile as the reference) and levels of subclinical infection (using levels 1 and 2 as the reference). For each analysis, data were first adjusted for age and sex and then analyzed in models including BMI, alcohol use, smoking status, physical activity, and fatty acid intake. In all computations (ANOVA and logistic regression), the transformed variables for HOMA-IR and CRP were used. Participants with CRP >10 mg/l (n = 21 [1.7%]) were excluded from CRP analyses. Because participants were members of extended families, kinship was accounted for in additional analyses.

, and fatty acid intake. In all computations (ANOVA and logistic regression), the transformed variables for HOMA-IR and CRP were used. Participants with CRP >10 mg/l (n = 21 [1.7%]) were excluded from CRP analyses. Because participants were members of extended families, kinship was accounted for in additional analyses. RESULTS Of the 1,174 participants (55% women), mean BMI was 27.5 kg/m2, 31.5% were overweight, and 29.1% were obese. Metabolic syndrome was present in 14.3% (11% men, 17% women). Insulin was low (median 8.1 μU/ml). Mean CRP was 1.6 mg/l (median 0.9 mg/l), and mean HCY was 7.3 μmol/l. Measures for five common pathogens averaged 3.4 per participant, and CRP was positively correlated with pathogen burden (P = 0.02). BMI, percent women, IFG (in women), CRP, glucose, insulin, and waist circumference were higher and current smoking, drinking, and physical activity were lower with increasing tertiles of HOMA-IR or in metabolic syndrome (all P < 0.001). With increasing quartiles of CRP in the simple models (Table 1), there was a significant increasing trend in HOMA-IR (P < 0.0001) and in probabilities of metabolic syndrome and IFG (P < 0.0001 and 0.003, respectively). However, in the multivariate models, relations between HOMA-IR, metabolic syndrome, IFG, and CRP were not significant (P = 0.341, 0.137, and 0.379, respectively). Relationships of pathogen burden with HOMA-IR, metabolic syndrome, and IFG were not significant in the simple or multivariate models. For HCY, neither model was significant.

multivariate models, relations between HOMA-IR, metabolic syndrome, IFG, and CRP were not significant (P = 0.341, 0.137, and 0.379, respectively). Relationships of pathogen burden with HOMA-IR, metabolic syndrome, and IFG were not significant in the simple or multivariate models. For HCY, neither model was significant. CONCLUSIONS GOCADAN provides a unique setting for exploring relationships between inflammation, insulin resistance, and metabolic syndrome. Obesity is not pervasive, diabetes rates are low, rates of insulin resistance and other metabolic syndrome components vary, and chronic inflammation is prevalent in the population. Although CRP in univariate comparisons was higher in those with insulin resistance and metabolic syndrome, pathogen burden was not. After adjustment for confounders, no consistent relationships were observed between HOMA-IR, metabolic syndrome, or IFG and CRP or subclinical pathogen burden.

prevalent in the population. Although CRP in univariate comparisons was higher in those with insulin resistance and metabolic syndrome, pathogen burden was not. After adjustment for confounders, no consistent relationships were observed between HOMA-IR, metabolic syndrome, or IFG and CRP or subclinical pathogen burden. Low-grade chronic inflammation has been shown in vitro to promote insulin resistance (10). Although cross-sectional studies have demonstrated relationships between CRP and measures of insulin resistance or metabolic syndrome (11–13), longitudinal studies have varied, with CRP predictive of diabetes in some studies (14,15) but not others (3,13). In the Insulin Resistance and Atherosclerosis Study (IRAS), Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) study, Augsburg study, and Atherosclerosis Risk in Communities (ARIC) study, odds of diabetes with increasing CRP became nonsignificant after adjustment for BMI or other covariates (3,13,14), suggesting that elevations in inflammatory markers seen in insulin resistance, metabolic syndrome, and diabetes are not causative but a consequence of these abnormalities. Some (15) argue that inclusion of BMI in the models represents an overadjustment, proposing that inflammation precedes obesity; however, until a causal relationship has been established, it seems prudent to focus on the complete models.

syndrome, and diabetes are not causative but a consequence of these abnormalities. Some (15) argue that inclusion of BMI in the models represents an overadjustment, proposing that inflammation precedes obesity; however, until a causal relationship has been established, it seems prudent to focus on the complete models. The lack of a relationship between chronic inflammation and insulin resistance and/or metabolic syndrome in this population is supported by the observation that, despite chronic subclinical infection, this population has low incidence of insulin resistance and diabetes. As in other populations, diabetes is associated with the female sex, obesity, and greater insulin resistance (6); CRP is positively associated with BMI and is higher with diabetes, smoking, and increasing pathogen burden (data not shown); thus, CRP appears to be an indicator of chronic inflammation. The reasons for the high subclinical pathogen burden are unclear; they may be related to confinement in close quarters in winter and lack of adequate medical care. This study is limited by its cross-sectional design. Also, the high prevalence of subclinical infection may obscure relationships between obesity and insulin resistance on secretion of inflammatory cytokines. Finally, inflammatory mediators other than CRP that were not measured may be mediators.

The lack of a relationship between chronic inflammation and insulin resistance and/or metabolic syndrome in this population is supported by the observation that, despite chronic subclinical infection, this population has low incidence of insulin resistance and diabetes. As in other populations, diabetes is associated with the female sex, obesity, and greater insulin resistance (6); CRP is positively associated with BMI and is higher with diabetes, smoking, and increasing pathogen burden (data not shown); thus, CRP appears to be an indicator of chronic inflammation. The reasons for the high subclinical pathogen burden are unclear; they may be related to confinement in close quarters in winter and lack of adequate medical care. This study is limited by its cross-sectional design. Also, the high prevalence of subclinical infection may obscure relationships between obesity and insulin resistance on secretion of inflammatory cytokines. Finally, inflammatory mediators other than CRP that were not measured may be mediators. In summary, while analyses of unique populations often lead to unexpected findings, they can help to further understanding of complex disorders. Our study provides evidence that the inflammatory process may reflect diabetes or insulin resistance but is most likely not their cause and suggests that subclinical infection should be considered in further explorations of the inflammatory process in people with insulin resistance or metabolic syndrome.

lex disorders. Our study provides evidence that the inflammatory process may reflect diabetes or insulin resistance but is most likely not their cause and suggests that subclinical infection should be considered in further explorations of the inflammatory process in people with insulin resistance or metabolic syndrome. This study was funded by grants RO1-HL64244, U01 HL082458, and M10RR0047-34 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland. We thank Norton Sound Health Corporation, the village leadership, and Rachel Schaperow for her editorial services.

Inhibition of dipeptidyl peptidase-4 (DPP-4) increases the concentration of glucagon-like peptide-1, an incretin hormone that stimulates glucose-dependent insulin release, suppresses glucagon production, slows gastric emptying, reduces appetite, and may promote preservation of β-cell function in patients with type 2 diabetes (1). Alogliptin is a novel, high-affinity, high-specificity DPP-4 inhibitor that produces rapid and sustained DPP-4 inhibition and significantly reduces postprandial plasma glucose concentrations in patients with type 2 diabetes (2,3). A phase-three study was conducted to evaluate the efficacy and safety of alogliptin in adults with type 2 diabetes that was inadequately controlled with diet and exercise. RESEARCH DESIGN AND METHODS Eligible patients were treatment-naïve (i.e., no current antidiabetes therapy and <7 days of therapy in the past 3 months) men and women, aged 18–80 years, with type 2 diabetes. Key inclusion criteria included A1C 7–10%, BMI 23–45 kg/m2, treatment with diet and exercise for ≥1 month, and systolic/diastolic blood pressure ≤180/≤110 mmHg. Patients received counseling on diet and exercise. Patients who completed a 4-week, single-blind run-in period with fasting plasma glucose (FPG) <275 mg/ml (15.27 mmol/l) and ≥75% compliance (by tablet count) were randomized (2:2:1) to 26 weeks of double-blind treatment with 12.5 mg alogliptin, 25 mg alogliptin, or placebo taken once daily before the first meal. Additional antidiabetes agents were prohibited.

lind run-in period with fasting plasma glucose (FPG) <275 mg/ml (15.27 mmol/l) and ≥75% compliance (by tablet count) were randomized (2:2:1) to 26 weeks of double-blind treatment with 12.5 mg alogliptin, 25 mg alogliptin, or placebo taken once daily before the first meal. Additional antidiabetes agents were prohibited. Efficacy assessments included all randomized patients who received the double-blind study drug. The primary end point was mean change from baseline in A1C at week 26. Other efficacy measures included changes in FPG, clinical response rates, incidences of marked hyperglycemia (FPG ≥200 mg/dl [11.10 mmol/l]) and hyperglycemic rescue, and changes in body weight. Exploratory end points included changes in measures of pancreatic function (fasting insulin, fasting proinsulin, and homeostasis model assessment of β-cell function) and lipid profiles. Treatment group differences for the primary end point were analyzed through ANCOVA, with treatment and geographic region as variables and baseline A1C and diabetes duration as covariates. The last observation carried forward method was used for imputing missing data; testing was two-sided at a significance level of 0.05. Continuous secondary efficacy analyses were performed as for the primary analysis, except that the baseline covariate corresponds with the tested end point. Incidence variables were compared using nonparametric, extended Mantel-Haenszel X2 tests; covariates were the same as for the primary end point.

e was significantly less with alogliptin (12.5 mg, 9.8%; 25 mg, 7.6%; P ≤ 0.001) than with placebo (29.7%). Minor reductions in weight with alogliptin were neither clinically nor statistically significant relative to placebo. Results for secondary efficacy measures and exploratory assessments are summarized in Table 1. Overall incidences of adverse events (67.4–70.3%) and proportions of patients who discontinued because of adverse events (1.5–2.3%) were similar across treatment groups. Most adverse events were mild or moderate in intensity. Serious adverse events occurred without relation to dose (12.5 mg, 3.8%; 25 mg, 0.8%; and placebo, 3.1%) and were considered unrelated to treatment. No patient died during the study. Adverse events from the most commonly observed categories occurred with similar or lower frequency in those given alogliptin versus placebo (infection, 28.0–37.6%; gastrointestinal, 12.1–14.3%). Headache occurred more frequently with alogliptin (6.8–7.5%) than with placebo (4.7%). Despite increased surveillance for skin-related adverse events, their overall incidence remained low (12.5%), albeit higher with alogliptin (12.8–15.2%) than with placebo (6.3%), mostly because of pruritic events. Two patients discontinued because of skin-related adverse events: one adverse event was considered possibly related to the study drug (25 mg, moderate subcorneal pustular dermatosis); the other was judged unrelated to the study drug (12.5 mg, moderate exacerbation of contact dermatitis). No skin lesions resembling those noted in nonclinical studies of other DPP-4 inhibitors were observed. Hypoglycemia was rare (1.5–3.0%), and no hypoglycemic event was considered an adverse event or was severe enough to require assistance. No clinically meaningful changes in laboratory test results, vital sign measurements, or electrocardiogram recordings were observed.

al studies of other DPP-4 inhibitors were observed. Hypoglycemia was rare (1.5–3.0%), and no hypoglycemic event was considered an adverse event or was severe enough to require assistance. No clinically meaningful changes in laboratory test results, vital sign measurements, or electrocardiogram recordings were observed. CONCLUSIONS Alogliptin monotherapy administered for 26 weeks to treatment-naïve patients with type 2 diabetes produced significant and clinically meaningful improvements in A1C. Glycemic improvements with alogliptin were rapid, sustained, and independent of age, race, and sex. Alogliptin treatment was well tolerated and was not associated with treatment-related serious adverse events. A low incidence of hypoglycemia occurred with alogliptin and with placebo. Weight gain is common among patients taking sulfonylureas and thiazolidinediones and may reduce treatment adherence (7,8); thus, the weight neutrality of alogliptin and other DPP-4 inhibitors may offer a therapeutic advantage (9–15). Increases in the proinsulin-to-insulin ratio with alogliptin versus placebo and a trend toward increased homeostasis model assessment of β-cell function suggest that alogliptin, similarly to other DPP-4 inhibitors (9–11,14,15), may modestly improve pancreatic function. In summary, the efficacy and safety of alogliptin monotherapy were comparable with those of other DPP-4 inhibitors (9–15). Alogliptin represents an effective treatment option whether given alone or in combination with antihyperglycemic agents from other classes.

Weight gain is common among patients taking sulfonylureas and thiazolidinediones and may reduce treatment adherence (7,8); thus, the weight neutrality of alogliptin and other DPP-4 inhibitors may offer a therapeutic advantage (9–15). Increases in the proinsulin-to-insulin ratio with alogliptin versus placebo and a trend toward increased homeostasis model assessment of β-cell function suggest that alogliptin, similarly to other DPP-4 inhibitors (9–11,14,15), may modestly improve pancreatic function. In summary, the efficacy and safety of alogliptin monotherapy were comparable with those of other DPP-4 inhibitors (9–15). Alogliptin represents an effective treatment option whether given alone or in combination with antihyperglycemic agents from other classes. Supplementary Material Online-Only Appendix Financial support for completion of this study and for data analysis was provided by Takeda Global Research & Development Center, Deerfield, Illinois. This study was presented at the 68th Annual Scientific Sessions of the American Diabetes Association, San Francisco, CA, 6–10 June 2008, and at the 44th European Association for the Study for Diabetes Annual Meeting, Rome, Italy, 7–11 September 2008. Editorial assistance with manuscript preparation was provided by Scientific Connections, Newtown, PA.

Foot ulcers are common in diabetic patients, with prevalence as high as 25% (1). These ulcers frequently become infected, and spread of infections to soft tissue and to bony structures is a major causal factor for lower-limb amputation (2). Early diagnosis and adequate treatment are essential. Because microorganisms are always present on skin wounds, diagnosis of infection must be based not on microbiological findings but on clinical criteria, as emphasized by the Infectious Diseases Society of America and the International Working Group on the Diabetic Foot (IWGDF) and more recently by the French Society for Infectious Pathology (3–5). However, because of the confounding impact of neuropathy and ischemia on local and systemic inflammatory response, diagnosing foot infection at an early stage in diabetic individuals may be difficult.

onal Working Group on the Diabetic Foot (IWGDF) and more recently by the French Society for Infectious Pathology (3–5). However, because of the confounding impact of neuropathy and ischemia on local and systemic inflammatory response, diagnosing foot infection at an early stage in diabetic individuals may be difficult. Recently, we demonstrated the value of using a miniaturized oligonucleotide array covering different genes of Staphylococcus aureus, by far the most common and virulent pathogen in diabetic foot infection (3), and we showed that the virulence gene profile of S. aureus enables us to distinguish grade 1 from grades 2–4 ulcers because the former generally displayed a very low level of virulence genes (6). One of the main limitations in that work was the limited panel of genes used. Here we analyzed the most prevalent virulence-associated genes and the in vivo virulence potential of the different S. aureus strains isolated from diabetic foot ulcers. The aim was to detect genetic markers to distinguish noninfected and infected ulcer and to predict outcome of grade 1 ulcers. RESEARCH DESIGN AND METHODS From 1 March 2004 through 31 July 2007, a prospective longitudinal study of patients with diabetic foot ulcers at Nîmes University Hospital was conducted as described previously (6). Seventy-four patients (63%), hospitalized before or during April 2006, were enrolled previously (6). This study was approved by the local ethics committee and performed in accordance with the Declaration of Helsinki as revised in 2000.

foot ulcers at Nîmes University Hospital was conducted as described previously (6). Seventy-four patients (63%), hospitalized before or during April 2006, were enrolled previously (6). This study was approved by the local ethics committee and performed in accordance with the Declaration of Helsinki as revised in 2000. Bacterial isolation After wound debridement, samples for bacterial culture were obtained by swabbing the wound base, needle aspiration, or tissue biopsies and were sent immediately to the bacteriology department. Only patients with monomicrobial cultures positive for S. aureus were included in the study. Patients with grade 1 ulcers were closely followed over a period of 6 months to confirm the wound status (infected/noninfected ulcer). If the wound healed, a microbiological specimen was obtained 1 month later. If the wound did not heal, antibiotic therapy was initiated and surgical debridement or minor amputation was performed according to the wound status; a sample for bacteriological culture was obtained before antibiotic treatment was begun and the ulcer grade was updated. Microbiological study Genus, species, and antibiotic susceptibilities were determined using the Vitek 2 card (bioMérieux, Marcy-l'Etoile, France) and interpreted according to the recommendations of the French Society for Microbiology (7). Susceptibility to methicillin was screened by agar diffusion using cefoxitin disks (Bio-Rad, Marnes-La-Coquette, France) (7).

tibiotic susceptibilities were determined using the Vitek 2 card (bioMérieux, Marcy-l'Etoile, France) and interpreted according to the recommendations of the French Society for Microbiology (7). Susceptibility to methicillin was screened by agar diffusion using cefoxitin disks (Bio-Rad, Marnes-La-Coquette, France) (7). Virulence profile of S. aureus strains To assess the virulence potential of strains, the presence of 31 among the most prevalent virulence-associated genes was evaluated by PCR as described previously (8,9): staphylococcal enterotoxins A, B, C, D, E, G, H, I, J, K, and Q (se), toxic shock syndrome toxin 1 (tst), exfoliative toxins A and B (etA and etB), Panton Valentine leukocidin (PVL) (lukS-PV_lukF-PV), LukDE leukocidin (lukE), β- and two γ-hemolysins (hlb, hlg, and hlgv), epidermal cell differentiation inhibitor (edinC), nine microbial surface components recognizing adhesive matrix molecules (MSCRAMM) (bbp, cna, ebpS, clfA, clfB, fib, fnbA, fnbB, and eno), and capsular types 5 and 8 (cap5 and cap8). The accessory gene regulator (agr) allele group was determined by multiplex PCR (10). Clonality of S. aureus strains To eliminate a bias in the distribution of virulence genes due to the presence of clonal bacteria in each grade, we used different epidemiological methods. This analysis also allowed us to compare strains isolated at admission and during follow-up to determine whether they were identical.

Clonality of S. aureus strains To eliminate a bias in the distribution of virulence genes due to the presence of clonal bacteria in each grade, we used different epidemiological methods. This analysis also allowed us to compare strains isolated at admission and during follow-up to determine whether they were identical. Macrorestriction analysis of SmaI-digested chromosomal DNA was performed by pulsed field gel electrophoresis (PFGE) with the CHEF DRII system (Bio-Rad) (11). The PFGE patterns were analyzed by GelCompar software (Applied Math, Kortrijk, Belgium) and compared by the algorithmic clustering method known as the unweighted-pair group method using arithmetic averages with the Dice coefficient of similarity. Staphylococcal chromosomal cassette (SCCmec) type was determined by PCR typing according to a simplified strategy of Kondo's typing system without determining the differences in the junkyard region (see supplemental glossary, available in an online appendix at http://dx.doi.org/10.2337/dc08-1010) (12). The spa sequence (see supplemental glossary) typing was performed according to the Ridom Staph Type standard protocol (http://www.ridom.com) and by using the Ridom SpaServer, which automatically analyzes spa repeats, assigns spa types, and clusters related spa types in a spa group (http://spa.ridom.de/index.shtml).

2). The spa sequence (see supplemental glossary) typing was performed according to the Ridom Staph Type standard protocol (http://www.ridom.com) and by using the Ridom SpaServer, which automatically analyzes spa repeats, assigns spa types, and clusters related spa types in a spa group (http://spa.ridom.de/index.shtml). Caenorhabditis elegans in vivo model C. elegans (see supplemental glossary) has been used to develop an easy model system of host-pathogen interactions to identify basic evolutionarily conserved pathways associated with microbial pathogenesis. This test is based on the capacity of pathogens ingested by C. elegans nematodes to infect and ultimately kill the worms (13). The survival of nematodes fed on different S. aureus strains is an indirect marker of their virulence potential. Fer-15 worms were infected as previously described (14). In brief, nematode growth medium plates were inoculated with 10 μl of an overnight culture of S. aureus strains and incubated at 37°C for 8 h. Between 25 and 30 C. elegans worms were transferred from a lawn of Escherichia coli OP50 to a lawn of the bacterium to be tested, incubated at 25°C, and examined at 24-h intervals with a stereomicroscope (Leica MS5) for viability. Nematodes were classified as dead if they failed to respond to touch. All experiments were conducted in triplicate and repeated at least five times for each selected strain. S. aureus virulence was assessed using the nematode survival curve and calculating the LT50 and LT100 (the times required to kill 50 and 100% of the worms, respectively).

d as dead if they failed to respond to touch. All experiments were conducted in triplicate and repeated at least five times for each selected strain. S. aureus virulence was assessed using the nematode survival curve and calculating the LT50 and LT100 (the times required to kill 50 and 100% of the worms, respectively). Statistical analysis For each qualitative variable (virulence genotyping), comparison between ulcer grades was assessed using Fisher's exact test. The ability to diagnose infection of a wound was expressed by sensitivity, specificity, and positive and negative predictive values; area under the receiver operating characteristic (AUCROC) curve was calculated by the nonparametric Hanley method. To assess the utility of combining several virulence markers, we used a logistic regression with a backward procedure to select the most relevant markers; only markers for which AUCROC was >0.80 were initially entered as explanatory variables in the regression analysis. An ROC curve was then generated for the combination derived from the regression model, and its area was compared with that of every single virulence marker by a nonparametric method adapted to paired data (15). To compare overall survival curves in the nematode killing assay, a Cox regression was used. Statistical analysis was performed using S-Plus 2000 software (Insightful, Seattle, WA), and results were considered significant for P < 0.05.

every single virulence marker by a nonparametric method adapted to paired data (15). To compare overall survival curves in the nematode killing assay, a Cox regression was used. Statistical analysis was performed using S-Plus 2000 software (Insightful, Seattle, WA), and results were considered significant for P < 0.05. RESULTS Clinical and bacteriological data From 513 selected patients, 118 were included because they had been free of any antibiotic treatment for at least 6 months, and S. aureus was the single organism isolated from the bacterial culture of their wound (Fig. 1). In 69, the current wound was the first episode of ulceration, whereas in 49 it was a recurrence. The characteristics of the study population are shown in Fig. 1 and Table 1. Of the wounds, 24 (20%) were classified as grade 1 and were followed for 6 months. Of the 118 S. aureus strains, 48 (41%) were methicillin-resistant (MRSA). During the follow-up period, 9 grade 1 ulcers healed (38%), whereas 15 worsened. In two healing ulcers, samples remained positive for S. aureus compared with positive results for 12 of the 15 nonhealed ulcers. In total, 132 S. aureus strains were isolated (118 on the initial cultures and 14 during the follow-up period) (Fig. 1). All recurrent ulcers were positive for MRSA, whereas methicillin-susceptible S. aureus (MSSA) was isolated in all but one of the ulcers appearing as the first episode. A higher prevalence of MRSA with increasing severity of infection was also noted, with a statically significant difference for grade 1 compared with grades 2–4 (P = 0.021).

were positive for MRSA, whereas methicillin-susceptible S. aureus (MSSA) was isolated in all but one of the ulcers appearing as the first episode. A higher prevalence of MRSA with increasing severity of infection was also noted, with a statically significant difference for grade 1 compared with grades 2–4 (P = 0.021). Virulence profile Virulence genotyping of the 132 strains evaluated by PCR is shown in Table 2. Individual gene analysis showed that the prevalence rates of 10 genes (sea, seb, sec, sei, sej, hlb, hlg, hlgv, cap5, and lukE) were significantly more often associated with strains isolated from grades 2–4 ulcers, whereas the cap8 gene was most frequently found in strains from grade 1 ulcers (P < 0.05). From the logistic model analysis, a five-gene combination of sea, sei, lukE, hlgv, and cap8 was the most predictive for differentiating grade 1 from grades 2–4 ulcers: mean ± SD AUCROC was 0.940 ± 0.028 (95% CI 0.885–0.995), which is significantly greater than that from combining all of the 30 virulence genes (0.810 ± 0.078) (P < 0.05); sensitivity was 0.977 ± 0.025, specificity was 0.871 ± 0.063, and positive and negative predictive values were 0.884 and 0.975, respectively. Using the logistic regression equation, three grade 1 ulcers were misclassified owing to absence of cap8 but presence of sea, sei, lukE, and hlgv genes; interestingly, these ulcers rapidly worsened. Conversely, 17 grades 2–4 ulcers were misclassified owing to absence of sea, sei, lukE, and hlgv genes.

75, respectively. Using the logistic regression equation, three grade 1 ulcers were misclassified owing to absence of cap8 but presence of sea, sei, lukE, and hlgv genes; interestingly, these ulcers rapidly worsened. Conversely, 17 grades 2–4 ulcers were misclassified owing to absence of sea, sei, lukE, and hlgv genes. Clonality study By using PFGE, a wide genomic diversity was shown among the 132 S. aureus isolates (data not shown), allowing us to exclude a bias in the statistical analysis. PFGE also revealed the spread of three clonal MRSA groups (44 strains; 33% of the isolates). A major group, clustering 29 strains (66% of MRSA strains) matched the Lyon clone (agr allele 1, spa type t008, and SCCmec type IV). The first minor clone (eight strains, 16% of MRSA) was the “classic pediatric” clone (agr allele 2, spa type related to t311, and SCCmec type IV). The second minor clone (three strains, 6% of MRSA) was the “new pediatric” clone (agr2 allele, spa type t777, and SCCmec type VI). Comparison between strains isolated from patients’ ulcers at admission and follow-up demonstrated that those isolated from healing and slowly worsening ulcers had no clonal link. On the other hand, the strains from the three rapidly worsening ulcers were similar in each ulcer (data not shown). S. aureus–mediated killing of C. elegans When feeding on E. coli OP50, C. elegans has a 2-week life span (LT100 of worms varying between 11 and 14 days and LT50 between 3 and 5 days). When feeding on a pathogen, worms die far more rapidly, with a life span between 3 and 7 days (LT50 between 1 and 2 days).

Comparison between strains isolated from patients’ ulcers at admission and follow-up demonstrated that those isolated from healing and slowly worsening ulcers had no clonal link. On the other hand, the strains from the three rapidly worsening ulcers were similar in each ulcer (data not shown). S. aureus–mediated killing of C. elegans When feeding on E. coli OP50, C. elegans has a 2-week life span (LT100 of worms varying between 11 and 14 days and LT50 between 3 and 5 days). When feeding on a pathogen, worms die far more rapidly, with a life span between 3 and 7 days (LT50 between 1 and 2 days). Three strains (two MSSA and one MRSA) chosen at random among those isolated at admission in each of the four ulcer grades were tested for their capacity to kill C. elegans. In addition, strains initially isolated from grade 1 ulcers were compared with those isolated at the follow-up: strains from three ulcers with different outcomes were chosen: one that healed, one that worsened rapidly, and one that degraded slowly (Fig. 2). According to their killing ability, two populations of bacteria can be observed: the first with LT50 <2 days, suggesting a high virulence potential, and the second with LT50 >3 days, suggesting lower virulence (P < 0.001). All of the strains with LT50 <2 days were isolated from ulcers grades 2–4 except for one (NSA22465); conversely, all of the strains but one with LT50 >3 days were isolated from grade 1 ulcers. Interestingly, no significant difference in the killing potentials of MRSA and MSSA was observed within ulcers of the same grade: both strains from grade 1 had a similarly long LT50 (3.5 ± 0.3 days), and MRSA and MSSA from ulcers grades 2–4 had a similarly short LT50 (1.7 ± 0.2 days). Finally, both at entry and at follow-up, the LT50 for strains isolated from healing ulcers was relatively long (3.7 vs. 3.6 days, NS), whereas it was short for the strains from rapidly worsening ulcers (1.67 vs. 1.71 days, NS). For the strain from a slowly worsening ulcer, the LT50 was significantly shorter at follow-up than at admission (1.8 vs. 3.3 days, P < 0.001).

for strains isolated from healing ulcers was relatively long (3.7 vs. 3.6 days, NS), whereas it was short for the strains from rapidly worsening ulcers (1.67 vs. 1.71 days, NS). For the strain from a slowly worsening ulcer, the LT50 was significantly shorter at follow-up than at admission (1.8 vs. 3.3 days, P < 0.001). CONCLUSIONS This study demonstrated for the first time the existence of two populations of S. aureus strains in diabetic foot ulcers: strains isolated from noninfected ulcers with a low virulence potential (as shown by the in vivo nematode model results) as opposed to strains isolated from infected ulcers with a high virulence potential. Moreover, the presence or absence of five virulence genes separated the two populations and allowed us to distinguish noninfected from infected wounds.

ers with a low virulence potential (as shown by the in vivo nematode model results) as opposed to strains isolated from infected ulcers with a high virulence potential. Moreover, the presence or absence of five virulence genes separated the two populations and allowed us to distinguish noninfected from infected wounds. The fact that determining five virulence genes may help to differentiate noninfected from infected wound is an attractive result. Among infection-associated genes, four corresponded to MRSA markers (sea, sei, lukE, and hlgv) and hlgv, sea, and lukE-lukD genes are shared in the Lyon clone. Moreover, hospital-acquired MRSA strains shared the enterotoxin gene cluster locus, notably sei (16). sea is the most studied and interesting gene; its product has a strong proinflammatory effect (17). In our study, this gene was not exclusively related to the virulence of MRSA infection, as it was also detected in MSSA strains. Recently, an innate immune evasion cluster located on β-hemolysin–converting bacteriophages and carrying sea was discovered. It is easily transferred among strains (18). This potential of transfer via bacteriophage could explain our results and, notably, the presence of the sea gene in MSSA strains. The higher prevalence of sea, sei, lukE, and hlgv genes among the strains isolated from grades 2–4 compared with grade 1 ulcers and the absence of differences in MRSA clones between grade 1 and grades 2–4 ulcers suggest that these markers are really interesting, as they actually could be virulence markers. Finally, a number of studies demonstrated that the noninfected ulcer marker, capsular polysaccharide T8 (Cap8) was strongly associated with MSSA strains, as suggested by our study (19). However, its role in virulence has not been clearly defined. Production of Cap8 appears to be regulated by various environmental cues, and its overproduction might be implicated in S. aureus virulence (20,21).

, capsular polysaccharide T8 (Cap8) was strongly associated with MSSA strains, as suggested by our study (19). However, its role in virulence has not been clearly defined. Production of Cap8 appears to be regulated by various environmental cues, and its overproduction might be implicated in S. aureus virulence (20,21). Another interesting point is the low level of clonal strains isolated in this study (33%). Among MRSA strains, results indicated that a major clone matching the Lyon clone (66%) was widely distributed. Our study is in accordance with a recent report showing that this clone is currently the most prevalent MRSA clone in France (16). Interestingly, although dissemination of PVL-producing clones has been extensively reported and discussed since the beginning of the new millennium, these strains were not detected in our study. The use of the C. elegans model demonstrated that S. aureus virulence was not dependent on methicillin resistance as suggested previously (22,23). This result is interesting because nearly half of the S. aureus strains isolated were methicillin resistant. We can speculate that within the same bacterial species there are pathogens with different virulence potential against the host. These bacterial populations with variable virulence represent a new challenge in terms of pathogenicity, treatment, and prevention of transmission.

trains isolated were methicillin resistant. We can speculate that within the same bacterial species there are pathogens with different virulence potential against the host. These bacterial populations with variable virulence represent a new challenge in terms of pathogenicity, treatment, and prevention of transmission. Our study suggests that testing for the presence of five genes may not only help clinicians to distinguish grade 1 from grades 2–4 ulcers but will also predict wound outcome. At follow-up, S. aureus was isolated from 13 grade 1 ulcers. cap8 was detected in S. aureus from the two healing ulcers, but the strains had pulsotypes and genotypes different from those of the baseline sample. Three of the 15 worsening ulcers (corresponding to the three false-positive results) rapidly degraded; both pulsotype and virulence profiles were found to be unchanged, suggesting that the isolates were the same and the wound was actually infected at baseline and not simply colonized. Finally, S. aureus from slowly worsening recalcitrant ulcers harbored virulence markers that were absent at baseline: pulsotypes and genotypes were different in every case, suggesting that new, more virulent S. aureus strains had colonized the ulcer (Fig. 1).

was actually infected at baseline and not simply colonized. Finally, S. aureus from slowly worsening recalcitrant ulcers harbored virulence markers that were absent at baseline: pulsotypes and genotypes were different in every case, suggesting that new, more virulent S. aureus strains had colonized the ulcer (Fig. 1). In summary, the increasing prevalence of resistant staphylococci and the small number of new antimicrobial drugs must stimulate the discovery of new solutions for diabetic foot infections in the near future. Testing for the presence of five genes will be a useful tool in management of diabetic foot ulcers. One-step multiplex PCR assays are relatively easy and rapid to perform (2 h after obtaining a specimen) at low cost (∼5 USD). This testing will allow early discrimination between noninfected grade 1 and infected grades 2–4 diabetic foot ulcers, such that antibiotic treatment is prescribed for those most likely to benefit. Supplementary Material Online-Only Appendix This work was supported by the Coloplast Foundation, the French Speaking Association for Diabetes and Metabolic Diseases (ALFEDIAM-Aventis grant), the Foundation for Medical Research of Languedoc-Roussillon-Rouergue the Institut National de la Santé et de la Recherche Médicale, la Region Languedoc Roussillon, and the Montpellier 1 University. Fer-15 nematodes were provided by the Caenorhabditis Genetics Center, a foundation of the National Institutes of Health National Center for Research Resources. We thank the team of the Department of Diabetology for help in recruiting patients.

The importance of insulin and glucagon in maintaining postprandial glycemic excursions within a narrow range is well established (1). However, alterations in gastric emptying, another potentially important factor (2), are not generally considered to be of clinical significance for postprandial hyperglycemia in diabetes unless diabetes late complications, such as gastroparesis, have emerged (3,4). Gastroparesis is a relatively rare diabetes late complication resulting from irreversible intestinal nerve damage (5) and has to be distinguished from the physiological inhibitory effects of acute hyperglycemia on gastric motility (6,7). The latter has been proposed as a defense mechanism to minimize postprandial hyperglycemia by reducing the rate of efflux of glucose into the circulation from the gut (8). This process may be of special importance for patients with type 1 diabetes because they have been reported to have a reduced ability to delay gastric emptying in response to hyperglycemia (9).

hanism to minimize postprandial hyperglycemia by reducing the rate of efflux of glucose into the circulation from the gut (8). This process may be of special importance for patients with type 1 diabetes because they have been reported to have a reduced ability to delay gastric emptying in response to hyperglycemia (9). The pancreatic β-cell hormone islet amyloid polypeptide (IAPP) is cosecreted with insulin in a fixed molar ratio (10) and reduces gastric emptying. Thus, patients with type 1 diabetes even without concomitant enteric neuropathy should have increased rather than delayed rates of gastric emptying, because they are IAPP deficient (11). Accordingly, the present studies were undertaken to test the hypothesis that impairment in hyperglycemia-induced delay in gastric emptying should result in greater meal-derived glucose appearance in the systemic circulation and thus should contribute to postprandial hyperglycemia in patients with type 1 diabetes.

(11). Accordingly, the present studies were undertaken to test the hypothesis that impairment in hyperglycemia-induced delay in gastric emptying should result in greater meal-derived glucose appearance in the systemic circulation and thus should contribute to postprandial hyperglycemia in patients with type 1 diabetes. RESEARCH DESIGN AND METHODS Written informed consent was obtained from 10 healthy subjects and 15 subjects with type 1 diabetes after the Ludwig Maximilians University of Munich Institutional Review Board had approved the protocol. Healthy subjects (seven men and three women, 39 ± 4 years of age, body weight 80 ± 4 kg) had normal routine laboratory blood test results as well as no family history of diabetes and had normal glucose tolerance (assessed by oral glucose tolerance tests according to World Health Organization criteria [12]). Healthy subjects and type 1 diabetic subjects (eight men and seven women, 37 ± 2 years of age, body weight 76 ± 3 kg) had normal physical examinations and no gastrointestinal symptoms. No type 1 diabetic subject had symptoms of cardiac autonomic or peripheral sensory neuropathy assessed by electrocardiogram breath-dependent variability and clinical examination of peripheral sensory function. None had symptoms or a history of gastrointestinal neuropathy (i.e., bloating, vomiting, constipation, diarrhea, or redundant postprandial hypoglycemia) or evidence for nephropathy, assessed by microalbumin excretion. Retinal fundus photography showed only mild background retinopathy in three type 1 diabetic subjects. All type 1 diabetic subjects had been receiving continuous subcutaneous insulin infusion therapy for at least 3 months before the study with good glycemic control as assessed by A1C (7.3 ± 0.2%) and seven-point self-obtained blood glucose profiles without severe hypoglycemic episodes. Three days before the study, all subjects had been consuming a weight-maintaining diet containing at least 200 g of carbohydrates and had abstained from alcohol, smoking, and exercise.

emic control as assessed by A1C (7.3 ± 0.2%) and seven-point self-obtained blood glucose profiles without severe hypoglycemic episodes. Three days before the study, all subjects had been consuming a weight-maintaining diet containing at least 200 g of carbohydrates and had abstained from alcohol, smoking, and exercise. Protocol Healthy subjects were studied on two occasions (euglycemia and hyperglycemia), and type 1 diabetic subjects were studied on three occasions (euglycemia and hyperglycemia with and without 30 μg pramlintide [Amylin Pharmaceuticals, San Diego, CA], injected subcutaneously in the lower abdominal wall with the meal, which was designed to replace absent IAPP secretion as indicated by its effects on gastric emptying compared with hyperglycemia in healthy subjects estimated from previous studies [13,14]). Healthy subjects received only placebo injections. Experiments were separated by at least 1 day and undertaken in a randomized, single-blinded order. Subjects refrained from food intake at least 10 h before admission to the clinical research unit between 7:00 and 7:30 a.m. on the study day. A dorsal hand vein was cannulated, and temperature was maintained at 40°C with a thermoregulated lamp for arterialized venous blood sampling (8). On the nights before the study days, type 1 diabetic subjects were instructed to measure plasma glucose concentrations at 10 p.m. and 2 a.m. to adjust basal insulin infusion with the aim of average plasma glucose concentrations of ∼5 or 10 mmol/l, respectively (15). After admission to the clinical research unit, the continuous subcutaneous insulin infusion was discontinued, and an intravenous insulin infusion was adjusted accordingly to maintain preprandial plasma glucose concentrations at ∼5 and ∼11 mmol/l, respectively.

ucose concentrations of ∼5 or 10 mmol/l, respectively (15). After admission to the clinical research unit, the continuous subcutaneous insulin infusion was discontinued, and an intravenous insulin infusion was adjusted accordingly to maintain preprandial plasma glucose concentrations at ∼5 and ∼11 mmol/l, respectively. Postprandial intravenous insulin infusion rates were based on the individual bolus requirements from the subcutaneous pump therapy. Postprandial insulin infusion rates of euglycemic and hyperglycemic experiments were 6.4 ± 0.9, 6.8 ± 0.8, 3.5 ± 0.8, 1.4 ± 0.3, and 0.9 ± 0.1 islet equivalents/h at 0–30, 35–60, 65–120, 125–240, and 240–330 min, respectively.

ion rates were based on the individual bolus requirements from the subcutaneous pump therapy. Postprandial insulin infusion rates of euglycemic and hyperglycemic experiments were 6.4 ± 0.9, 6.8 ± 0.8, 3.5 ± 0.8, 1.4 ± 0.3, and 0.9 ± 0.1 islet equivalents/h at 0–30, 35–60, 65–120, 125–240, and 240–330 min, respectively. At 8 a.m., a primed (25 μmol) continuous (∼0.25 μmol · kg−1 · min−1) infusion of [1-13C] glucose was started via a forearm vein for measurements of plasma glucose turnover in all except the hyperglycemic experiments of healthy subjects. At least 3 h were allowed for achievement of isotope equilibration. Before meal ingestion, three baseline blood samples were collected for fasting glucose, insulin, glucagon, and IAPP concentrations and [1-13C]glucose enrichments. Thereafter, subjects ingested a standardized meal within 5 min. The meal (450 kcal, ∼45% carbohydrates, ∼30% fat, and ∼25% protein) consisted of three scrambled eggs and 100 ml jelly containing 50 g of glucose enriched with 3 g 6,6-dideuteroglucose. Consumption of 200 ml sparkling water was allowed with the meal. Subjects remained in a semisupine position throughout the study period. Over the initial 90 min of the postprandial period, blood samples were taken at 15-min intervals and thereafter at 30-min intervals until completion of the experiment at 330 min except for the hyperglycemic experiments in healthy subjects, which were terminated at 240 min. The scrambled eggs and jelly were additionally labeled with 99mTc-Sn-colloid (∼70 MBq) for measurements of gastric emptying by high-resolution scintigraphy (20 images/min; Orbiter, Siemens, Erlangen, Germany), starting immediately after food ingestion for the remainder of the experiment.

ich were terminated at 240 min. The scrambled eggs and jelly were additionally labeled with 99mTc-Sn-colloid (∼70 MBq) for measurements of gastric emptying by high-resolution scintigraphy (20 images/min; Orbiter, Siemens, Erlangen, Germany), starting immediately after food ingestion for the remainder of the experiment. After the hyperglycemic experiments in type 1 diabetic subjects had been completed, we performed hyperglycemic experiments in healthy subjects isoglycemic to those of the hyperglycemic placebo experiments of type 1 diabetic subjects using a glucose infusion algorithm as described previously (16,17). Fasting plasma glucose was clamped at 11.03 ± 0.1 mmol/l, not significantly different from preprandial plasma glucose concentrations in type 1 diabetic subjects. Glucose infusion rates in hyperglycemic isoglycemic experiments in healthy subjects were started 2.5 h before meal ingestion. No insulin was infused in healthy subjects. Mean glucose infusion rates before meal ingestion were 49.1 ± 8.4 μmol · kg−1 · min−1 after an intravenous bolus of 16.2 ± 2.2 g glucose. Postprandial rates of glucose infusion in healthy subjects were adjusted to produce postprandial plasma glucose excursions not significantly different from those in hyperglycemic experiments in type 1 diabetic subjects. Mean postprandial rates of glucose infusion were 96.5 ± 9.8 μmol · kg−1 · min−1. The hyperglycemic experiments in healthy subjects were terminated 4 h after meal ingestion. In type 1 diabetic subjects, no glucose was infused in the postprandial state. Healthy subjects served in part as a control group in studies in which effects of pramlintide administration on postprandial glucose fluxes in 14 healthy subjects are reported (8).

n healthy subjects were terminated 4 h after meal ingestion. In type 1 diabetic subjects, no glucose was infused in the postprandial state. Healthy subjects served in part as a control group in studies in which effects of pramlintide administration on postprandial glucose fluxes in 14 healthy subjects are reported (8). Analytical procedures Gastric outlines on the maximum-intensity image were defined as the region of interest. Completion of gastric emptying was assumed at a reduction of the initial activity to <5%. Loss of activity was corrected for the radioactive half-life of 99mTc (8). Blood samples were collected for plasma glucose concentrations and [1-13C]- and [6,6-2H2]glucose enrichments in oxalate-fluoride tubes and for plasma insulin, glucagon, and IAPP concentrations in EDTA tubes containing a protease inhibitor. Samples were immediately placed on ice, and plasma was separated within 30 min by centrifugation at 4°C. Plasma glucose concentrations and [1-13C]- and [6,6-2H2]glucose enrichments were measured as described previously (8). Plasma insulin and glucagon concentrations were determined by standard radioimmunoassay (8), and IAPP concentrations were determined using an enzyme-linked immunosorbent assay (Linco Research).

at 4°C. Plasma glucose concentrations and [1-13C]- and [6,6-2H2]glucose enrichments were measured as described previously (8). Plasma insulin and glucagon concentrations were determined by standard radioimmunoassay (8), and IAPP concentrations were determined using an enzyme-linked immunosorbent assay (Linco Research). Calculations Systemic release and uptake of glucose were calculated with steady-state equations before meal ingestion and subsequently with the non–steady-state equations of DeBodo et al. (8) using a pool fraction of 0.65 and a volume of distribution of 200 ml/kg. Rates of appearance of the oral glucose load in the systemic circulation were calculated from [6,6-2H2]glucose enrichments using the equation of Chiasson et al. (8). The endogenous glucose release was calculated as the difference between the overall rate of plasma glucose appearance and the rate of appearance of exogenous glucose (8). Splanchnic glucose disposal of the ingested glucose load was calculated as the difference between the amount of glucose emptied by the stomach and the amount of glucose that appeared in the systemic circulation at the end of the experiments.

plasma glucose appearance and the rate of appearance of exogenous glucose (8). Splanchnic glucose disposal of the ingested glucose load was calculated as the difference between the amount of glucose emptied by the stomach and the amount of glucose that appeared in the systemic circulation at the end of the experiments. Statistical analysis Data are means ± SEM unless otherwise specified using Statistica statistical software (1998 edition; Statsoft, Tulsa, OK). Normality of the distribution was assessed using the Kolmogorov-Smirnov test. Comparisons between the groups and baseline with postprandial values were performed using ANOVA for exclusion of carryover effects followed by post hoc comparison with paired and unpaired t tests within and between patient groups, respectively. P < 0.05 was considered statistically significant. Correlation between variables was performed using Spearman's regression analysis.

dial values were performed using ANOVA for exclusion of carryover effects followed by post hoc comparison with paired and unpaired t tests within and between patient groups, respectively. P < 0.05 was considered statistically significant. Correlation between variables was performed using Spearman's regression analysis. RESULTS Gastric retention, lag period, 50% retention time, 60-min retention Initial rates of gastric emptying during euglycemia were greater in type 1 diabetic subjects, with significantly greater rates between 90 and 120 min (all P < 0.03). However, lag periods, 60-min retention, and 50% retention times (T50) were not statistically different in healthy subjects and type 1 diabetic subjects (all P > 0.3) (Fig. 1). Hyperglycemia markedly slowed gastric emptying in healthy subjects with greater lag periods, 60-min retention, and T50 (all P < 0.001), with the most pronounced effects occurring within the first 60 min after meal ingestion. In contrast, hyperglycemia had no effects on these parameters in type 1 diabetic subjects. Pramlintide administration markedly delayed gastric emptying in type 1 diabetic subjects (lag period, 60-min retention, and T50, all P < 0.001) compared with placebo, with the most pronounced effects occurring within the initial 60 min after meal ingestion. As a consequence, lag period, 60-min retention time, and T50 were restored to values comparable to those observed in hyperglycemic experiments in healthy subjects (P = 0.29, 0.06, and 0.29, respectively).

) compared with placebo, with the most pronounced effects occurring within the initial 60 min after meal ingestion. As a consequence, lag period, 60-min retention time, and T50 were restored to values comparable to those observed in hyperglycemic experiments in healthy subjects (P = 0.29, 0.06, and 0.29, respectively). Plasma glucose and correlation between peak plasma glucose at 60 min and gastric content at 45 min Fasting plasma glucose concentrations in the euglycemic experiments were slightly but significantly greater in type 1 diabetic subjects (area under the curve [AUC] −60 to 0 min: 265 ± 7 vs. 334 ± 13 mmol · l−1 · min−1, P < 0.001); however, after meal ingestion plasma glucose increased comparably in type 1 diabetic and healthy subjects (increase in AUC 0 to 60 min: 131 ± 15 vs. 162 ± 20 mmol · l−1 · min−1, P = 0.45) (Fig. 2). In the hyperglycemic experiments, fasting and postprandial glucose concentrations did not differ between healthy subjects and type 1 diabetic subjects in the placebo experiments (AUC −60 to 0 min and 0–330 min: 665 ± 4 vs. 686 ± 9 and 2,959 ± 45 vs. 2,915 ± 80 mmol · l−1 · min−1, P = 0.08 and 0.69, respectively).

. 2). In the hyperglycemic experiments, fasting and postprandial glucose concentrations did not differ between healthy subjects and type 1 diabetic subjects in the placebo experiments (AUC −60 to 0 min and 0–330 min: 665 ± 4 vs. 686 ± 9 and 2,959 ± 45 vs. 2,915 ± 80 mmol · l−1 · min−1, P = 0.08 and 0.69, respectively). Pramlintide administration completely prevented any increase in postprandial plasma glucose above baseline during the entire postprandial period so that mean plasma glucose concentrations were reduced by 3 mmol/l (P < 0.001). The AUCs for 0–330 min were 3,739 ± 123 vs. 2,995 ± 86 mmol · l−1 · min−1(P < 0.001). Changes in postprandial plasma glucose concentrations at 60 min correlated with gastric content at 45 min (r = 0.52, P = 0.001) (Fig. 1). Plasma insulin, glucagon, and IAPP Plasma insulin concentrations paralleled those of plasma glucose concentrations in healthy subjects and were on average ∼11-fold greater in the hyperglycemic than in the euglycemic experiments. AUCs over the entire experimental period were 52,838 ± 3,975 versus 606,929 ± 135,590 pmol · l−1 · min−1 (P < 0.001). In type 1 diabetic subjects, plasma insulin concentrations were comparable in all experiments (AUCs over the entire experimental period: 114,146 ± 9,344, 113,551 ± 10,157, and 114,778 ± 7,514 pmol · l−1 · min−1, all P > 0.70) but on average approximately twofold greater than those for healthy subjects (P < 0.001) (Fig. 2).

ype 1 diabetic subjects, plasma insulin concentrations were comparable in all experiments (AUCs over the entire experimental period: 114,146 ± 9,344, 113,551 ± 10,157, and 114,778 ± 7,514 pmol · l−1 · min−1, all P > 0.70) but on average approximately twofold greater than those for healthy subjects (P < 0.001) (Fig. 2). Plasma glucagon concentrations decreased significantly after meal ingestion in healthy subjects (AUC −60 to 0 min vs. 0–60 min: 3,574 ± 279 vs. 3,283 ± 294 pg · ml−1 · min−1, P = 0.01) but not in type 1 diabetic subjects (AUC −60 to 0 min vs. 0–60 min: 2,433 ± 196 vs. 2,597 ± 203 pg · ml−1 · min−1, P = 0.16) (Fig. 2). When pramlintide was given in type 1 diabetic subjects, plasma glucagon concentrations decreased to a nadir at 60 min and were significantly lower within the first 90 min (AUC 0–90 min: 3,829 ± 280 vs. 3,085 ± 289 pg · ml−1 · min−1, placebo versus pramlintide, P = 0.01). In euglycemic experiments in healthy subjects, plasma IAPP concentrations increased from 2.1 ± 0.6 pmol/l to peak values of 12.4 ± 1.3 pmol/l at 90 min, averaging 8.8 ± 0.9 pmol/l postprandially. In the hyperglycemic experiments, plasma IAPP concentrations were significantly greater before meal ingestion (on average 18.9 ± 4.2 pmol/l, AUC −30 to 0 min: 63 ± 17 vs. 566 ± 126 pmol · l−1 · min−1, P < 0.001) and increased to an average of 42.8 ± 7.3 pmol/l (AUC 0−240 min: 2,104 ± 206 vs. 10261 ± 1,746 pmol · l−1 · min−1, P < 0.0001) (Fig. 2). In type 1 diabetic subjects, plasma IAPP was undetectable under euglycemic and hyperglycemic conditions both pre- and postprandially.

± 17 vs. 566 ± 126 pmol · l−1 · min−1, P < 0.001) and increased to an average of 42.8 ± 7.3 pmol/l (AUC 0−240 min: 2,104 ± 206 vs. 10261 ± 1,746 pmol · l−1 · min−1, P < 0.0001) (Fig. 2). In type 1 diabetic subjects, plasma IAPP was undetectable under euglycemic and hyperglycemic conditions both pre- and postprandially. Rates of exogenous (meal) and endogenous plasma glucose appearance In type 1 diabetic subjects, preprandial endogenous glucose production was significantly greater in both the euglycemic and hyperglycemic experiments than that in healthy subjects (AUC −60 to 0 min: 545 ± 21 μmol/kg in healthy subjects vs. 788 ± 41, 963 ± 38, and 911 ± 65 μmol/kg, respectively, in the experiments with type 1 diabetic subjects, all P < 0.001). In healthy subjects, postprandial endogenous glucose production was suppressed by 60.6 ± 1.5%. In type 1 diabetic subjects, the degree of suppression was comparable in the euglycemic and hyperglycemic experiments (45.4 ± 2.5 and 45.1 ± 1.6%, respectively, P > 0.50) and lower than that in healthy subjects (both P < 0.001). Postprandial endogenous glucose production did not differ after placebo and pramlintide administration in type 1 diabetic subjects (AUC 0–330 min: 2,991 ± 154 vs. 2,773 ± 199 μmol/kg, P > 0.21) (Fig. 2).

ments (45.4 ± 2.5 and 45.1 ± 1.6%, respectively, P > 0.50) and lower than that in healthy subjects (both P < 0.001). Postprandial endogenous glucose production did not differ after placebo and pramlintide administration in type 1 diabetic subjects (AUC 0–330 min: 2,991 ± 154 vs. 2,773 ± 199 μmol/kg, P > 0.21) (Fig. 2). Rates of exogenous glucose appearance (meal-derived glucose) increased from 30 to 90 min in type 1 diabetic subjects compared with those of healthy subjects in the euglycemic placebo experiments (AUC 30–90 min: 743 ± 47 vs. 1,012 ± 67 μmol/kg, P < 0.005) and were unaffected by hyperglycemia (AUC 30–90 min in the hyperglycemic placebo experiment: 1,054 ± 50 μmol/kg, P > 0.80 compared with placebo euglycemia in type 1 diabetes). Administration of pramlintide significantly reduced rates of exogenous glucose appearance in type 1 diabetic subjects from 15 to 90 min (AUC 15–90 min: 1,220 ± 58 vs. 269 ± 47 μmol/kg, P < 0.001) with slightly but significantly greater rates for the remainder of the experiment. Overall, they were not significantly different from those of healthy subjects during euglycemia (AUC 0–330 min in healthy subjects and type 1 diabetic subjects after pramlintide administration, respectively: 2,750 ± 144 vs. 2,251 ± 241 μmol/kg, P = 0.145) (Fig. 2).

tly greater rates for the remainder of the experiment. Overall, they were not significantly different from those of healthy subjects during euglycemia (AUC 0–330 min in healthy subjects and type 1 diabetic subjects after pramlintide administration, respectively: 2,750 ± 144 vs. 2,251 ± 241 μmol/kg, P = 0.145) (Fig. 2). CONCLUSIONS Major findings of these studies are that the physiological defense mechanism to delay gastric emptying in response to postprandial hyperglycemia is impaired in patients with type 1 diabetes. Moreover, our studies demonstrate that this impairment leads to exaggerated rates of meal-derived glucose appearance in plasma and thus contributes to postprandial hyperglycemia. IAPP increased markedly in healthy subjects and was associated with a profound delay in gastric emptying. A delay in gastric emptying in type 1 diabetic patients comparable to that found in healthy subjects markedly improved postprandial glucose excursions in type 1 diabetic patients. Delayed gastric emptying decreased profoundly initial appearance of meal-derived glucose in the systemic circulation in conjunction with increased splanchnic glucose sequestration.

abetic patients comparable to that found in healthy subjects markedly improved postprandial glucose excursions in type 1 diabetic patients. Delayed gastric emptying decreased profoundly initial appearance of meal-derived glucose in the systemic circulation in conjunction with increased splanchnic glucose sequestration. These findings appear at first glance to contradict those of previously published studies reporting that physiological hyperglycemia delays gastric emptying in patients with type 1 diabetes (9). As in our studies, gastric emptying was not found to be delayed in patients with type 1 diabetes under euglycemic conditions, but a blunted responsiveness to hyperglycemia was reported with a >50% reduction in the delay of gastric emptying in patients with type 1 diabetes. These findings and the results of the present studies support the concept that the physiological defense mechanism to delay gastric emptying in response to postprandial hyperglycemia is impaired in patients with type 1 diabetes. The complete unresponsivness to hyperglycemia in our studies may be explained by different study designs. In our studies typical postprandial hyperglycemic glucose fluctuations were allowed, whereas Schvarcz et al. (9) used the continuous hyperglycemic clamp technique.

emia is impaired in patients with type 1 diabetes. The complete unresponsivness to hyperglycemia in our studies may be explained by different study designs. In our studies typical postprandial hyperglycemic glucose fluctuations were allowed, whereas Schvarcz et al. (9) used the continuous hyperglycemic clamp technique. Under euglycemic conditions, plasma IAPP concentrations paralleled those of insulin concentrations in healthy subjects. When exposed to hyperglycemia, IAPP concentrations increased from ∼9 to 43 pmol/l postprandially. No apparent IAPP secretion could be detected either under euglycemic or hyperglycemic conditions in patients with type 1 diabetes. Pramlintide administration in type 1 diabetic patients delayed gastric emptying to an extent comparable to that with hyperglycemia in healthy subjects. This result, however, does not necessarily imply concentrations comparable to those of endogenous IAPP found in healthy subjects but rather a comparable pharmacodynamic effect on gastric emptying. Taking this into consideration, IAPP deficiency may be seen as one cause for the unresponsiveness in delaying gastric emptying. To prove that IAPP deficiency was the sole reason for the lack of delay in gastric emptying, direct inhibition of the IAPP effects using a specific antagonist would be required. Such an antagonist, however, is not available for human use, but the view that IAPP deficiency plays an important role in the lack of delay in gastric emptying in response to hyperglycemia is further supported by the highly potent inhibitory effects of IAPP on gastric emptying (18).

ecific antagonist would be required. Such an antagonist, however, is not available for human use, but the view that IAPP deficiency plays an important role in the lack of delay in gastric emptying in response to hyperglycemia is further supported by the highly potent inhibitory effects of IAPP on gastric emptying (18). In studies suggesting that insulin may be another important regulator of gastric emptying in healthy volunteers (19), the effects of hyperinsulinemia on gastric emptying were found to be marginal (19). Interestingly, in studies undertaken in patients with type 1 diabetes, no effect of insulin on gastric emptying was detected (20). Gastric emptying was found to be increased in patients with type 1 diabetes under euglycemic conditions in our studies even though plasma insulin concentrations were significantly higher compared with those of healthy volunteers. This result also argues against an important effect of insulin on gastric emptying in patients with type 1 diabetes.

to be increased in patients with type 1 diabetes under euglycemic conditions in our studies even though plasma insulin concentrations were significantly higher compared with those of healthy volunteers. This result also argues against an important effect of insulin on gastric emptying in patients with type 1 diabetes. Gastric emptying is modulated by feedback mechanisms arising from the interaction of nutrients with the small intestine (21). Both intestinal vagus nerve activity and intestinal peptides regulate gastric emptying. Glucagon-like peptide 1 (GLP-1) inhibits gastric emptying (22). Its secretion, however, is stimulated by the intestinal nutrient content and flow rather than by the plasma glucose concentration itself (23). Furthermore, the fact that gastric emptying and thus nutrient flow to the intestine, which should reduce direct intestinal L-cell–stimulated and also cholinergic GLP-1 secretion, was delayed in the hyperglycemic experiments in healthy subjects speaks against a major role of GLP-1 as a mediator of hyperglycemia-induced delay in gastric emptying.

gastric emptying and thus nutrient flow to the intestine, which should reduce direct intestinal L-cell–stimulated and also cholinergic GLP-1 secretion, was delayed in the hyperglycemic experiments in healthy subjects speaks against a major role of GLP-1 as a mediator of hyperglycemia-induced delay in gastric emptying. Direct inhibition of vagal nerve activity induced by hyperglycemia could be another important factor to delay gastric emptying. To our knowledge there is no convincing evidence that hyperglycemia per se as applied in our studies affects vagal activity. Interestingly, in healthy humans hyperglycemia has been reported to cause profound inhibition of vagal activity accompanied by substantial IAPP secretion (24). In contrast, however, in IAPP-deficient patients with type 1 diabetes, hyperglycemia did not affect vagal activity (9). Because the inhibitory effect of IAPP on gastric emptying seems to be mediated via inhibition in vagal nerve activity (24), our experiments are consistent with the concept that the hyperglycemia-induced delay in gastric emptying may be at least partially regulated via an IAPP-mediated inhibitory effect on vagal nerve activity.

the inhibitory effect of IAPP on gastric emptying seems to be mediated via inhibition in vagal nerve activity (24), our experiments are consistent with the concept that the hyperglycemia-induced delay in gastric emptying may be at least partially regulated via an IAPP-mediated inhibitory effect on vagal nerve activity. Recent studies showed that IAPP and pramlintide suppress postprandial glucagon secretion (25,26). Indeed, we found greater suppression of postprandial glucagon in type 1 diabetic patients when pramlintide was given. This suppression may have occurred either directly by an inhibitory effect of pramlintide on the pancreatic α-cell or indirectly via reduced efflux of nutrients from the gut, because amino acids such as arginine are known to stimulate glucagon secretion (27). Thus, it remains unclear whether the greater suppression of glucagon secretion is attributable to a direct inhibition of the pancreatic α-cells or to reduced influx of nutrients from the gut. However, because endogenous glucose production was comparable in the placebo and pramlintide experiments in type 1 diabetic patients, we believe that the pramlintide-induced reduction of postprandial glucose concentrations was primarily due to the delay in gastric emptying.

to reduced influx of nutrients from the gut. However, because endogenous glucose production was comparable in the placebo and pramlintide experiments in type 1 diabetic patients, we believe that the pramlintide-induced reduction of postprandial glucose concentrations was primarily due to the delay in gastric emptying. Rates of meal-derived glucose appearance were significantly greater in the early postprandial period in type 1 diabetic subjects when placebo was given and would correspond to a ∼30% reduction in splanchnic glucose sequestration. If we assume that all of this glucose had been used for glycogen formation, our estimates are in close agreement with nuclear magnetic resonance spectroscopic studies revealing a 30% reduction in glycogen content in moderately hyperglycemic patients with type 1 diabetes (28). Interestingly, in the pramlintide experiments a nearly identical proportion of hepatic glucose sequestration was found in patients with type 1 diabetes compared with our healthy subjects. This result could have been related to the greater suppression of postprandial glucagon secretion or initially reduced influx of glucose from the gut and more efficient uptake by the liver or increased rates of glycolysis in the gut.

tration was found in patients with type 1 diabetes compared with our healthy subjects. This result could have been related to the greater suppression of postprandial glucagon secretion or initially reduced influx of glucose from the gut and more efficient uptake by the liver or increased rates of glycolysis in the gut. Teleologically, the slowing of gastric emptying during hyperglycemia can be seen as an important defense mechanism to prevent hyperglycemia. IAPP secretion is linked to insulin release (10). As a response to hyperglycemia, the pancreatic β-cell with its glucose sensor increases insulin and IAPP secretion. Insulin suppresses hepatic glucose output and increases peripheral glucose uptake (17). IAPP reduces the release of nutrition from the gut and thus reduces the efflux of glucose into the system and thereby prevents aggravation of postprandial hyperglycemia. Taken together, these studies highlight the importance of a delay in gastric emptying as a response to hyperglycemia to minimize postprandial glucose excursions, a defense mechanism not operative in patients with type 1 diabetes, which may be explained at least partially by IAPP deficiency. The present work was supported by Deutsche Forschungsgemeinschaft Grant SchI527/5-2 to H.J.W. and J.S., by an unrestricted grant from Amylin Pharmaceuticals (San Diego, CA), and by National Institute of Diabetes and Digestive and Kidney Disease Grant DK-20411 to J.E.G. We thank the laboratory staff of the Clinical Research Unit and are especially indebted to Rita Schinkmann and Silke Herrmann for their superb help.

Impairment of insulin sensitivity is considered the background defect that interplays with the add-on progressive β-cell dysfunction to underlie the development of type 2 diabetes (1,2). An atypical form of diabetes, ketosis-prone diabetes (KPD), has been described over the past 2 decades and may represent a significant proportion of diabetes cases in people of sub-Saharan African origin (3,4). Patients with KPD present at onset with acute hyperglycemia, usually >30 mmol/l, and ketosis or ketoacidosis as type 1 diabetes but do not have autoimmune markers against the islet β-cell (3,5–7). The correction of those insulin-requiring acute-phase disorders is followed in >50% of cases by an insulin-free near-normoglycemic remission weeks to months later (8–10), thus resembling the course of type 2 diabetes. The pathogenesis and, consequently, the classification of KPD are still debated. It was classified under idiopathic type 1 diabetes or type 1B diabetes (11). However, growing evidence based on clinical and metabolic studies suggests its high phenotypical likeness to type 2 diabetes, and “ketosis-prone type 2 diabetes” has been proposed as a provisional name and is being used elsewhere (4,8,12). Metabolic studies have evidenced insulin secretion deficiency as the major determinant of the ketotic onset (8–10). This deficit is marked by a loss of acute-phase insulin secretion in response to intravenous glucose (10) or a decrease in C-peptide response to glucagon (9,10). The subsequent remission process is due to a restoration, at least partial, of the β-cell insulin secretory capacity after achievement of good metabolic control (8,10). Insulin action was assessed in three reports, but only toward glucose metabolism, and was found to be normal or decreased while patients were in good metabolic control (6,8,10). Moreover, most studies on KPD have been reported in African-Americans who are more overweight than native Africans and may be metabolically different from them, as suggested earlier (13).

but only toward glucose metabolism, and was found to be normal or decreased while patients were in good metabolic control (6,8,10). Moreover, most studies on KPD have been reported in African-Americans who are more overweight than native Africans and may be metabolically different from them, as suggested earlier (13). In this study, we aimed at characterizing all aspects of insulin action in Africans with KPD when in the near-normoglycemic state without insulin treatment compared with control subjects of the same geographic origin.

but only toward glucose metabolism, and was found to be normal or decreased while patients were in good metabolic control (6,8,10). Moreover, most studies on KPD have been reported in African-Americans who are more overweight than native Africans and may be metabolically different from them, as suggested earlier (13). In this study, we aimed at characterizing all aspects of insulin action in Africans with KPD when in the near-normoglycemic state without insulin treatment compared with control subjects of the same geographic origin. RESEARCH DESIGN AND METHODS This study was undertaken at the Clinical Investigation Center of Saint-Louis University Hospital, Paris, France. We studied 15 subjects of sub-Saharan African origin with KPD who were in insulin-free remission, along with 17 healthy control subjects of the same geographic origin, with normal glucose tolerance. All participants were from West and Central Africa; they were born in Africa with no other racial antecedent in their ancestry and had migrated to France at adult age. KPD was defined as previously described (8). All patients had been diagnosed, had received insulin treatment at diagnosis, and were followed in the Department of Diabetes and Endocrinology of our hospital. Insulin-free remission was defined as maintenance of an A1C level ≤7.0% for at least 3 months after the withdrawal of insulin treatment, which was initiated at onset or relapse. Healthy control subjects were recruited by advertisement; they were matched to patients for age, sex, and BMI and were free of any family history of type 2 diabetes among their first-degree relatives. The age at inclusion averaged 44 years in each group. Diabetes was of short duration (6–72 months, mean 25.2), and patients had been in insulin-free remission for 3–45 months (mean 10.5). As antidiabetic treatment, most patients were on metformin alone (n = 8) or combined to a sulfonylurea (n = 2). Two patients were on diet alone, one on a sulfonylurea alone, one on a glinide, and one on acarbose. Four patients had presented with diabetes ketoacidosis (at least 2+ ketonuria and plasma bicarbonate <15 mmol/l and/or arterial pH <7.30), whereas the others had ketosis (at least 2+ ketonuria).

a sulfonylurea (n = 2). Two patients were on diet alone, one on a sulfonylurea alone, one on a glinide, and one on acarbose. Four patients had presented with diabetes ketoacidosis (at least 2+ ketonuria and plasma bicarbonate <15 mmol/l and/or arterial pH <7.30), whereas the others had ketosis (at least 2+ ketonuria). Participants underwent a screening, and those included had normal liver, cardiovascular, pulmonary, and kidney function assessed by medical history, physical examination, electrocardiography, and routine biochemical and hematological tests, as well as negative hepatitis B and C and human immunodeficiency viruses’ serological tests. The A1C level was also confirmed in patients during that visit. Patients on oral antidiabetic drugs were asked to stop them at least 5 days before the procedures, and no healthy control subject was taking a drug known to affect glucose or lipid homeostasis. Fasting blood glucose of all patients remained below 8.2 mmol/l. The study was approved by the ethics committee of Paris Saint-Louis, and each participant gave a written informed consent to participate.

before the procedures, and no healthy control subject was taking a drug known to affect glucose or lipid homeostasis. Fasting blood glucose of all patients remained below 8.2 mmol/l. The study was approved by the ethics committee of Paris Saint-Louis, and each participant gave a written informed consent to participate. Anthropometric measurements and dual-energy X-ray absorptiometry In all participants, height (to the nearest 0.5 cm) was measured using a wall-stuck stadiometer, and weight was measured to the nearest 0.1 kg (SECA scale, Hamburg, Germany). The BMI was calculated as the weight (in kilograms) divided by the square of the height (in meters). The waist circumference (to the nearest 1 cm) was measured at the midway between the lower costal margin and the iliac crest, while the person was in the upright position, using a nonstretchable tape. Percent fat, fat mass, and fat-free mass were measured by dual-energy X-ray absorptiometry using an absorptiometer (Hologic QDR-1000/W; Wilmington, MA). The anthropometric characteristics and body mass distribution were comparable between the two groups.

person was in the upright position, using a nonstretchable tape. Percent fat, fat mass, and fat-free mass were measured by dual-energy X-ray absorptiometry using an absorptiometer (Hologic QDR-1000/W; Wilmington, MA). The anthropometric characteristics and body mass distribution were comparable between the two groups. Metabolic assessments Oral glucose tolerance test. To confirm normal glucose tolerance in control subjects and to estimate the β-cell function in all participants, an oral glucose tolerance test was performed on the screening visit after a 12-h overnight fast. Blood samples were collected before (t0) and 30 (t30) and 120 min (t120) after a 75-g oral glucose load, for determination of plasma glucose and insulin concentrations. The glucose tolerance status was classified according to the current American Diabetes Association criteria (1).

ing visit after a 12-h overnight fast. Blood samples were collected before (t0) and 30 (t30) and 120 min (t120) after a 75-g oral glucose load, for determination of plasma glucose and insulin concentrations. The glucose tolerance status was classified according to the current American Diabetes Association criteria (1). Euglycemic clamp. A two-step euglycemic-hyperinsulinemic clamp was performed within the week after the screening visit, after a 12-h in-hospital overnight fast. It consisted of a first step (low-dose insulin infusion) at 10 mU/m2 body surface per min for 100 min to measure the effects of insulin on plasma nonesterified fatty acids (NEFAs). This was followed by a primed 100-min step (high-dose insulin infusion) at 80 mU/m2 per min insulin infusion to evaluate the effects of insulin on glucose disposal. The endogenous glucose production was also measured during the whole procedure. Fasting (basal) blood samples were collected at −70 and −60 min before the starting of the clamp. At −60 min, a priming bolus of 3 mg/kg d-[6,6-2H2]glucose (deuterated glucose) (96 molar percent excess) (Assistance Publique-Hôpitaux de Paris, Paris, France) was injected, followed by a continuous infusion at 0.05 mg · kg−1 · min−1 for 260 min. Continuous insulin infusion was then started at t0, and glucose concentration was measured every 5 min during the whole procedure. Blood glucose level was clamped at 5.5 mmol/l using a variable infusion of 20% glucose, based on the negative feedback principle (14). Arterialized blood samples were drawn at baseline for the measurement of basal [6,6-2H2]glucose enrichment and every 10 min during the last 20 min of each step (80th, 90th, and 100th min and 180th, 190th, and 200th min) for the measurement of plasma insulin, NEFA, and [6,6-2H2]glucose enrichment.

edback principle (14). Arterialized blood samples were drawn at baseline for the measurement of basal [6,6-2H2]glucose enrichment and every 10 min during the last 20 min of each step (80th, 90th, and 100th min and 180th, 190th, and 200th min) for the measurement of plasma insulin, NEFA, and [6,6-2H2]glucose enrichment. Analytical techniques All assays were run in duplicate. Plasma glucose was measured by the hexokinase method (Roche Diagnostics, Mannheim, Germany). The high-performance liquid chromatography method was used to measure A1C. Plasma insulin was measured using immuno-radiometric assays (BI-INSULIN IRMA; Cis Bio-International, Gif-Sur-Yvette, France) with a detection limit of 0.2 μU/ml and an intra- and interassay coefficient of variation (CV) <9.5%. Plasma NEFA concentrations were determined using the colorimetric method (Randox Laboratories, Antrim, U.K.). Plasma [2H2]glucose enrichment was measured by selected ion monitoring electron impact gas chromatography–mass spectrometry (5971A; Hewlett-Packard, Palo Alto, CA) as previously described (15). All other biochemical tests were done using routine laboratory methods. Calculations Basal concentration of each biochemical parameter was calculated as the mean of two values obtained from blood samples collected 10 min apart, and the steady-state concentrations were the average of the three values measured 10 min apart during the last 20 min of each step.

Plasma [2H2]glucose enrichment was measured by selected ion monitoring electron impact gas chromatography–mass spectrometry (5971A; Hewlett-Packard, Palo Alto, CA) as previously described (15). All other biochemical tests were done using routine laboratory methods. Calculations Basal concentration of each biochemical parameter was calculated as the mean of two values obtained from blood samples collected 10 min apart, and the steady-state concentrations were the average of the three values measured 10 min apart during the last 20 min of each step. For endogenous glucose production (EGP), after an overnight fast, steady-state conditions for the deuterated glucose prevailed, and the basal EGP (bEGP) equaled the rate of glucose appearance (Ra). It was therefore calculated as the deuterated glucose infusion rate (mg · kg−1 · min−1) divided by the plasma enrichment of [6,6-2H2]glucose. During the two steps of the euglycemic clamp, due to the non–steady-state conditions, the Steele's equation was used to estimate the Ra (16). The residual EGP at the last 20 min of the first (residual EGP1 [rEGP1]) and second (residual EGP2 [rEGP2]) steps of the glucose clamp were therefore obtained by subtracting the unlabeled glucose infusion rate from the total Ra at each step of the clamp. Negative values of EGP, observed only at the high-infusion step, were considered as nil EGP. The product EGP × plasma insulin at basal and at the end of each step was used as the hepatic insulin resistance index (17).

tained by subtracting the unlabeled glucose infusion rate from the total Ra at each step of the clamp. Negative values of EGP, observed only at the high-infusion step, were considered as nil EGP. The product EGP × plasma insulin at basal and at the end of each step was used as the hepatic insulin resistance index (17). The insulin-stimulated glucose disposal rate (M) was calculated from the glucose infusion rate during the last 20 min of the second step of the glucose clamp after accounting for inter-individual differences in glucose space (14) and was expressed in milligram per kilogram of body fat-free mass per minute. Total glucose disposal (TGD) rate was obtained by adding rEGP2 to M. The insulinogenic index (Δinsulin0-30/Δglucose0-30) was used to estimate the early insulin secretion during the oral glucose tolerance test and the product of the TGD and the insulinogenic index as a measure of β-cell function. The homeostasis model assessment–insulin resistance (HOMA-IR) index was calculated as fasting glucose (mmol/l) × fasting insulin (μU/ml)/22.5. At baseline and at the end of each clamp step, we used the product mean NEFA concentration × mean plasma insulin as a surrogate of the adipose insulin resistance (that is, the resistance to NEFA suppression by insulin).

The homeostasis model assessment–insulin resistance (HOMA-IR) index was calculated as fasting glucose (mmol/l) × fasting insulin (μU/ml)/22.5. At baseline and at the end of each clamp step, we used the product mean NEFA concentration × mean plasma insulin as a surrogate of the adipose insulin resistance (that is, the resistance to NEFA suppression by insulin). Statistical analysis Results are represented as means ± SE and percentage, unless stated otherwise. Statistical analysis was performed using SPSS software version 12.0 (SPSS, Chicago, IL). We used the Fisher's exact test to compare categorical variables and the nonparametric Mann-Whitney U test for quantitative variables. The level of significance was set at P < 0.05.

ans ± SE and percentage, unless stated otherwise. Statistical analysis was performed using SPSS software version 12.0 (SPSS, Chicago, IL). We used the Fisher's exact test to compare categorical variables and the nonparametric Mann-Whitney U test for quantitative variables. The level of significance was set at P < 0.05. RESULTS Characteristics of the participants and biochemical and metabolic parameters The age at inclusion and BMI were comparable between the two groups (Table 1). Fasting plasma glucose level was <7.0 mmol/l in 87% (n = 13) of patients. During the oral glucose tolerance test, two patients had the profile of impaired glucose tolerance and others fulfilled the criteria of diabetes. As shown in Table 1, there was no significant difference between patients with KPD and controls with respect to plasma lipid parameters. Patients had a significantly higher fasting plasma glucose concentration; fasting plasma insulin levels were also higher in patients, although the difference was not significant (Table 2). The insulinogenic index was significantly higher in the control group (Table 2). The insulinogenic index of patients who presented initially with ketoacidosis was not different from that of individuals who presented with ketosis alone (4.5 ± 1.6 vs. 4.7 ± 1.1; P = 0.95). During insulin infusion, steady-state plasma insulin concentrations at the last 20 min of step one (SSPI1) and step two (SSPI2) were comparable between the two groups (Table 2).

d initially with ketoacidosis was not different from that of individuals who presented with ketosis alone (4.5 ± 1.6 vs. 4.7 ± 1.1; P = 0.95). During insulin infusion, steady-state plasma insulin concentrations at the last 20 min of step one (SSPI1) and step two (SSPI2) were comparable between the two groups (Table 2). Insulin-mediated glucose disposal Mean TGD rate was reduced by 30% in KPD patients (P = 0.018) (Table 2). This difference remained significant after adjustment for BMI (P = 0.034). The HOMA-IR was accordingly higher in KPD patients. The product of TGD and insulinogenic index was also markedly reduced in KPD patients compared with control subjects (Table 2). Endogenous glucose production The EGP was significantly higher in patients at baseline (bEGP) and at the end of the first step (rEGP1) of the glucose clamp (Table 2). During the last step of the clamp, rEGP2 was 0 in control subjects, whereas it was still positive in four patients, and the difference between the two groups was significant (P = 0.007). The EGP as a function of the plasma insulin level (dose-response curve) is displayed in Fig. 1A. The hepatic insulin resistance index was 33% higher in patients at basal (bEGP × fasting plasma insulin). During both low-dose (rEGP1 × SSPI1) and high-dose (rEGP2 × SSPI2) insulin infusions, this index remained higher in patients (Table 2).

ion of the plasma insulin level (dose-response curve) is displayed in Fig. 1A. The hepatic insulin resistance index was 33% higher in patients at basal (bEGP × fasting plasma insulin). During both low-dose (rEGP1 × SSPI1) and high-dose (rEGP2 × SSPI2) insulin infusions, this index remained higher in patients (Table 2). NEFA suppression Basal NEFA concentration was 57% higher in patients compared with control subjects (Table 2), even after adjustment for BMI (P = 0.024). It remained significantly elevated at the end of the low-dose insulin infusion step (SSNEFA1) and at the borderline significance level at the end of the high-dose insulin infusion step (SSNEFA2) of the euglycemic clamp (Table 2). By contrast, the relative decline from baseline was similar between the two groups, whatever the step (in patients and control subjects, respectively: first step 61.5 ± 5.5 vs. 64.9 ± 4.2%, P = 0.7; second step 89.5 ± 1.8 vs. 88.2 ± 1.6%, P = 0.5). Figure 1B presents the plasma NEFA concentration as a function of plasma insulin levels. Insulin resistance index to NEFA disappearance (IRNEFA) was doubled in patients compared with controls at basal (fasting IRNEFA) and during the clamp (IRNEFA1 and IRNEFA2), with the difference being statistically significant at basal but at borderline significance at the first and second steps (Table 2).

nsulin levels. Insulin resistance index to NEFA disappearance (IRNEFA) was doubled in patients compared with controls at basal (fasting IRNEFA) and during the clamp (IRNEFA1 and IRNEFA2), with the difference being statistically significant at basal but at borderline significance at the first and second steps (Table 2). CONCLUSIONS Because ketosis is the hallmark of KPD, the role of insulin secretion has been widely studied at onset and in the long-term course of the disease (6,8–10). In two among these studies, insulin sensitivity toward glucose metabolism was evaluated using either the euglycemic clamp (6) or the minimal model (10). Compared with control subjects’ values, it was reported to be similar (10) or reduced (6) in patients with or without insulin treatment after they recovered from the acute ketotic episode. We also previously used the intravenous insulin tolerance test to evaluate insulin sensitivity in a larger cohort of KPD patients (8). We showed that, although markedly impaired during the ketotic phase, insulin sensitivity improved significantly after 6 months of follow-up and almost reached nondiabetic values in patients who became insulin independent but not in patients who still required insulin for metabolic control.

r cohort of KPD patients (8). We showed that, although markedly impaired during the ketotic phase, insulin sensitivity improved significantly after 6 months of follow-up and almost reached nondiabetic values in patients who became insulin independent but not in patients who still required insulin for metabolic control. In the present study, to assess various aspects of insulin action in patients with KPD in near-normoglycemic remission, we used a two-step euglycemic-hyperinsulinemic clamp to compare them with matched control subjects with normal glucose tolerance. This allowed us to provide a direct and complete characterization of insulin sensitivity in a phenotype of diabetes that still requires thorough insight for appropriate classification and treatment. The first finding of our study is that despite insulin-free near-normoglycemic remission in sub-Saharan African adults with KPD, insulin-mediated glucose disposal is markedly reduced. This was previously suggested by Banerji et al. (6) in patients with KPD in good metabolic control, either insulin treated or not.

ment. The first finding of our study is that despite insulin-free near-normoglycemic remission in sub-Saharan African adults with KPD, insulin-mediated glucose disposal is markedly reduced. This was previously suggested by Banerji et al. (6) in patients with KPD in good metabolic control, either insulin treated or not. Another important finding is that despite insulin-free near-normoglycemic remission, in the postabsorptive state, patients with KPD display a higher endogenous glucose production rate, which mostly corresponds to hepatic glucose production (18). This is related to an increased hepatic insulin resistance as evidenced by the higher basal hepatic insulin resistance index (bEGP × fasting plasma insulin). During both low- and high-dose insulin infusions, this index remained higher in patients, indicating that for a given insulin concentration, the suppression of EGP was less marked in them compared with control subjects.

e as evidenced by the higher basal hepatic insulin resistance index (bEGP × fasting plasma insulin). During both low- and high-dose insulin infusions, this index remained higher in patients, indicating that for a given insulin concentration, the suppression of EGP was less marked in them compared with control subjects. Our last important finding is that at fasting and during the low-dose insulin infusion, plasma NEFA concentrations were higher in patients than in control subjects. This was substantiated by the higher basal adipose insulin resistance index, demonstrating that KPD patients in near-normoglycemic remission display adipose insulin resistance. Of note, increased circulating NEFAs may in turn worsen insulin resistance and the insulin secretion defect. The lack of difference in NEFA response to insulin infusion may be related to the good metabolic control of patients or to the small number of subjects studied. Absolute basal NEFA concentrations seem quite high in our subjects. This may be because we did not use a lipolysis inhibitor during blood sample collection, although higher NEFA levels have been previously reported in black populations (19).

good metabolic control of patients or to the small number of subjects studied. Absolute basal NEFA concentrations seem quite high in our subjects. This may be because we did not use a lipolysis inhibitor during blood sample collection, although higher NEFA levels have been previously reported in black populations (19). Reduced muscle glucose uptake and decreased liver and adipose insulin sensitivity have also been reported in Caucasians with type 2 diabetes by Groop et al. (20) and are usually considered as characteristic features of type 2 diabetes (21,22). However, in these reports, diabetic patients were hyperglycemic, with mean fasting plasma glucose at 10.5 mmol/l (22) or A1C level averaging 9.6% (20). In our study, patients were in near-normoglycemia with mean A1C level at 6.2%. Although all antidiabetic medications, if any, were discontinued many days before the investigation, fasting plasma glucose averaged 6.3 mmol/l and was below the diabetic-defining cutoff of 7 mmol/l in 87% of patients. Thus, this is the first investigation of diabetic patients in a metabolic state close to normoglycemia without insulin treatment. The multiorgan insulin resistance observed in near-normoglycemia suggests that these defects are primary rather than secondary to the diabetic state. We also show that the impairment of insulin sensitivity is associated with a decreased early-phase insulin secretion resulting in a drastically reduced index of β-cell function (TGD × insulinogenic index). Indeed, it is now recognized that type 2 diabetes develops when insulin secretion is unable to compensate for insulin resistance (23,24).

mpairment of insulin sensitivity is associated with a decreased early-phase insulin secretion resulting in a drastically reduced index of β-cell function (TGD × insulinogenic index). Indeed, it is now recognized that type 2 diabetes develops when insulin secretion is unable to compensate for insulin resistance (23,24). The fact that patients with KPD display metabolic abnormalities that characterize type 2 diabetes does not explain the ketotic onset or relapses that define KPD. A genetic defect or an environmental factor making the β-cells more susceptible to gluco- and/or lipotoxicity may be a potential factor. To date, no prominent genetic factor has been identified. We recently proposed that an endemic asymptomatic viral infection may be the ketotic precipitating factor in such patients from sub-Saharan Africa. Among possible candidates, we focused on human herpesvirus-8 (HHV-8). We found a high association between KPD and HHV-8 infection and evidenced that this virus was able to infect human β-cells in vitro (25). We acknowledge that the small number of subjects may have minimized the role of BMI and/or percent body fat on the insulin sensitivity defects. Indeed, these anthropometric parameters were slightly higher in patients, although not significantly. However, adjustment for BMI or percent body fat did not change the significance of our results. Also, it should be noted that despite the near-normoglycemic state, the higher fasting plasma glucose in patients might have by itself worsened the insulinogenic index (26).

ly higher in patients, although not significantly. However, adjustment for BMI or percent body fat did not change the significance of our results. Also, it should be noted that despite the near-normoglycemic state, the higher fasting plasma glucose in patients might have by itself worsened the insulinogenic index (26). In conclusion, in the context of near-normoglycemic remission, we observe insulin resistance at the level of muscle, adipose tissue, and liver. As suggested in reports on the recovery of β-cell function after the acute phase of KPD (8,10), these findings strongly indicate that KPD is a subtype of type 2 diabetes in which an uncommon factor triggers the ketotic onset or relapses. This work was supported by an institutional grant (PHRC) from Assistance Publique–Hôpitaux de Paris, the French Diabetes Association (AFD), a nonprofit organization, and the French-speaking Association for the Study of Diabetes and Metabolic Diseases (ALFEDIAM). Part of these results was presented at the 43rd European Association for the Study of Diabetes (EASD) annual meeting [S. Choukem et al. Diabetologia 50 (Suppl. 1):S277, 2007]. The authors are grateful to the participants, the nurse staff of the Clinical Investigation Center, and the technical staff of the Hormones Laboratory at Saint-Louis Hospital for their dedication.

With the explosive growth of incident diabetes, type 2 diabetes has become a major international public health challenge. Moreover, an increasing number of individuals have evidence of a pre-diabetic state, which indicates significant future risk of developing diabetes. Fortunately, accumulating evidence suggests that type 2 diabetes can be delayed or prevented in individuals with pre-diabetes by either lifestyle modification or medication (1–2). However, because prevention is a fundamental public health goal, there is clearly a great need for effective strategies to identify high-risk individuals. Unfortunately, the best available risk stratification method is an oral glucose tolerance test (OGTT), which is both costly and difficult to perform in a clinical setting. The albumin-to-creatinine ratio (ACR) in a single untimed urinary specimen is a reflection of urinary albumin excretion and is increasingly being accepted as a marker that predicts several important health outcomes, including hypertension, kidney failure, cardiovascular events, and mortality (3–5). These associations have been observed throughout the biological range, even at levels far below those previously considered to be pathological (e.g., microalbuminuria) (6).

ed as a marker that predicts several important health outcomes, including hypertension, kidney failure, cardiovascular events, and mortality (3–5). These associations have been observed throughout the biological range, even at levels far below those previously considered to be pathological (e.g., microalbuminuria) (6). The ACR is also closely linked to cardiometabolic risk factors, vascular disease, and insulin resistance (7–9) and might therefore play a clinically important role in predicting future onset of diabetes. Observational studies have shown an association between ACR and other markers of urinary albumin excretion and incident diabetes (10–12). In addition, observations that proteinuria-reducing therapies (e.g., ACE inhibitors and angiotensin II receptor blockers) delay progression to diabetes also support this hypothesis, albeit indirectly (13,14). However, the observational studies were heterogeneous in terms of study design and risk for incident diabetes, did not include proteinuria throughout its biologic range, and/or recruited individuals from ethnic groups distinct from the general U.S. population. In addition, a randomized clinical trial did not find that ramipril, a proteinuria-reducing agent, altered the incidence of diabetes (15). The Diabetes Prevention Program (DPP) enrolled a large and well-characterized cohort of adults who were at high-risk for developing diabetes based on having elevated fasting glucose and impaired glucose tolerance. We tested the hypothesis that ACR, throughout its biological range, improves the prediction of future diabetes.

iabetes Prevention Program (DPP) enrolled a large and well-characterized cohort of adults who were at high-risk for developing diabetes based on having elevated fasting glucose and impaired glucose tolerance. We tested the hypothesis that ACR, throughout its biological range, improves the prediction of future diabetes. RESEARCH DESIGN AND METHODS The eligibility criteria, design, and methods of the DPP have been reported elsewhere (16). In brief, eligibility criteria included age ≥25 years, BMI ≥24 kg/m2 (≥22 kg/m2 in Asian Americans), and fasting plasma glucose levels between 95 and 125 mg/dl (lower limit did not apply to the American Indian centers) in addition to impaired glucose tolerance (IGT) by an oral glucose tolerance test (OGTT) (plasma glucose of 140–199 mg/dl 2 h after a 75-g oral glucose load). Participants were recruited from 27 U.S. study sites and were excluded if they had conditions that would impair their ability to participate or took certain medicines, including thiazide diuretics and β-blockers (16). All participants gave informed consent and signed documents approved by the institutional review board at each center. Eligible participants received standard advice on a healthy diet and physical activity and were randomly assigned to one of three additional interventions (intensive lifestyle intervention versus metformin versus matching placebo).

consent and signed documents approved by the institutional review board at each center. Eligible participants received standard advice on a healthy diet and physical activity and were randomly assigned to one of three additional interventions (intensive lifestyle intervention versus metformin versus matching placebo). Measurements and laboratory tests Development of diabetes was determined by an annual OGTT or by a semiannual fasting plasma glucose level with confirmation by a second test, using the criteria of the American Diabetes Association (21) and the World Health Organization (17). Urinary albumin excretion was estimated from a morning fasting spot urine sample by the urinary albumin-to-urinary creatinine ratio (i.e., milligrams of albumin per gram of creatinine). Urinary albumin was measured using Behring reagents on the BN II nephelometer (Dade Behring, Deerfield, IL) (interassay coefficient of variation [CV] 4.4% and intra-assay CV 4.3%). Microalbuminuria was defined using standard criteria as ACR between 30 and 300 mg/day, whereas macroalbuminuria was defined as ACR >300 mg/day. Serum and urinary creatinine concentrations were measured using Roche reagents on the Hitachi 917 autoanalyzer (Boehringer Mannheim, Mannheim, Germany) (serum creatinine: interassay CV 3.5% and intra-assay CV 3.2%; urinary creatinine: interassay CV 1.8% and intra-assay CV 1.2%).

uminuria was defined as ACR >300 mg/day. Serum and urinary creatinine concentrations were measured using Roche reagents on the Hitachi 917 autoanalyzer (Boehringer Mannheim, Mannheim, Germany) (serum creatinine: interassay CV 3.5% and intra-assay CV 3.2%; urinary creatinine: interassay CV 1.8% and intra-assay CV 1.2%). Immunoreactive insulin was measured in plasma. Measurement methods for glucose and insulin have been published previously (18). Insulin secretion and sensitivity were expressed using glucose and insulin measured in conventional units (milligrams per deciliter and microunits per milliliter, respectively). Insulin secretion was measured using the corrected insulin response = (100 × 30-min insulin)/(30-min glucose × [30-min glucose − 70 mg/dl]) (19). Insulin sensitivity was measured using the insulin sensitivity index, which is 22.5/(fasting insulin × [fasting glucose/18.0]), the reciprocal of which is the homeostasis model assessment of insulin resistance (HOMA-IR) (20). Baseline demographic and anthropometric (i.e., BMI measured as weight in kilograms divided by height in meters and waist circumference) data were also measured.

ex, which is 22.5/(fasting insulin × [fasting glucose/18.0]), the reciprocal of which is the homeostasis model assessment of insulin resistance (HOMA-IR) (20). Baseline demographic and anthropometric (i.e., BMI measured as weight in kilograms divided by height in meters and waist circumference) data were also measured. Statistical methods Because of its skewed distribution, ACR was analyzed both as a categorical (i.e., by quartiles) and as a continuous (i.e., after base 2 logarithm transformation) variable. Baseline characteristics were described by means ± SD for continuous variables and percentages for categorical variables. Spearman's correlation coefficients between continuous variables and ACR were reported. For categorical variables, ANOVA was used to identify differences in the base 2 logarithm-transformed ACR. Because a significant interaction (P < 0.05) was noted between treatment group and incident diabetes when ACR was divided into quartiles (but not when it was examined as a continuous variable after base 2 logarithm transformation), the former analysis was stratified by treatment arm. Cox proportional hazards models were used to evaluate the association between ACR and risk of developing diabetes by both univariate and multivariate means. The first multivariate model adjusted for the demographic factors age, sex, and race/ethnicity alone, whereas the second included demographics with the time-dependent covariates exercise and weight loss as well as baseline characteristics that were associated with ACR in a statistically significant manner (at the 0.05 level). All analyses were performed using SAS (SAS Institute, Cary, NC). Nominal P values are presented without adjustment for multiplicity of testing.

time-dependent covariates exercise and weight loss as well as baseline characteristics that were associated with ACR in a statistically significant manner (at the 0.05 level). All analyses were performed using SAS (SAS Institute, Cary, NC). Nominal P values are presented without adjustment for multiplicity of testing. RESULTS Descriptive data The DPP randomly assigned 3,234 participants to one of three treatment arms (placebo, metformin, or lifestyle). Our analysis included only the 3,188 individuals with ACR data at baseline. The distribution of baseline ACR was highly skewed at the upper end. Overall, 2,997 participants (94%) had ACR levels below the microalbuminuria cutoff point of 30 mg/g, and only 14 (0.4%) had macroalbuminuria (ACR level >300 mg/g). The medians and ranges of ACR within quartiles are shown in Table 1. The cross-sectional association between baseline characteristics and ACR is shown in Table 1. ACR was positively associated with age and markers of adiposity and insulin secretion and resistance, blood pressure, and use of antihypertensive agents with antiproteinuric effects and was inversely related to male sex and serum creatinine, both of which influence urinary excretion of creatinine, the denominator in the ACR. Race/ethnicity, prior gestational diabetes, family history of diabetes, serum lipids, and smoking were not significantly associated with ACR.

ents with antiproteinuric effects and was inversely related to male sex and serum creatinine, both of which influence urinary excretion of creatinine, the denominator in the ACR. Race/ethnicity, prior gestational diabetes, family history of diabetes, serum lipids, and smoking were not significantly associated with ACR. ACR and development of diabetes The 3,188 participants were followed for a mean of 3.2 years (range 0–5.0 years), during which 674 (21%) developed diabetes. A test of heterogeneity revealed a significant interaction between ACR, treatment group, and diabetes risk, so the analysis was stratified by treatment group. Table 2 shows HRs for incident diabetes by ACR quartile for the unadjusted and fully adjusted stratified models. No consistent pattern was seen for either adjusted or unadjusted models or with a model adjusting for demographic characteristics alone (data not shown). When ACR was examined as a continuous variable, the unadjusted model showed a 7% increase in incident diabetes with every doubling of ACR (hazard ratio [HR] 1.07 [95% CI 1.0–1.1]), but statistical significance was lost (0.98 [0.91–1.06]) when the model was fully adjusted for the covariates baseline age, sex, race, BMI, waist circumference, fasting insulin, insulin sensitivity/secretion, systolic and diastolic blood pressure, serum creatinine, and ACE inhibitor and calcium channel blocker use and for time-dependent changes in weight and physical activity.

en the model was fully adjusted for the covariates baseline age, sex, race, BMI, waist circumference, fasting insulin, insulin sensitivity/secretion, systolic and diastolic blood pressure, serum creatinine, and ACE inhibitor and calcium channel blocker use and for time-dependent changes in weight and physical activity. To test the possibility that examining quartiles of ACR may not have been sensitive enough to reveal an association between incident diabetes and ACR at its higher range (e.g., microalbuminuria), as suggested by other studies (10,11), HRs were examined after each DPP treatment cohort was separately divided into 10 equal groups by ACR (∼100 participants/group) (Fig. 1). No consistent pattern was observed between ACR and incident diabetes before or after full adjustment for covariates in the highest decile, in which 191 of 318 subjects had micro- or macroalbuminuria. CONCLUSIONS Identifying a simple, safe, and inexpensive tool to improve prediction of future diabetes would be an important public health achievement, especially in light of the ongoing diabetes pandemic. Preliminary results from several observational studies raise the possibility that low levels of ACR could play such a role, with further (indirect) evidence being that proteinuria-reducing antihypertensive agents are associated with a reduced risk of incident diabetes (13,14). However, in the present study, the largest study of pre-diabetic individuals to date, we did not find that ACR had any independent predictive value.

play such a role, with further (indirect) evidence being that proteinuria-reducing antihypertensive agents are associated with a reduced risk of incident diabetes (13,14). However, in the present study, the largest study of pre-diabetic individuals to date, we did not find that ACR had any independent predictive value. Our study hypothesis was not necessarily dependent upon a causal link between increasing ACR and onset of diabetes. For example, one possible premise is that exposure to levels of glycemia below what is conventionally considered pathological induces changes in renal handling of urinary albumin, which in turn would lead to increased ACR. Alternatively, increased ACR could simply be one of a number of early, organ-specific manifestations of insulin resistance that herald the onset of diabetes. Regardless, we felt it important to test this hypothesis, especially given prior findings.

handling of urinary albumin, which in turn would lead to increased ACR. Alternatively, increased ACR could simply be one of a number of early, organ-specific manifestations of insulin resistance that herald the onset of diabetes. Regardless, we felt it important to test this hypothesis, especially given prior findings. Results from previous observational studies have supported an association between urinary albumin excretion and incident diabetes. In a longitudinal study of 2,205 American Indians, ACR of ≥30 mg/day (i.e., microalbuminuria or macroalbuminuria) predicted incident diabetes over an average of 4 years of follow-up in a combined group of men with normal or IGT (odds ratio 2.19 [95% CI 1.48–3.21]) and women with baseline IGT only (2.69 [1.41–5.21]) (11). A prospective, community-based Dutch study of 5,654 individuals with normal or impaired glucose tolerance found a stepwise increase in the 4-year risk of incident diabetes by baseline urinary albumin excretion as measured by 24-h urine collections (tertile 1, ≤6.9 mg/kg, 1.8%; tertile 2, 6.9–12.4 mg/kg, 2.3%; tertile 3, ≥12.4 mg/kg, 5.8%; P < 0.001) (12). Results were relatively unchanged when individuals with baseline IGT were excluded from the analysis. Mykkänen et al. (10) observed a significantly higher proportion of baseline microalbuminuria (44.4% vs. 30.4%, P = 0.017) in elderly Finns who developed diabetes after 3.5 years of follow-up (compared with those who did not), although there was no actual difference in mean ACR between groups. In this study, the increased odds of developing diabetes were no longer statistically significant after adjustment for fasting plasma glucose and insulin.

erly Finns who developed diabetes after 3.5 years of follow-up (compared with those who did not), although there was no actual difference in mean ACR between groups. In this study, the increased odds of developing diabetes were no longer statistically significant after adjustment for fasting plasma glucose and insulin. Our study contributes new information that has been lacking in several important ways. First, our cohort was composed exclusively of individuals who were at high risk of developing diabetes on the basis of elevated fasting glucose and IGT plus overweight or obesity (for the majority). Because our sample size was far larger than all the IGT subgroups from the previously mentioned studies combined, it is unlikely that our negative findings were related to insufficient statistical power. Interestingly, despite their elevated risk for diabetes, the great majority of our cohort did not have microalbuminuria. Second, we analyzed ACR throughout its continuous range, avoiding artificial cutoff values, such as microalbuminuria, that could limit its descriptive utility (6). On the other hand, the DPP included very few subjects with micro- or macroalbuminuria, which could have reduced the power to show an association in these ranges and the generalizability of our results. Indeed, the predictive power of ACR could have been obscured by the increased risk for development of diabetes present in the DPP cohort at baseline. The fact that ACR was so closely associated with insulin or glycemic parameters supports this hypothesis. In addition, we excluded individuals with chronic kidney disease, further reducing generalizability, and measured ACR only once at baseline. ACR can be affected by diet, physical activity, and other habits, which may have introduced some variability into our findings, although this would be limited somewhat by the uniform collection criterion (i.e., fasting morning sample). Finally, major differences between the DPP and prior studies were the ethnically and culturally diverse cohort and the fact that we detected early diabetes by annual or semiannual surveillance glucose tolerance tests using accepted criteria (17,21), thus reducing the likelihood that ACR reflected prior exposure to severe hyperglycemia.

fferences between the DPP and prior studies were the ethnically and culturally diverse cohort and the fact that we detected early diabetes by annual or semiannual surveillance glucose tolerance tests using accepted criteria (17,21), thus reducing the likelihood that ACR reflected prior exposure to severe hyperglycemia. The study hypothesis was based on the presumption that subtle damage occurs within the kidney in the pre-diabetic state—whether from chronic exposure to abnormal glycemia that is below the formal threshold for diabetes, elevated intrarenal blood pressure, or oxidative stress (22), among other causes—that manifests itself as elevated ACR. ACR, as a subtle marker of incipient damage, could in turn herald the onset of diabetes. Although we did not confirm such a relationship, this does not exclude the possibility that ACR can predict hard outcomes in this population, as it has in others. In summary, in a population of subjects at elevated risk for diabetes, ACR below the microalbuminuria range does not predict incident diabetes. Supplementary Material Online-Only Appendix A.N.F. is supported by the National Institutes of Health (K23 RR019615).

Approximately 10% of urban Indian men and women aged 40–49 years have type 2 diabetes, and a rising prevalence is predicted to produce 80 million diabetic patients in India by 2030 (1–3). Cardiovascular disease is also rising (4). Similar trends, thought to reflect increasing obesity, are occurring in other developing countries undergoing economic transition, and interventions to prevent disease are urgently needed. Research in high-income countries has shown that factors linked to weight gain in early life contribute to the risk of developing diabetes and cardiovascular disease. Low birth weight (5,6) and accelerated gain in BMI during childhood and adolescence predict increased risk (7,8). The optimal pattern of infant weight gain (the first 1–2 postnatal years) is unclear; studies of adults suggest that low infant weight gain is a risk factor for later disease (7–9), whereas studies of children suggest the opposite (10,11). There are few data from developing countries. In the New Delhi birth cohort (12,13), children were measured at birth and every 6 months throughout infancy, childhood, and adolescence. We reported earlier that low BMI in infancy and rapid childhood BMI gain were associated with an increased risk of adult diabetes or impaired glucose tolerance (IGT) (12). We have now examined other cardiovascular risk factors and the cluster of risk factors known as the metabolic syndrome.

childhood, and adolescence. We reported earlier that low BMI in infancy and rapid childhood BMI gain were associated with an increased risk of adult diabetes or impaired glucose tolerance (IGT) (12). We have now examined other cardiovascular risk factors and the cluster of risk factors known as the metabolic syndrome. RESEARCH DESIGN AND METHODS During 1969–1972, married women living in a 12-km2 area of Delhi (n = 20,755) were followed up (12,13). There were 9,169 pregnancies and 8,181 live births. Trained personnel recorded the babies’ weight and length within 72 h of birth and every 6 months until age 14–21 years. Gaps in funding interrupted measurements in 1972–1973 and 1980–1982. At recruitment, 60% of families had incomes of >50 rupees/month (national average 28 rupees/month) and 15% of parents were illiterate (national average 66%). Nevertheless, 43% of families lived in one room. Hindus were the majority religious group (84%), followed by Sikhs (12%), Christians (2%), Muslims (1%), and Jains (1%).

ruitment, 60% of families had incomes of >50 rupees/month (national average 28 rupees/month) and 15% of parents were illiterate (national average 66%). Nevertheless, 43% of families lived in one room. Hindus were the majority religious group (84%), followed by Sikhs (12%), Christians (2%), Muslims (1%), and Jains (1%). Current study In 1998–2002 we retraced 2,584 (32%) of the cohort and 1,583 agreed to participate. Data on schooling, occupation, household possessions, alcohol consumption, tobacco use, physical activity, and family history were obtained by questionnaire (12,13). Weight and height were measured using standardized techniques. Waist circumference was measured using fiberglass tape, in expiration, midway between the lower lateral costal margin and the iliac crest, with the subject standing. Blood pressure was recorded using an automated device (Omron 711) with the subject seated, after 5 minutes of rest (mean of two readings). Plasma glucose concentrations were measured fasting and 120-min after a 75-g glucose load. Glucose and fasting cholesterol and triglyceride concentrations were analyzed by enzymatic methods using Randox kits on a Beckman AutoAnalyzer, and HDL cholesterol was measured using the same method after phosphotungstate precipitation. IGT and diabetes were defined using World Health Organization criteria (14). Metabolic syndrome was defined using National Cholesterol Education Program Adult Treatment Panel III criteria (15,16). Insulin resistance (homeostasis model assessment [HOMA]) was estimated (17). The study was approved by the All India Institute of Medical Sciences research ethics committee, and informed verbal consent was obtained.

ome was defined using National Cholesterol Education Program Adult Treatment Panel III criteria (15,16). Insulin resistance (homeostasis model assessment [HOMA]) was estimated (17). The study was approved by the All India Institute of Medical Sciences research ethics committee, and informed verbal consent was obtained. Statistical analyses Data from the whole original cohort was used to derive individual SD scores for BMI and weight at 6 months and birthdays from 1 to 21 years (12). Participants had a mean ± SD of 23 ± 5.5 observations. Interpolated values were used if measurements were made within 6 months (up to 1 year), 1 year (aged 2 years), 1.5 years (aged 3 years), and 2 years (all older ages). Back-transformation provided estimates of measurements at these ages. To measure changes in measurements in early life (e.g., from 2 to 11 years), we regressed SD scores at the end of the interval (11 years) on SD scores at the beginning (2 years) and at all preceding time points (birth, 6 months, and 1 year) and expressed the residuals as SD scores. This method produces uncorrelated variables describing change between specific ages (conditional SD scores). Quadratic terms were included when relationships between measurements at different ages were nonlinear. Associations between size in early life and adult outcomes were examined using regression. Outcomes with skewed distributions (HDL cholesterol and insulin resistance) were log-transformed.

es (conditional SD scores). Quadratic terms were included when relationships between measurements at different ages were nonlinear. Associations between size in early life and adult outcomes were examined using regression. Outcomes with skewed distributions (HDL cholesterol and insulin resistance) were log-transformed. RESULTS Of the 1,526 subjects attending the clinic, glucose tolerance category was definable for 1,442 and metabolic syndrome for 1,492. Compared with the remainder of the original cohort, more participants were male (58 vs. 52%, P < 0.001), mean birth weight was heavier (2,851 vs. 2,818 g, P = 0.046), and maternal literacy was 6% higher. BMI SD scores differed by −0.10 to 0.06 (mean −0.04) between birth and 21 years and were statistically significant at 11 and 12 years. The children were short, light, and thin according to an international reference (18), but as adults almost half were overweight or obese (Table 1). BMI (Table 2) and weight at birth and 2 years were positively related to adult waist circumference and inversely related to 120-min glucose concentrations. Eleven-year BMI was positively related to all outcomes except HDL cholesterol and glucose, and adult BMI was positively related to all outcomes except HDL cholesterol, to which it was inversely related. After adjustment for adult BMI, the associations with earlier BMI reversed, becoming significantly inverse for many outcomes. The associations were little changed after further adjustment for adult lifestyle factors (data not shown).

ely related to all outcomes except HDL cholesterol, to which it was inversely related. After adjustment for adult BMI, the associations with earlier BMI reversed, becoming significantly inverse for many outcomes. The associations were little changed after further adjustment for adult lifestyle factors (data not shown). Subjects with high adult waist circumference, triglycerides, blood pressure, and insulin resistance and those with metabolic syndrome had a higher mean BMI than the cohort mean at all ages from birth (Figs. 1 and 2A). The pattern for metabolic syndrome (Fig. 2A) matched that for overweight/obesity (Fig. 2B) but differed from that for IGT/diabetes (Fig. 2C), which was associated with high BMI during childhood and adolescence but a low BMI from 1 to 4 years. Conditional regression analyses showed (Table 3) that greater BMI gain from birth to 2 years was associated with higher adult waist circumference and systolic blood pressure and lower 120-min glucose concentration. Weight gain in infancy was more strongly related to adult risk factors than BMI gain (Table 4, Fig. 2) and showed additional positive associations with triglycerides, insulin resistance, and metabolic syndrome. Greater BMI/weight gain from 2 to 11 years was associated with higher waist circumference, triglycerides, systolic blood pressure, and insulin resistance and a higher risk of IGT/diabetes and metabolic syndrome. Greater BMI/weight gain between 11 years and adulthood was associated with an increase in all risk factors (lower HDL cholesterol).

from 2 to 11 years was associated with higher waist circumference, triglycerides, systolic blood pressure, and insulin resistance and a higher risk of IGT/diabetes and metabolic syndrome. Greater BMI/weight gain between 11 years and adulthood was associated with an increase in all risk factors (lower HDL cholesterol). The inverse association between infant BMI gain and adult IGT/diabetes was stronger in subjects with lower birth weight (OR 0.74 [95% CI 0.58–0.95] for subjects weighing <2,850 g [median] compared with 1.05 [0.81–1.36] for subjects weighing ≥2,850 g, Pinteraction = 0.01). There were no significant interactions at other ages or for other outcomes. Mean ± SD age at adiposity rebound (lowest recorded childhood BMI) was 6.6 ± 1.7 years. Earlier rebound was associated with increased adult metabolic syndrome (P = 0.07) and IGT/diabetes (P = 0.04) and higher waist circumference (P < 0.001), systolic blood pressure (P = 0.052), triglyceride concentration (P = 0.054), and 120-min glucose concentration (P = 0.01). These associations became nonsignificant after adjustment for adult BMI. CONCLUSIONS The Delhi cohort represents an affluent, well-educated section of Indian society that has undergone considerable “transition.” As children they were thin, but as young adults almost half were overweight and 29% had metabolic syndrome. Higher levels of all risk factors except IGT/diabetes were associated with BMI or weight above the average for the cohort as a whole (Figs. 1 and 2) and more rapid BMI or weight gain than the cohort average (Tables 3 and 4) from birth onward.

t as young adults almost half were overweight and 29% had metabolic syndrome. Higher levels of all risk factors except IGT/diabetes were associated with BMI or weight above the average for the cohort as a whole (Figs. 1 and 2) and more rapid BMI or weight gain than the cohort average (Tables 3 and 4) from birth onward. Strengths of the study were that it was population-based and children were measured by trained personnel, with exceptionally frequent follow-up throughout childhood. As with other birth cohorts, there was considerable loss to follow-up and participants are likely to be unrepresentative of the original sample. However, differences in their childhood sizes were small, and in a within-sample analysis, loss to follow-up would introduce bias only if associations between early BMI/weight and later disease differed between those studied and not studied, which seems unlikely given that inclusion was based only on subjects’ availability. Birth Studies in high-income countries have shown increased metabolic syndrome in adults of lower birth weight (19). In Delhi, after adjustment for adult BMI, there were inverse associations with BMI at birth for metabolic syndrome and its components (Table 2), but these resulted from positive associations with childhood BMI gain, not from lower BMI at birth (Table 3). The absence of associations between metabolic syndrome and small size at birth in this population may be due to their young age, low mean birth weight, or different newborn body composition (20).

nts (Table 2), but these resulted from positive associations with childhood BMI gain, not from lower BMI at birth (Table 3). The absence of associations between metabolic syndrome and small size at birth in this population may be due to their young age, low mean birth weight, or different newborn body composition (20). Infancy Consistent with studies of adults in high-income countries (8,9), greater infant BMI/weight gain was associated with a lower risk of diabetes, especially in lower-birth-weight infants. However, it was associated with an increased risk of metabolic syndrome and its components, which is consistent with recent studies showing higher BMI, blood pressure, and insulin concentrations in children who had greater infant weight gain (10,11,21). Understanding these apparently paradoxical findings is important. Effects may differ among populations according to body composition at birth and fat and lean mass accrual during infancy and may vary for different outcomes according to critical periods of development for different tissues. In developing countries, greater infant weight gain is beneficial for survival, growth, and neurocognitive development (22). However, it may become disadvantageous as obesity-related adult chronic diseases emerge (23). The balance of benefits and risks will become clearer as the cohort ages enough to assess cardiovascular disease and mortality. In an intervention study with relevant adult outcomes, protein-energy supplementation in infancy produced no increase in adult cardiovascular risk factors (24).

t chronic diseases emerge (23). The balance of benefits and risks will become clearer as the cohort ages enough to assess cardiovascular disease and mortality. In an intervention study with relevant adult outcomes, protein-energy supplementation in infancy produced no increase in adult cardiovascular risk factors (24). Childhood and adolescence A clear message from our study, consistent with studies in high-income countries, is that rapid BMI gain in childhood and adolescence and earlier adiposity rebound are associated with adult metabolic syndrome and IGT/diabetes. This result probably reflects the known correlation between childhood and adult BMI. Thus, even in underweight children in developing countries, increasing BMI SD scores (“becoming obese relative to oneself”) is a risk factor for later disease. Reinforced by evidence that risk factors in Indian children are already high (25), our study supports efforts to prevent childhood obesity. It also suggests that interventions to control adiposity should be targeted not only to obese children, but also to “normal” weight children with rising BMI SD scores. The original study was funded by the Indian Council of Medical Research and National Institutes of Health (U.S.); the current study was funded by the British Heart Foundation and the Medical Research Council (U.K.).

Childhood and adolescence A clear message from our study, consistent with studies in high-income countries, is that rapid BMI gain in childhood and adolescence and earlier adiposity rebound are associated with adult metabolic syndrome and IGT/diabetes. This result probably reflects the known correlation between childhood and adult BMI. Thus, even in underweight children in developing countries, increasing BMI SD scores (“becoming obese relative to oneself”) is a risk factor for later disease. Reinforced by evidence that risk factors in Indian children are already high (25), our study supports efforts to prevent childhood obesity. It also suggests that interventions to control adiposity should be targeted not only to obese children, but also to “normal” weight children with rising BMI SD scores. The original study was funded by the Indian Council of Medical Research and National Institutes of Health (U.S.); the current study was funded by the British Heart Foundation and the Medical Research Council (U.K.). We thank the participants and field and laboratory staff. We acknowledge Dr. Shanti Ghosh and I.M. Moriyama who initiated the cohort study with Dr. Bhargava, Vinod Kapani for technical input, and Rajeshwari Verma and Bhaskar Singh for maintaining liaison with the cohort.

There is now a pandemic of diabetes and obesity in both developing and developed countries (1,2). Apart from predicting type 2 diabetes and cardiovascular and all-cause mortality (3), metabolic syndrome has also been linked to chronic kidney disease (CKD) (4,5). However, most of these latter studies were either cross-sectional in nature or conducted in general populations. Diabetes and hypertension are the main driving forces for the rising epidemic of CKD and end-stage renal disease (ESRD) (6). Given that components of metabolic syndrome such as hypertension, hyperlipidemia, and obesity are common among subjects with type 2 diabetes, the impacts of clustering of these risk factors in the form of metabolic syndrome on CKD in individuals with type 2 diabetes remain to be established. Furthermore, there is now consistent data showing that Asian diabetic populations, including Chinese, have a higher risk of renal complications than their Caucasian counterparts (7). In a large-scale multinational survey, up to 60% of Asian patients with type 2 diabetes had albuminuria, compared with 30–40% reported in the Western population (8). Against this background, we examined the independent risk associations of metabolic syndrome and its components with CKD in Chinese patients with type 2 diabetes, using the Hong Kong Diabetes Registry.

Furthermore, there is now consistent data showing that Asian diabetic populations, including Chinese, have a higher risk of renal complications than their Caucasian counterparts (7). In a large-scale multinational survey, up to 60% of Asian patients with type 2 diabetes had albuminuria, compared with 30–40% reported in the Western population (8). Against this background, we examined the independent risk associations of metabolic syndrome and its components with CKD in Chinese patients with type 2 diabetes, using the Hong Kong Diabetes Registry. RESEARCH DESIGN AND METHODS The Hong Kong Diabetes Registry was established in 1995 as part of a continuous quality-improvement program at the Prince of Wales Hospital, Hong Kong. Between 1995 and 2005, a total of 7,838 patients from community and hospital clinics, as well as patients newly discharged from the hospital, was enrolled in this Registry. Patients with type 1 diabetes (n = 334) defined as acute presentation with diabetic ketoacidosis, heavy ketonuria (>3+), or a continuous requirement of insulin within 1 year of diagnosis were excluded from this analysis. Patients with estimated glomerular filtration rate (eGFR) <60 ml/min per 1.73 m2 (n = 540) and a preexisting macrovascular complication (n = 1,135) at enrollment were not included. A macrovascular complication was defined as ischemic heart disease, cerebrovascular disease, or peripheral vascular disease. A total of 5,829 Chinese patients with type 2 diabetes were included in the final analysis. All subjects had a comprehensive assessment of risk factors and complications based on the European DiabCare protocol (9). Ethics approval was obtained from the Chinese University of Hong Kong Clinical Research Ethics Committee, and written consent was obtained from all patients for data analysis and research purposes.

All subjects had a comprehensive assessment of risk factors and complications based on the European DiabCare protocol (9). Ethics approval was obtained from the Chinese University of Hong Kong Clinical Research Ethics Committee, and written consent was obtained from all patients for data analysis and research purposes. Details of the clinical assessments and laboratory assays were described previously (10). In brief, sitting blood pressure was measured in both arms using a Dinamap machine after at least 5 min of resting, and the mean value was used for analysis. Waist circumference (to the nearest centimeter) was measured by a plastic tape at the narrowest level between the xiphisternum and umbilicus. Funduscopy was performed by physicians with training in diabetes or ophthalmologists through dilated pupils. Retinopathy was defined by the presence of dot and blot hemorrhages, hard exudates, cotton wool spots, neovascularization, laser scars, or a history of vitrectomy. Sensory neuropathy was diagnosed by two of these three features: reduced sensation to monofilament examination in any part of the sole with normal skin, a score of ≤6 of 8 (aged <65 years) or ≤4 of 8 (aged ≥65 years old) using a graduated tuning fork, or abnormal sensation in lower limbs.

s, or a history of vitrectomy. Sensory neuropathy was diagnosed by two of these three features: reduced sensation to monofilament examination in any part of the sole with normal skin, a score of ≤6 of 8 (aged <65 years) or ≤4 of 8 (aged ≥65 years old) using a graduated tuning fork, or abnormal sensation in lower limbs. Fasting blood samples were taken for measurement of plasma glucose, A1C, lipid profile (total cholesterol, HDL cholesterol, triglycerides, and calculated LDL cholesterol), and renal function. A sterile random spot urine sample was used to measure the albumin-to-creatinine ratio (ACR). Microalbuminuria was defined as either ACR of 2.5–30 mg/mmol in women or 3.5–30 mg/mmol in men. Macroalbuminuria was defined as ACR >30 mg/mmol. All laboratory assays were performed at the Department of Chemical Pathology, the Prince of Wales Hospital, which is accredited by the Royal College of Pathologists of Australasia. Glomerular filtration rate was calculated using the abbreviated equation developed by the Modification of Diet in Renal Disease (MDRD) study with modification for the Chinese population: eGFR = 186 × [SCR × 0.011]−1.154 × [age]−0.203 × [0.742 if female] × 1.233, where SCR is serum creatinine expressed as micromoles per liter (original milligrams per deciliter converted to micromoles per liter) and 1.233 is the adjusting coefficient for Chinese (11).

) study with modification for the Chinese population: eGFR = 186 × [SCR × 0.011]−1.154 × [age]−0.203 × [0.742 if female] × 1.233, where SCR is serum creatinine expressed as micromoles per liter (original milligrams per deciliter converted to micromoles per liter) and 1.233 is the adjusting coefficient for Chinese (11). Metabolic syndrome was diagnosed according to the National Cholesterol Education Program Adult Treatment Panel III criteria with Asian modifications for waist circumferences (12). Because patients already had diabetes, metabolic syndrome was considered to be present if two or more of the following four criteria were met: waist circumference >80 cm in women and >90 cm in men, fasting plasma triglyceride ≥1.7 mmol/l, fasting HDL cholesterol <1.0 mmol/l in men and <1.3 mmol/l in women, and blood pressure ≥130/85 mmHg. Patients who were taking antihypertensive drugs including ACE inhibitors and angiotensin receptor blockers were recorded as having elevated blood pressure. Patients who were using fibrates and/or statins were recorded as fulfilling one lipid criterion.

l in men and <1.3 mmol/l in women, and blood pressure ≥130/85 mmHg. Patients who were taking antihypertensive drugs including ACE inhibitors and angiotensin receptor blockers were recorded as having elevated blood pressure. Patients who were using fibrates and/or statins were recorded as fulfilling one lipid criterion. Clinical outcomes All clinical end points including hospital admissions and mortality were censored on 30 July 2005 using databases from the HA Central Computer System, which records admissions to all public hospitals. These databases, including the Hong Kong Death Registry, were matched by a unique identification number, the Hong Kong Identity Card number, which is compulsory for all residents in Hong Kong and used by all government departments and major organizations. Serum creatinine measurements collected within 6 months of the censored date and hospitalization discharge diagnoses were used to derive eGFR for end point definition. Hospital discharge diagnoses were coded by the ICD-9. Hospitalization with CKD event was defined as 1) hospitalization with a diagnosis of diabetes with renal manifestations (code 250.4), CKD (code 585), or unspecified renal failure (code 586) or 2) dialysis (ICD-9 procedure code 39.95) or peritoneal dialysis (ICD-9 procedure code 54.98). In this study, the end point was defined as the first eGFR <60 ml/min per 1.73 m2 or the first hospitalization with CKD event.

with renal manifestations (code 250.4), CKD (code 585), or unspecified renal failure (code 586) or 2) dialysis (ICD-9 procedure code 39.95) or peritoneal dialysis (ICD-9 procedure code 54.98). In this study, the end point was defined as the first eGFR <60 ml/min per 1.73 m2 or the first hospitalization with CKD event. Statistical analysis All data are means ± SD or median (interquartile range). For between-group comparisons, the χ2 test was used for categorical variables, and the Student's t test or ANOVA was used for continuous variables. Cox regression analysis was applied by the backward stepwise method with CKD as the dependent variable. The metabolic syndrome (or its components) and other known risk factors for CKD were included in the model. Hazard ratios (HRs) with 95% CIs were calculated. Before analysis, the skewed distribution of ACR and triglyceride was normalized by logarithmic transformation. A Kaplan-Meier curve was used to show the risk relationship between patients with and without the metabolic syndrome and patients with different components of the metabolic syndrome. P (two sided) <0.05 was considered to be significant. Statistical analysis was performed using the Statistical Package for Social Science (version 13.0 for Windows; SPSS, Chicago, IL).

relationship between patients with and without the metabolic syndrome and patients with different components of the metabolic syndrome. P (two sided) <0.05 was considered to be significant. Statistical analysis was performed using the Statistical Package for Social Science (version 13.0 for Windows; SPSS, Chicago, IL). RESULTS The mean ± SD age of the cohort was 54.1 ± 13.0 years, and 45.5% were men. The mean duration of diabetes was 6.23 ± 6.17 years. The frequency of metabolic syndrome was 51.2% (n = 2,985). After a median observation period of 4.6 years (interquartile range 1.9–7.3), 741 patients (12.7%) developed CKD. Table 1 compares the baseline demographic, clinical, and biochemical characteristics of patients with and without metabolic syndrome. Patients in the former group were older and had a longer duration of diabetes and worse glycemic control. In addition, compared with the group without metabolic syndrome, those with metabolic syndrome had twice the frequency of albuminuria and had a lower eGFR at baseline.