Browse the corpus

The provision, administration, and evaluation of contemporary health care services generate an enormous quantity of accessible data for research. Access to “big data” by clinicians and researchers increases every year, with tools such as electronic health records (EHRs) and open-access publication of research data and findings gaining rapid availability.1

aluation of contemporary health care services generate an enormous quantity of accessible data for research. Access to “big data” by clinicians and researchers increases every year, with tools such as electronic health records (EHRs) and open-access publication of research data and findings gaining rapid availability.1 Big data in health care often refer to EHR databases, patient registries, and administrative claims, among others. These data sources are also used in a variety of observational research study designs and clinical specialty areas. Research using these data may be undertaken by medical students or residents, junior investigators, or experienced research teams with varying levels of experience and knowledge using these types of data and study design approaches. This “consumer guide” will prepare readers to navigate and interpret research using observational data while providing explanations and context for related terminology. A second objective of this guide is to help the reader, either a junior investigator or an experienced clinician, critique studies in obstetrics and gynecology (OB/GYN) that use observational data, which will be accomplished by outlining how to assess common pitfalls of quasi-experimental study designs that use observational data. The guide then demonstrates the application of this knowledge using a hypothetical case study of an OB/GYN research article using observational data and a quasi-experimental, or nonrandomized, design. Finally, a compendium of observational data resources commonly used within OB/GYN research is provided.

rvational data. The guide then demonstrates the application of this knowledge using a hypothetical case study of an OB/GYN research article using observational data and a quasi-experimental, or nonrandomized, design. Finally, a compendium of observational data resources commonly used within OB/GYN research is provided. METHODS Terminology and definitions relevant to observational data research were identified via literature and keyword searches for terms using PubMed/MEDLINE and Google Scholar. Examples of the relevant identified terminology were based on reviews of the literature in OB/GYN research–focused journals. Categories of observational data were defined and reported in summary tables, with examples of studies using each type of data. Additional Web searches were conducted to identify data resources, and the identified government-sponsored data collection agencies were cross-referenced to find further data resources. Researcher recommendations were also used to gather observational data sources, along with review of current literature for observational studies or secondary data analysis in OB/GYN-focused journals. The contents of each data resource were summarized, and the cost and accessibility of each resource were verified with the data provider or government data collection agency. A flowchart was constructed to allow the user to quickly interpret the study design for a quasi-experimental study design using observational data.

e contents of each data resource were summarized, and the cost and accessibility of each resource were verified with the data provider or government data collection agency. A flowchart was constructed to allow the user to quickly interpret the study design for a quasi-experimental study design using observational data. RESULTS Review of the literature resulted in the identification of 22 frequently used terms relevant to interpreting observational data research (Table 1). These terms will be used throughout the guide. TABLE 1 Useful Definitions in Observational Research Selecting the Appropriate Study Design The reader should use this section to aid in distinguishing quasi-experimental designs using observational data from randomized controlled trials (RCTs) or other experimental designs. Experimental studies, such as RCTs, rely on an investigator-assigned exposure, whereas an observational study collects or reviews data from an extant phenomenon or occurrence; that is, the exposure is assigned “naturally” by clinical decision-making processes, policy changes, and so on (Fig. 1). Figure 2 provides the reader with a “checklist” for quickly interpreting observational data research and will be consulted in later discussion to aid in the critique of an example study abstract.

ce; that is, the exposure is assigned “naturally” by clinical decision-making processes, policy changes, and so on (Fig. 1). Figure 2 provides the reader with a “checklist” for quickly interpreting observational data research and will be consulted in later discussion to aid in the critique of an example study abstract. FIG. 1 Classifying study design. *If the investigator assigns the exposure, as in the case of a nonrandomized treatment versus control group study, then the study design would be classified as having an investigator-assigned quasi-experimental design. If the investigator does not assign the exposure in a treatment versus control group study, then the study design would be classified as an observational quasi-experimental design. An example of this is natural experiments, where some change like a new clinical guideline is enacted and the investigator compares the outcome on groups affected differently after the change. FIG. 2 A 4-step approach for identifying and interpreting observational studies.

FIG. 1 Classifying study design. *If the investigator assigns the exposure, as in the case of a nonrandomized treatment versus control group study, then the study design would be classified as having an investigator-assigned quasi-experimental design. If the investigator does not assign the exposure in a treatment versus control group study, then the study design would be classified as an observational quasi-experimental design. An example of this is natural experiments, where some change like a new clinical guideline is enacted and the investigator compares the outcome on groups affected differently after the change. FIG. 2 A 4-step approach for identifying and interpreting observational studies. For obstetrician-gynecologists, primary data collection and the use of RCTs are more appropriate for answering research questions such as determining whether a newly developed treatment or medication has the intended effect on patients. This type of research question requires a demonstration of causality, which can be more strongly inferred from RCTs because of the use of randomization, and is most often referred as an evaluation of “efficacy,” or answering the question of “Can this work?” Randomization of the “exposure” (for this type of research question, the exposure is the treatment or medication) reduces or eliminates the possibility of selection bias when determining what groups in the study population receive the treatment and what groups are used for comparison (eg, no treatment, placebo, or alternative treatment). This process of randomization allows the investigator to rule out observed effects for reasons other than the treatment. Randomization is thus a crucial component in the investigation of new medications and medical devices, which is why the US Food and Drug Administration requires the use of RCT study designs during the approval process. Randomized controlled trials maximize internal validity, that is, lack selection bias, but have limited external validity due to strict inclusion/exclusion criteria that limit the most sick patients and patients at extremes of age (very young or very old) and operate in a clinical environment that is not comparable to everyday practice.42 As such, many research questions can be left unanswered in underrepresented or disparate patient groups and must be answered using “real-world” data and observational study designs.

patients at extremes of age (very young or very old) and operate in a clinical environment that is not comparable to everyday practice.42 As such, many research questions can be left unanswered in underrepresented or disparate patient groups and must be answered using “real-world” data and observational study designs. Obstetrician-gynecologists also encounter research questions that may be preferable to investigate using observational study designs or secondary data analysis. Observational data are generated in the real world either through active data collection for research purposes or passive archiving of data for administrative purposes that can be reused. Real-world data are the product of everyday medical practice and not tightly controlled like that of an RCT. Patient medical treatments are part of a complex decision-making process and are nonrandomized. Patients will receive treatments based on evaluation of risk factors, comorbid conditions, medical history, behavioral aspects, and other factors. Thus, there are significant threats to validity (ie, confounding and bias), associated with patient treatment and outcomes that have to be factored into study design and statistical analysis. However, observational data have strong external validity in the population from which they were generated and offer a relatively inexpensive and time-efficient means to conduct clinical research.

bias), associated with patient treatment and outcomes that have to be factored into study design and statistical analysis. However, observational data have strong external validity in the population from which they were generated and offer a relatively inexpensive and time-efficient means to conduct clinical research. For example, it would be impractical, expensive, and potentially impossible to calculate the prevalence of cervical cancer in the general population by performing a cervical cancer screening and medical chart review for every female. We can, however, estimate cervical cancer prevalence using a variety of observational study design methods and data sources, such as querying patient registries, isolating claims for cervical cancer screenings and treatments in administrative data, surveying patients or clinicians or health systems, or conducting a secondary data analysis of reported surveillance data from government agencies. As another example, an institutional review board would be hesitant to allow an RCT to compare birth defects caused by antidepressant use in pregnant women because of ethical and liability concerns. However, pregnant women often receive antidepressants in routine clinical practice and could be evaluated using observational data from a claims database linked to vital statistics. Observational data could then allow for a simultaneous comparison of all antidepressants, and active comparator treatment groups could be factored in to provide stronger evidence. Thus, confounding and bias concerns could be mitigated, for example, by strong study design and/or statistical adjustments using techniques such as regression, propensity score matching, or weighting. Such studies could be conducted in a timely manner using existing resources and without posing direct risk to patients.

Thus, confounding and bias concerns could be mitigated, for example, by strong study design and/or statistical adjustments using techniques such as regression, propensity score matching, or weighting. Such studies could be conducted in a timely manner using existing resources and without posing direct risk to patients. Types of Observational Data Seven types of observational data categories were identified from literature searches (Table 2), and each has several strengths and limitations. The summary definitions for these data types (surveys, admissions/discharge data sets, administrative claims data, registries, surveillance data, electronic medical records [EMRs], and linked data sets) are found in Table 2, with further discussion below. The strengths and limitations with the measurement of exposures, outcomes, and other explanatory factors are discussed in the following sections for each of the 7 types of observational data categories. TABLE 2 Types of Observational Data

Types of Observational Data Seven types of observational data categories were identified from literature searches (Table 2), and each has several strengths and limitations. The summary definitions for these data types (surveys, admissions/discharge data sets, administrative claims data, registries, surveillance data, electronic medical records [EMRs], and linked data sets) are found in Table 2, with further discussion below. The strengths and limitations with the measurement of exposures, outcomes, and other explanatory factors are discussed in the following sections for each of the 7 types of observational data categories. TABLE 2 Types of Observational Data Surveys Surveys constitute an entire family of investigatory tools that sample individual patients, clinicians, or health systems using questionnaires or interviews about a topic defined by the investigator. Sampling strategy, survey instrument design, respondent participation rates, and level of reporting subjectivity are all critical elements of interpreting observational study designs that use surveys. For these reasons, surveys tend to be one of the more complicated types of observational data to collect and interpret, and the resulting data are subject to varying levels of both validity and reliability. Secondary data analysis using publicly available survey data sources (eg, see National Ambulatory Medical Care Survey, National Survey on Drug Use and Health, and National Health Interview Survey in Supplemental Table 1, Supplemental Digital Content, http://links.lww.com/OBGYNSURV/A28) tends to be a better choice for OB/GYN investigators seeking to analyze survey data that were collected in a methodologically rigorous manner. These surveys are often weighted so that survey respondents, once the weights are applied in the statistical analysis, are representative of the general population. Survey data such as those collected by various government agencies are collected yearly and also provide a useful source of data to track trends across time in the whole US population. These publicly available sources of survey data, however, come with limitations. None of the examples above were collected explicitly in populations relevant to OB/GYN research, although gender, sex, pregnancy status, and type/frequency of medical visits are available in each of these data resources. These surveys may not be appropriate for rare conditions or very restrictive patient populations because these may not have adequate sample sizes for study even with survey weights applied. In addition, survey responses may rely on unreliable self-reports and may not include data points relevant to your research question.

se surveys may not be appropriate for rare conditions or very restrictive patient populations because these may not have adequate sample sizes for study even with survey weights applied. In addition, survey responses may rely on unreliable self-reports and may not include data points relevant to your research question. Administrative Claims Data In recent years, large health care administrative claims databases have been widely accepted for research in OB/GYN. Administrative data are derived from automated electronic recording of encounters with the health care system, including physician's office visits, hospitalizations, filled prescriptions, and diagnostic procedures, which are collected for administrative or billing purposes.46,55 These databases are readily available and relatively inexpensive to purchase/access and contain clinical information coded using accepted coding systems. They are also representative of routine clinical care and large populations.55,56 Claims data are usually generated from a single insured population (eg, Medicare, Medicaid, or private insurance), and the generalizability or external validity of research findings outside these discrete populations must be considered. Because administrative claims data are generated primarily for payment and administration of health services, these data are subject to inherent limitations because they are not collected for research purposes. For example, most administrative databases contain limited clinical information.56 Physiological measurements such as blood pressure and glucose levels, tumor stage, and so on are generally not captured. Furthermore, diagnosis and procedure information may not be comprehensive enough and have low sensitivity and specificity for the condition that are meant to capture, and certain chronic conditions could be underdiagnosed. In addition, the reliability of information in administrative databases largely depends on its impact on payment and can differ between fee-for-service or capitated insurance benefits.46 Treatment and services may not be reflected in the database if they are not covered by health insurance, such as elective cosmetic procedures or over-the-counter medications, or are simply paid in cash by the patient or administered for free (eg, vaccines or $4 generic prescriptions).

or capitated insurance benefits.46 Treatment and services may not be reflected in the database if they are not covered by health insurance, such as elective cosmetic procedures or over-the-counter medications, or are simply paid in cash by the patient or administered for free (eg, vaccines or $4 generic prescriptions). The practice of “up-coding” of patients in order to increase billing has been described and can be a major limitation to research, as well as constituting fraud.57 When evaluating a study using administrative databases, it is imperative to recognize the strengths and the limitations of these data.

or capitated insurance benefits.46 Treatment and services may not be reflected in the database if they are not covered by health insurance, such as elective cosmetic procedures or over-the-counter medications, or are simply paid in cash by the patient or administered for free (eg, vaccines or $4 generic prescriptions). The practice of “up-coding” of patients in order to increase billing has been described and can be a major limitation to research, as well as constituting fraud.57 When evaluating a study using administrative databases, it is imperative to recognize the strengths and the limitations of these data. Despite its limitations, administrative claims databases have a wide spectrum of research applications. Using administrative claims data allows for the study of large patient populations, including pregnant women and infants/children, who tend to be partially or totally excluded from clinical trials.55 Specifically, claims can be used to (1) examine drug utilization patterns (eg, the prevalence and duration of medication use during pregnancy), (2) study rare events (eg, breast cancers and congenital malformations), (3) evaluate appropriate prescribing (eg, antibiotic use and misuse among infants), and (4) link to other resources to expand the use of information in those databases (eg, linkage to birth certificate data and the National Death Index). Claims data are becoming more widely used to conduct comparative effectiveness and safety research, which generally aims to compare alternative treatments where RCTs may never be performed or may be unfeasible. The US Food and Drug Administration's Sentinel Initiative is such an effort, developed as an active surveillance network built using a consortium of claims data providers. Claims data can be supplemented by linked data or be used to supplement RCTs or patient registries, which could provide substantial number data fields for further study or could extend the follow-up period beyond what was originally defined for a clinical study because of budgetary or other constraints.

data providers. Claims data can be supplemented by linked data or be used to supplement RCTs or patient registries, which could provide substantial number data fields for further study or could extend the follow-up period beyond what was originally defined for a clinical study because of budgetary or other constraints. Admissions/Discharge Data Sets Health system admission or discharge data sets typically include diagnoses, procedures, and patient demographic characteristics from the hospital or health care facility in which the data were collected. These data are similar in format and content to administrative claims data and are thus subject to similar strengths and limitations. Discharge data are typically a 1-line, cross-sectional record of a hospitalization and are agnostic to when diagnoses and procedures occurred during the hospitalization. One important difference from administrative claims data is that admissions/discharges may include patients who have 1 or more types of insurance (or who have no insurance), which is typically not captured in administrative claims data that are derived from a single-payer source. Obstetrics and gynecology research that uses discharge data has found that some recorded procedures and diagnoses were highly accurate upon validation study for some conditions, but not as accurate for others.45 The most commonly used discharge database, the National Inpatient Sample, is a random sample of hospitalizations that is weighted to be nationally representative but contains no unique patient-identifying identification variable. Thus, such results must be interpreted with caution as the unit of analysis is a “hospitalization” rather than a single patient, and some patients may be double counted. Such data are reliable for evaluating trends in utilization or costs and have been used to evaluate variation in care quality. Newer data resources, such as the National Readmissions Database, include patient identifiers and would allow one to follow individuals longitudinally in the data to evaluate 30-day readmissions, mortality, or other outcomes.

ating trends in utilization or costs and have been used to evaluate variation in care quality. Newer data resources, such as the National Readmissions Database, include patient identifiers and would allow one to follow individuals longitudinally in the data to evaluate 30-day readmissions, mortality, or other outcomes. Registries Registries are popular tools in OB/GYN research and are built to track a certain type of condition, use of medication/device, or patient within participating health systems. Registries are often useful for establishing counts of observed adverse effects and outcomes and providing preliminary evidence for change over time in prevalence or incidence. One limitation is that registries tend to have voluntary enrollment, which means that participants might be inherently different from nonparticipants. Another is that registries are typically maintained by institutions with limited geographic penetration (eg, a single health system or professional organization). Surveillance Data Surveillance data are typically collected by a government or nonprofit agency to monitor a condition or attribute that is relevant to public health. Because a particular condition or attribute is the focus of data collection, an obstetrician-gynecologist may encounter limitations in using these data to answer research questions that require access to detailed clinical data. In addition, these data sources tend to be deidentified, which makes longitudinal tracking and data linkages difficult or impossible.

te is the focus of data collection, an obstetrician-gynecologist may encounter limitations in using these data to answer research questions that require access to detailed clinical data. In addition, these data sources tend to be deidentified, which makes longitudinal tracking and data linkages difficult or impossible. Electronic Medical Records Many obstetrician-gynecologists have access to EMRs (or EHRs) used by their respective health care institutions. These data contain detailed clinical information and may also contain extended follow-ups and full medical histories. This depth of information is very valuable to OB/GYN research, but the use of these data for research purposes requires substantial review and oversight by your institutional review board and may require substantial time and staffing resources to filter and condense into usable data for analysis. In addition, the EMRs of a single health system are limited to the geographic area and population serviced by the health system, as well as the local policies and practices of the hospital and physicians, which limits the generalizability of study results to the population as a whole. Unless part of a fully integrated health system (eg, Kaiser Permanente or the Veterans Affairs system), patient visits to other physician offices or admissions to other hospitals may not be captured in the medical record. Development of a medical record that follows a patient, or is easily accessible from a cloud-based system, is an eagerly anticipated development for researchers looking to use medical records because this will further centralize the data for these purposes.

to other hospitals may not be captured in the medical record. Development of a medical record that follows a patient, or is easily accessible from a cloud-based system, is an eagerly anticipated development for researchers looking to use medical records because this will further centralize the data for these purposes. Linked Data Sets Data linkages serve an important purpose in OB/GYN research because this is the primary tool for tracking females and progeny within single data sets and across multiple data sources. For example, observational study designs that require the linkage of state vital statistics data (birth certificates and death certificates) are a critical tool for calculating maternal and neonate mortality and morbidity. Use of a common identifier is required to conduct a data linkage, which is why data linkages are difficult to conduct using publicly available data sources (these data are typically deidentified). However, some publicly available observational data resources may conduct a linkage by requiring the investigator to use the services of an honest-broker third party who conducts the linkage and then removes all identifying data. Such policies will vary significantly based on government or institutional policies.

ied). However, some publicly available observational data resources may conduct a linkage by requiring the investigator to use the services of an honest-broker third party who conducts the linkage and then removes all identifying data. Such policies will vary significantly based on government or institutional policies. Observational Data Resources Supplemental Table 1 (Supplemental Digital Content, http://links.lww.com/OBGYNSURV/A28) contains a summary of observational data resources that are commonly used in OB/GYN research. This list is not meant to be exhaustive, and in some cases, a single prominent example was included for brevity (eg, commercial insurance claims, where a variety of vendors sell or allow access to different sets of administrative claims from commercial insurers). DISCUSSION Many researchers are engaged in translating big data resources into actionable benefit for health care quality and effectiveness. One study using the UK National Health Service administrative data demonstrated that observational study designs using “big data” could effectively be applied to what the authors called “small data” (eg, single-institution administrative data).58 In the following sections, additional pitfalls and benefits of observational data research are discussed, and then a case study is examined so that the reader may apply the terminology and interpretation of study design to evaluating an OB/GYN study abstract.

“small data” (eg, single-institution administrative data).58 In the following sections, additional pitfalls and benefits of observational data research are discussed, and then a case study is examined so that the reader may apply the terminology and interpretation of study design to evaluating an OB/GYN study abstract. Additional Pitfalls of Observational Data Research There are several elements of observational research that we have yet to discuss. While the potential power of big data to disseminate findings or to investigate research questions with larger population sizes is apparent, many urge caution regarding the pitfalls of conducting such research using health care data. Namely, large-scale collection and “data mining” (ie, a series of statistical techniques for filtering large databases to isolate a characteristic or trend of interest sometimes in the absence of a research hypothesis) practices with health care data may pose risks related to causal inference and patient privacy, and/or findings may be susceptible to misinterpretation if inappropriate or unfamiliar study designs or statistical techniques are applied.59 For many, the latter is sometimes referred to as the “black box” and is made easy by modern statistical software. In this section, we examine some additional pitfalls of working with observational data, namely, interpreting the relevance of effect size (eg, clinical vs statistical significance), the technology constraints associated with big data research, and the use of data designed for purposes other than research.

al software. In this section, we examine some additional pitfalls of working with observational data, namely, interpreting the relevance of effect size (eg, clinical vs statistical significance), the technology constraints associated with big data research, and the use of data designed for purposes other than research. Effect Size: Interpreting Clinical Versus Statistical Significance It is important for OB/GYN professionals to know how to interpret and implement results from observational data in their clinical decision making. Observational studies rely on statistical analysis, which simply address the acceptance or rejection of statistical hypothesis. Statistically significant differences are determined by the generally accepted level of probability (the “P level”), and P values provide guidance with respect to random error due to sampling rather than a true change of outcome between study groups. Readers should consider several assumptions when interpreting statistical significances, such as sample size or distribution of the population. The power to detect a given effect size, and observe a statistically significant result, is generally a product of sample size and variation. With larger samples available in big data, the likelihood of statistical significance is high. This means that nonchance differences will be reported as “statistically significant” in big data studies, but those differences may be so small that the effect size, or practical difference, has little value for application to clinical practice. In other words, the observed statistical significance from an observational study alone is insufficient to apply in the clinical practice and may be susceptible to misuse by health care providers.

ences may be so small that the effect size, or practical difference, has little value for application to clinical practice. In other words, the observed statistical significance from an observational study alone is insufficient to apply in the clinical practice and may be susceptible to misuse by health care providers. The results of an observational study can be statistically significant, but too minimal to be clinically important. However, there is no standard or defined understanding of clinically relevant changes. “Clinical significance” is generally defined as the smallest meaningful change in an observed effect, but this does not establish a standard effect size to be considered as worthwhile by the practitioner to enact a change in clinical care.60,61 There is a large degree of subjectivity in the judgment of clinical significance that could direct the course of patient care because of disparities in patients' characteristics, differences across health care fields, clinical experience, and goals of clinical care. The aware reader should not focus on measures of significance presented in research articles but should evaluate the effect size, and its corresponding confidence intervals, to determine if a given effect size is enough to elicit a change in clinical practice. Furthermore, readers should expect both relative effect sizes and absolute effect sizes from research in order to fully appreciate the magnitude of the effect in question.

the effect size, and its corresponding confidence intervals, to determine if a given effect size is enough to elicit a change in clinical practice. Furthermore, readers should expect both relative effect sizes and absolute effect sizes from research in order to fully appreciate the magnitude of the effect in question. Let us examine a hypothetical example for interpreting effect size and statistical significance. A big data observational study finds a statistically significant effect of an intervention to improve patient preventive care. The intervention in our example is a change in insurance coverage, which was enacted by a large managed care organization to fully cover and encourage annual Papanicolaou (Pap) smear examinations. Administrative claims data were reviewed before and after the change in insurance coverage for all adult female patients in the large managed care organization with samples of several thousand patients receiving examinations in a 5-year period prior to the change in coverage and several thousand additional patients receiving examinations in the 5-year period following the change in insurance coverage. The study reported 2 major findings, namely, a P value that demonstrated a statistically significant difference in the proportion of patients in the managed care organization receiving examinations following the insurance change and an odds ratio with confidence intervals for the likelihood of patients receiving abnormal Pap test results following the coverage change.

alue that demonstrated a statistically significant difference in the proportion of patients in the managed care organization receiving examinations following the insurance change and an odds ratio with confidence intervals for the likelihood of patients receiving abnormal Pap test results following the coverage change. These 2 statistical reports represent very different findings with varying applicability to clinical practice. First, the P value demonstrating a statistically significant difference in the proportion of patients who received Pap tests following the change does indicate that more patients received the tests, but does not indicate how many more patients (effect size) received those tests. This P value has limited clinical applicability because it demonstrates neither the size of the intervention effect nor the desirability of that effect (eg, more frequent Pap testing could potentially result in improved patient outcomes in high-risk patients, but what about low-risk patients? Is more frequent testing worthwhile for the whole population?). Second, let us suppose that the odds ratio in our example was reported as follows: the odds of an abnormal Pap test result following the insurance change was 1.03 (95% confidence interval, 1.01–1.05), which means that abnormal test results were found 0.03 times more often following the insurance change. These findings do give the reader information about the effect size of the intervention (ie, the odds of getting abnormal test results went up by 0.03 times, or 3%, following the insurance change); however, this result still lacks clinical context. We know that a greater proportion of patients received annual testing following the insurance coverage change, but we do not know if the additional testing resulted in more false-positives (also known as type I errors) or a true capture of additional patients with abnormal test results. It is also unclear whether a 3% change in effect size is large enough to warrant change in practice—this would remain a clinical decision.

nge, but we do not know if the additional testing resulted in more false-positives (also known as type I errors) or a true capture of additional patients with abnormal test results. It is also unclear whether a 3% change in effect size is large enough to warrant change in practice—this would remain a clinical decision. This example brings to light another consideration for working with big data and statistics, that of type I error, or false-positives. There are numerous methods for safeguarding against type I errors while working with exploratory statistical testing, so best practices would indicate that a statistician or methodologist be consulted to ensure that the selected study design and analysis methods are appropriate for the data available to the researcher.

There are numerous methods for safeguarding against type I errors while working with exploratory statistical testing, so best practices would indicate that a statistician or methodologist be consulted to ensure that the selected study design and analysis methods are appropriate for the data available to the researcher. Technology Constraints and Big Data “Big data” is aptly named, in part, because many sources of observational big data are dramatically larger in sample size than what is typically used in experimental designs (RCTs). For example, a typical RCT conducted in OB/GYN research may enroll a study population of 200 female patients, whereas an extract of administrative claims from a large commercial or public insurer data set may contain millions of female patients who fit the study inclusion criteria with tens of millions of individual medical or pharmacy billing records. A data set of this size is not easily stored or analyzed using conventional computer hardware and software. A large data set (or database) may require the use of storage servers, querying software, sophisticated statistical analysis software packages, and the assistance of database administrators and/or statisticians to appropriately manage and analyze data of that magnitude. Access to the infrastructure required to work with big data may be cost or time prohibitive for an obstetrician-gynecologist practicing outside a research institution environment with established resources. Other data sources mentioned, such as surveys or discharge data, can be managed on local personal computers and are amenable to commonly used software packages that are not technically restrictive (eg, SPSS).

an obstetrician-gynecologist practicing outside a research institution environment with established resources. Other data sources mentioned, such as surveys or discharge data, can be managed on local personal computers and are amenable to commonly used software packages that are not technically restrictive (eg, SPSS). Additional Benefits of Observational Data Research In this section, we discuss some additional benefits of working with observational data. First, the methods and procedures typically used in secondary data analysis tend to be highly reproducible. This is especially true when working with publicly available data and when scholarly publications require thorough reporting of the research protocols used by observational study designs. As an obstetrician-gynecologist engaged in research, reproducibility is of tantamount importance when confirming or dismissing emerging evidence. Efforts by several independent groups and scientific societies have sought to standardize reporting of observational studies. Readers should be familiar with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines as a checklist for reporting observational study methods.62 Other resources are available for evaluating research using administrative claims data specifically.63

orting of observational studies. Readers should be familiar with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines as a checklist for reporting observational study methods.62 Other resources are available for evaluating research using administrative claims data specifically.63 Next, the use of extant data sources is a cost-effective strategy for engaging in research activities. A data source that is both publicly available and kept updated with contemporary entries (eg, annual surveillance data collected by government agencies) is of high value, but has a low cost of access for the investigator. In some cases, the purchase of extant data sources (eg, commercial insurance administrative claims) may be a more cost-effective means to answering research question than collecting primary data for an RCT, which tend to be expensive even when conducted in small populations. Observational data sometimes can allow for research questions to be evaluated that are deemed unethical or risky for approval of a clinical trial but have occurred in the real world (eg, birth defects associated with antipsychotic medications).64

RCT, which tend to be expensive even when conducted in small populations. Observational data sometimes can allow for research questions to be evaluated that are deemed unethical or risky for approval of a clinical trial but have occurred in the real world (eg, birth defects associated with antipsychotic medications).64 Some observational data resources also allow for the use of large population sizes, which are desirable for the high statistical power that reduces uncertainty and produces more precise estimates of observed phenomenon. A well-designed, highly powered “large N” observational study may produce evidence of greater value than an underpowered “small N” RCT in some cases, particularly among populations where it is difficult to conduct RCTs (eg, pregnant women). To reiterate, however, readers should remain aware of statistical versus clinical significance in studies with large sample sizes. Case Study Application: Evaluating an Abstract That Uses Observational Data Now that the reader is familiar with terminology, definitions, and evaluation strategies for interpreting observational data research, we will apply the strategies discussed throughout this guide by critically evaluating the following study abstract relevant to obstetrician-gynecologists and other clinicians who treat women and/or neonates. The abstract uses observational data to calculate the incidence of neonatal abstinence syndrome and quantifies access to care differences for pregnant women seeking opioid abuse treatment:

valuating the following study abstract relevant to obstetrician-gynecologists and other clinicians who treat women and/or neonates. The abstract uses observational data to calculate the incidence of neonatal abstinence syndrome and quantifies access to care differences for pregnant women seeking opioid abuse treatment: Objective: Incidence of neonatal abstinence syndrome (NAS) is increasing because of the rise in opioid use. Rural states such as Kentucky have been disproportionally impacted by opioid abuse, and this study determines NAS burden nationally and in Kentucky while quantifying differences in access to care between Appalachian and non-Appalachian counties. Methods: Neonatal abstinence syndrome rates were calculated using national (2013) and Kentucky (2008–2014) National Inpatient Sample discharge data. Births were identified using the International Classification of Diseases, Ninth Revision code 779.5 and live birth codes V30.x–V38.x. Counties were classified as rural, micropolitan, or metropolitan using census data. Proximity analysis was conducted via mapping from ZIP code centroid to nearest opioid treatment facility. Distance to treatment facilities was calculated and then compared using nonparametric testing for counties by rural and Appalachian status.

were classified as rural, micropolitan, or metropolitan using census data. Proximity analysis was conducted via mapping from ZIP code centroid to nearest opioid treatment facility. Distance to treatment facilities was calculated and then compared using nonparametric testing for counties by rural and Appalachian status. Results: Neonatal abstinence syndrome cases tripled from 2008 to 2014 in Kentucky counties, with a 2013 NAS rate more than double the national NAS rate. Rural and Appalachian counties experienced an NAS increase per 1000 births that was 2 to 2.5 times higher than urban/non-Appalachian counties, with a greater number of NAS births overall in Appalachian counties. All opioid treatment facility types were farther from rural patients than micropolitan/metropolitan patients (P < 0.001), as well as farther for Appalachians versus non-Appalachians (P < 0.001, all facility types). Conclusions: Neonatal abstinence syndrome burden disparately affects rural and Appalachian Kentucky counties, whereas treatment options are disproportionately farther away for these residents. Policy efforts to increase NAS prevention and encourage opioid abuse treatment uptake in pregnant women should address rural and Appalachian disparities. Note: This abstract is reprinted with permission from the Journal of Rural Health (publisher: John Wiley and Sons, Inc; license no. 4151490014858).65

Conclusions: Neonatal abstinence syndrome burden disparately affects rural and Appalachian Kentucky counties, whereas treatment options are disproportionately farther away for these residents. Policy efforts to increase NAS prevention and encourage opioid abuse treatment uptake in pregnant women should address rural and Appalachian disparities. Note: This abstract is reprinted with permission from the Journal of Rural Health (publisher: John Wiley and Sons, Inc; license no. 4151490014858).65 Using Figures 1 and 2 as a guide, we have enough information from the study abstract to quickly dissect and interpret the study design and findings. First, we can note that the researchers did not assign an exposure, and subjects were not randomized into the study, which means that the study design is nonexperimental (or quasi-experimental). In addition, the study is retrospective and contains 2 distinct research objectives: to calculate incidence of neonatal abstinence syndrome in births and to quantify access to care for opioid abuse among pregnant women in a particular state. The design used to address the 2 research objectives would be best described as a descriptive, cross-sectional design, because of the lack of comparison or control groups and lack of longitudinal follow-up data of mothers or neonates and because the data source lacks individual patient identifiers.

a particular state. The design used to address the 2 research objectives would be best described as a descriptive, cross-sectional design, because of the lack of comparison or control groups and lack of longitudinal follow-up data of mothers or neonates and because the data source lacks individual patient identifiers. Next, the reader should note that administrative discharge data from the state of Kentucky and the National Inpatient Sample (see Supplemental Table 1, Supplemental Digital Content, http://links.lww.com/OBGYNSURV/A28, entry 12 for additional information about this data source from the Healthcare Cost Utilization Project family of data sets) were used in this article. National data are randomly sampled and would be nationally representative once these weights are applied, which can be used to generate estimates of neonatal abstinence syndrome for the entire country. However, these estimates cannot be calculated by state because of the way data are sampled. Kentucky data are complete records for the state, although Kentucky has been hit particularly hard by the opioid abuse epidemic, is rural, and has a distinct Appalachian disparate region and thus may not be comparable to other states. Noting the use of discharge data to define the patient population and to isolate diagnoses using billing codes is important, because diagnosis codes may not accurately capture all cases of pregnancy, substance abuse, and neonatal abstinence syndrome in these populations. Neonatal abstinence syndrome could be misdiagnosed, could be not caused by opioids (as was the focus of that research), and could vary in severity, which is not discernible from a single International Classification of Diseases, Ninth Revision diagnosis code. The aware reader should evaluate whether the authors have supported their choice of coding algorithms used either with prior research or with clear acknowledgement of the potential limitations of the research.

hich is not discernible from a single International Classification of Diseases, Ninth Revision diagnosis code. The aware reader should evaluate whether the authors have supported their choice of coding algorithms used either with prior research or with clear acknowledgement of the potential limitations of the research. The abstract findings report both incidence rates and P values for differences in distance to opioid abuse care facilities. However, the abstract's method section provides limited information about the statistical analysis tools used to calculate either of these findings in the study population, which could be cause for concern. Without these details, it is difficult to assess if the study is appropriately powered and whether the results are clinically meaningful (eg, how much higher would neonatal abstinence syndrome incidence in the study location [Kentucky] need to be compared to nationwide incidence to convince me it is a real problem?). Other results in this article show that women in rural areas would be on average 20 miles farther away from pregnancy-specific treatment centers than women in metropolitan areas, a statistically significant finding. While 20 miles may be a burdensome drive, is there evidence that this could affect treatment-seeking behaviors? Is it enough to enact policy changes from state or local government? In such situations, it is recommended to delve into the full-text article to find the missing information necessary to interpret the results, including references, which put the effect size in relative terms. If this information is unavailable in the full text, exercise significant caution in your interpretation and application of these study findings.

ded to delve into the full-text article to find the missing information necessary to interpret the results, including references, which put the effect size in relative terms. If this information is unavailable in the full text, exercise significant caution in your interpretation and application of these study findings. CONCLUSIONS Observational data research and big data can provide clinicians and researchers with a variety of options for conducting and interpreting OB/GYN research, with applications ranging from assessing patient health outcomes, identifying trends in utilization of medications or procedures, or for calculating prevalence estimates of disease states. Different types of observational data have varied strengths and limitations. For example, administrative claims data sources are useful for population-level prevalence estimates and utilization trends, whereas EHR-derived data and patient survey data may be more useful for exploring patient behaviors and trends in practice. In comparison with RCTs, quasi-experimental designs using observational data may afford the study of larger populations in a more cost-effective manner. Supplementary Material SUPPLEMENTARY MATERIAL ACKNOWLEDGMENTS Publication of this article was funded in part by the University of Florida Open Access Publishing Fund.

CONCLUSIONS Observational data research and big data can provide clinicians and researchers with a variety of options for conducting and interpreting OB/GYN research, with applications ranging from assessing patient health outcomes, identifying trends in utilization of medications or procedures, or for calculating prevalence estimates of disease states. Different types of observational data have varied strengths and limitations. For example, administrative claims data sources are useful for population-level prevalence estimates and utilization trends, whereas EHR-derived data and patient survey data may be more useful for exploring patient behaviors and trends in practice. In comparison with RCTs, quasi-experimental designs using observational data may afford the study of larger populations in a more cost-effective manner. Supplementary Material SUPPLEMENTARY MATERIAL ACKNOWLEDGMENTS Publication of this article was funded in part by the University of Florida Open Access Publishing Fund. All authors, faculty, and staff in a position to control the content of this CME activity and their spouses/life partners (if any) have disclosed that they have no financial relationships with, or financial interests in, any commercial organizations pertaining to this educational activity. Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.obgynsurvey.com).

EDITORIAL COMMENT (Historically, the advent of neonatal intensive care units (NICUs) has dramatically lowered the rate of neonatal mortality at low gestational ages. Around the world, in countries without the resources to have NICUs, babies born before 28 weeks’ gestation have essentially a 100% rate of mortality. Although we have been able to push the threshold of viability down to 24 and even 23 or 22 weeks of gestation, the long-term morbidity of these neonates has continued to be quite severe and in a large proportion. There continue to be improvements made in NICU care with less aggressive use of high levels of oxygen or ventilation. Thus, the current article from the BMJ, abstracted above, examined the long-term outcomes at 3 years of age in neonates born at 22 to 25 weeks’ gestation in 1995 and 2006. This study found that survival rates between the 2 time periods were improved from 39% to 52% with a great dependence on the specific gestational age (ranging from 7% at 22 weeks up to 66% at 25 weeks). Discouragingly, there was no difference in the rate of severe morbidity in survivors. However, there was an increase in neonates who survived without disability from 23% to 34%. Again, there was a strong relationship between gestational age and the prevalence of moderate or severe impairment (45% of survivors at 22/23 weeks to 25% at 25 weeks). However, much of this difference appeared due to improved survival rates as the rate of survival without severe disability was nonsignificantly increased from 59% to 65%.

strong relationship between gestational age and the prevalence of moderate or severe impairment (45% of survivors at 22/23 weeks to 25% at 25 weeks). However, much of this difference appeared due to improved survival rates as the rate of survival without severe disability was nonsignificantly increased from 59% to 65%. Although these data come from the United Kingdom, I think they are incredibly useful for counseling our patients here in the United States where long-term data are hard to come by. Given that about half of the survivors at 22 and 23 weeks’ gestation have moderate or severe morbidity, which includes cerebral palsy, blindness, deafness, or cognitive delay, how should patients be counseled? Furthermore, how much autonomy over decisions regarding resuscitation should patients have at various gestational ages? In my conversations with colleagues around the country, nearly all would not even offer/allow resuscitation before 22 weeks. Alternatively, at 26 weeks and beyond, it is often not a parent choice because the neonatal outcomes are deemed to be so good. Between these gestations, I believe a range of practices exist from institutions where resuscitation before 24 weeks is not offered but done begrudgingly if parents inquire, to routine resuscitation of all 23 weeks’ gestation neonates.

is often not a parent choice because the neonatal outcomes are deemed to be so good. Between these gestations, I believe a range of practices exist from institutions where resuscitation before 24 weeks is not offered but done begrudgingly if parents inquire, to routine resuscitation of all 23 weeks’ gestation neonates. At my institution, patients undergo informed consent and choose resuscitation versus comfort care at 22 through 25 weeks’ gestation. It is rare that we resuscitate a 22-week neonate and very rare that we don’t at 25 weeks, but patients are given the opportunity to choose. Seemingly, this is better care than an absolute rule because it allows patients to incorporate their beliefs and preferences regarding this decision. Unfortunately, these patients often have only a short time to make the decision, and it is unclear how well they understand the severity of the long-term outcomes and even how to utilize the varying probabilities to make the decision. It is a very different decision than one about whether to undergo pregnancy termination when diagnosed with a Down syndrome fetus. In that case, patients know for certain that their fetus has Down syndrome, and although there is a range of outcomes, the range is relatively narrow compared with the outcomes at 22 to 25 weeks’ gestation. Although the average outcome might be similar, the severe outcomes from 22- to 25-week neonates may be viewed as much worse than Down syndrome, and the survivors without disability have better outcomes than those with Down syndrome. So, not only do patients have to think about preferences toward the outcomes, but also they have to understand and use the probabilities related to the range of outcomes. I think this is an incredibly complex decision, one for which few expectant parents are prepared. Other than providing detailed information and responding to questions, we have little to help them make this monumental decision that can have enormous consequences. Furthermore, in this setting, it is not uncommon for these parents to interact with several different providers, most notably an obstetrician or MFM and a neonatologist. In my practice, I make every effort to do this counseling together to ensure a single message regarding this complicated information.—ABC)

Maternal nutritional requirements for optimal fetal development are difficult to discern in a human pregnancy. The ethical constraints regarding what can reasonably be put to experimental study, coupled with the length of gestation and our differing genetic strains, limit what can be stated with confidence for any given woman in any given location. Furthermore, we now know that our diet, our environment, and our responses to each in combination can alter the expression of our genes. This adds another layer of complexity to our findings. The purposes of this article are to summarize what appears to be known at this time about maternal nutritional requirements and to discuss what is not yet established. Fetal programming, or maternal epigenetic influence, occurs not by changing the genes themselves, but by altering how they are expressed. Methylation of histones, one example of epigenetic change, can influence gene expression. This epigenetic influence is transgenerational and long lasting.1

Maternal nutritional requirements for optimal fetal development are difficult to discern in a human pregnancy. The ethical constraints regarding what can reasonably be put to experimental study, coupled with the length of gestation and our differing genetic strains, limit what can be stated with confidence for any given woman in any given location. Furthermore, we now know that our diet, our environment, and our responses to each in combination can alter the expression of our genes. This adds another layer of complexity to our findings. The purposes of this article are to summarize what appears to be known at this time about maternal nutritional requirements and to discuss what is not yet established. Fetal programming, or maternal epigenetic influence, occurs not by changing the genes themselves, but by altering how they are expressed. Methylation of histones, one example of epigenetic change, can influence gene expression. This epigenetic influence is transgenerational and long lasting.1 The risk for the developing fetus of developing adult-onset disease is determined, at least in part, by maternal nutritional status at conception, during pregnancy, and in early infancy. The fetal strategy of limiting growth in the absence of adequate nutrition creates an infant with a higher insulin response to food and less growth of muscle (including heart), nephrons, and bone. If the diet improves in infancy or childhood, this infant will gain weight at a higher-than-normal rate and will be at higher risk of type 2 diabetes and the metabolic syndrome.2 As an adult, the lower number of nephrons and cardiac cells will set the stage for hypertension and cardiac failure.

rt), nephrons, and bone. If the diet improves in infancy or childhood, this infant will gain weight at a higher-than-normal rate and will be at higher risk of type 2 diabetes and the metabolic syndrome.2 As an adult, the lower number of nephrons and cardiac cells will set the stage for hypertension and cardiac failure. After fertilization, the fertilized ovum and early blastocyst appear to detect and respond to the nutritional quality of the fallopian tube environment, even before implantation. During this time, and throughout the first 10 weeks of gestation, nourishment of the developing fetus seems to be provided solely by the fluid produced by the glands of the endometrium. Studies of mice and other animal species show that there is no maternal bloodstream access to the embryo until 10 weeks of gestation, and the fluid found in the fallopian tubes matches that found in the endometrial glands.3 This is significant as all organ differentiation takes place by 11 weeks of gestation.4 Maternal nutrition must be optimized before conception, so that the preimplantation and early differentiation environment is ready to support early fetal and initial placental development. Maternal nutritional status influences nutrient partitioning to the placenta or fetus, which subsequently affects disease risk.5

.4 Maternal nutrition must be optimized before conception, so that the preimplantation and early differentiation environment is ready to support early fetal and initial placental development. Maternal nutritional status influences nutrient partitioning to the placenta or fetus, which subsequently affects disease risk.5 Overlying all of this is the current phenomenon of high-calorie malnutrition brought on by the increased reliance on processed foods and the drop in vegetable and fruit consumption around the world. The higher rate of maternal diabetes brought on by a high-calorie diet and the strong effect of diabetes on pregnancy outcome make this phenomenon particularly important. There is a growing appreciation of the chemical communication between the mother and the fetus and the competing interests of the mother versus those of the fetus. It appears that the developing fetus, from conception onward, evaluates the nutritional environment available to it and adjusts its rate of growth accordingly. At the same time, while the goal of the fetus is to maximize its chances of successful development and reproduction, the goal of the mother’s body is to maximize her long-term reproductive potential, even if it means sacrificing the current fetus to do so. This means that in the face of limited nutrition the placenta may limit what is available to the fetus, even if the mother receives supplementation.

l development and reproduction, the goal of the mother’s body is to maximize her long-term reproductive potential, even if it means sacrificing the current fetus to do so. This means that in the face of limited nutrition the placenta may limit what is available to the fetus, even if the mother receives supplementation. The following is a brief summary of what is known and what is not known about the impact of the maternal diet on fetal development and pregnancy outcomes organized by gestational weight gain goals, specific nutrient recommendations, and general dietary needs. General Overview Pregnancy and lactation are associated with major metabolic and physiologic changes in the mother. Nutritional requirements increase to optimize both maternal adaptation and fetal development. Improving the mother’s diet before and during pregnancy reduces the risk of medical problems for her and her infant.

General Overview Pregnancy and lactation are associated with major metabolic and physiologic changes in the mother. Nutritional requirements increase to optimize both maternal adaptation and fetal development. Improving the mother’s diet before and during pregnancy reduces the risk of medical problems for her and her infant. Gestational Weight Gain Recommendations The Institute of Medicine of the National Academy of Sciences provided gestational weight gain recommendations based on maternal prepregnancy body mass index (BMI) (Table 1; Fig. 1). These recommendations, however, specifically excluded women with diabetes, who now make up a significant portion of the US population. For diabetic women, there is evidence that weight gain at the lower edge of the recommended range or even below this range leads to better maternal and fetal outcomes.6,7 The additional energy increment needed to support appropriate weight gain during pregnancy is 90 to 125 kcal/d in the first trimester, 286 to 350 kcal/d in the second trimester, and 466 to 500 kcal/d in the third trimester.8,9 Thus, the old adage “eating for two” is not mathematically representational of the 10% to 25% increase in caloric intake actually needed to support a healthy pregnancy. Furthermore, given the average caloric intake of many individuals, there is no need to increase caloric intake in pregnancy, but rather to shift low-nutritional calories into more nutrient-dense calories. The protein requirement during pregnancy is not much higher than that needed by a nonpregnant woman, coming to 0.45 g/lb of maternal body weight, or an average of 71 g/d. High protein supplementation or balanced protein supplementation is not helpful and may be harmful to the pregnancy,10 whereas balanced energy/protein supplementation appears to lead to a lower risk of small-for-gestational-age (SGA) births, as well as a small increase in mean birth weight and maternal weight gain.10–13

igh protein supplementation or balanced protein supplementation is not helpful and may be harmful to the pregnancy,10 whereas balanced energy/protein supplementation appears to lead to a lower risk of small-for-gestational-age (SGA) births, as well as a small increase in mean birth weight and maternal weight gain.10–13 TABLE 1 Recommended Weight Gain in Pregnancy FIG. 1 Metabolic changes in pregnancy. Maternal metabolism is altered to redirect nutrients to the placenta and fetus. 40% of weight gain is associated with the fetus and placenta 60% of weight gain is associated with maternal changes Women with a high prepregnancy BMI have infants with higher birth weights, and heavier babies have a higher body fat mass. Except for the very overweight group, birth weight is strongly correlated to maternal weight gain.5,14 An elevated energy intake is associated with the following: increased maternal weight gain increased risk of hypertension gestational diabetes (GDM) cesarean deliveries, macrosomia (birth weight >4500 g) childhood obesity in the offspring An inadequate energy intake is associated with the following: SGA or low birth weight (LBW; <2500 g) SGA is associated with an increased risk of adult metabolic diseases including type 2 diabetes

increased risk of hypertension gestational diabetes (GDM) cesarean deliveries, macrosomia (birth weight >4500 g) childhood obesity in the offspring An inadequate energy intake is associated with the following: SGA or low birth weight (LBW; <2500 g) SGA is associated with an increased risk of adult metabolic diseases including type 2 diabetes The risk ratio for developing impaired glucose tolerance or type 2 diabetes mellitus is 4 times higher among babies with a birth weight of 5.5 lb or less than those with a birth weight of 7.5 to 8.5 lb.15 Intrauterine growth restriction is associated with an increased risk of adult onset of hypertension and stroke. Large-for-gestational-age infants are at increased risk of infant and adult obesity and other obesity-related comorbidities.15

birth weight of 5.5 lb or less than those with a birth weight of 7.5 to 8.5 lb.15 Intrauterine growth restriction is associated with an increased risk of adult onset of hypertension and stroke. Large-for-gestational-age infants are at increased risk of infant and adult obesity and other obesity-related comorbidities.15 Episodes of famine in otherwise well-nourished populations (the Netherlands in World War II and rural China in the 1950’s Great Leap Forward) provide valuable data to evaluate the impact of maternal diet on short- and long-term health outcomes in the offspring. Exposure to famine during gestation resulted in higher rates of adult-onset disease in the offspring including impaired glucose tolerance,16–18 obesity,19 coronary heart disease,19–23 atherogenic lipid profiles,19,22 hypertension,24,25 microalbuminuria,19,26 schizophrenia,19,22,27,28 antisocial personality and affective disorders,19,29–31 and a higher rate of addictive disorder in men.32 The offspring also tend to eat a higher fat diet.33 Women exposed to famine in utero tend to have more offspring, more twin gestations, and younger age at first pregnancy34 and show a higher incidence of metabolic syndrome.35 There are also differing effects of famine depending on when it occurs during the pregnancy. As organogenesis takes place along a rigid timeline, nutrient deficits early in the pregnancy may lead to compromised organ development, whereas deficits later in pregnancy may lead to an LBW infant with normal organ function.

There are also differing effects of famine depending on when it occurs during the pregnancy. As organogenesis takes place along a rigid timeline, nutrient deficits early in the pregnancy may lead to compromised organ development, whereas deficits later in pregnancy may lead to an LBW infant with normal organ function. Carbohydrates Carbohydrates are the body’s, and importantly the brain’s, main source of energy. Pregnant women need the energy provided by carbohydrates to grow a healthy baby, as glucose, derived from carbohydrate, is the main fuel used for intrauterine growth.36 The Daily Recommended Intake (DRI) for carbohydrates during pregnancy is 175 g/d. It is important that pregnant women choose high-quality carbohydrates with a low glycemic index (GI), found naturally in whole foods such as whole grains, nonstarchy vegetables, fruits, beans, peas, lentils, and low-fat dairy.36 Carbohydrates may also take the form of added sugars such as table sugar, honey, syrup, cane sugar, agave, high-fructose corn syrup, and fruit juice concentrate. All women, including pregnant women, should limit their intake of these sugars, as well as foods that have high amounts of added sugar such as candy, desserts, and sugar-sweetened beverages (soda, juice, lemonade, sweetened teas, and other fruit-flavored drinks). A study by Loy et al37 in Malaysia showed that increased consumption of fruits and vegetables was associated with an increase in birth weight, birth length, and head circumference, but was not associated with any specific measured micronutrient intake (Table 2).

monade, sweetened teas, and other fruit-flavored drinks). A study by Loy et al37 in Malaysia showed that increased consumption of fruits and vegetables was associated with an increase in birth weight, birth length, and head circumference, but was not associated with any specific measured micronutrient intake (Table 2). TABLE 2 Recommended Daily Intake During Pregnancy and Dietary Sources of Macronutrients and Micronutrients Protein Dietary protein provides important building blocks during times of growth, development, or repair and also provides structural components for human cells and for the synthesis of enzymes, which help to ensure proper function of these cells. A diet low in protein during pregnancy decreases insulin secretion in rats, whereas a diet high in protein induces changes in genes associated with energy metabolism in the liver in rats, mice, and pigs.38,39

mponents for human cells and for the synthesis of enzymes, which help to ensure proper function of these cells. A diet low in protein during pregnancy decreases insulin secretion in rats, whereas a diet high in protein induces changes in genes associated with energy metabolism in the liver in rats, mice, and pigs.38,39 In some cases, there appear to be important interrelated nutritional effects, such that supplementation with one nutrient has differing effects based on the availability of other nutrients. A good example of this would be providing only protein supplementation to a population that is both protein and energy deficient (dangerous) versus the effect of supplementing both protein and energy in a balanced manner (promotes growth) or by providing folate with or without iron supplementation. While the primary intent here is to discuss each nutrient separately, some discussion of their interactions is necessary. The DRI for protein during pregnancy is 1.1 g/kg per day or approximately 71 g protein per day starting in the second trimester, which is approximately 25 g more than what is recommended for nonpregnant women. Meats, poultry, seafood, eggs, milk and dairy products, beans, lentils, nuts, and seeds are rich sources of protein. Protein, in smaller amounts, is also found in vegetables and grains. Fat Fats are part of a healthy diet during pregnancy. Fat is primarily used as an energy source by the body, but is also used to transport the fat-soluble vitamins A, D, E, and K and to provide essential fatty acids that cannot be synthesized by the body.

In some cases, there appear to be important interrelated nutritional effects, such that supplementation with one nutrient has differing effects based on the availability of other nutrients. A good example of this would be providing only protein supplementation to a population that is both protein and energy deficient (dangerous) versus the effect of supplementing both protein and energy in a balanced manner (promotes growth) or by providing folate with or without iron supplementation. While the primary intent here is to discuss each nutrient separately, some discussion of their interactions is necessary. The DRI for protein during pregnancy is 1.1 g/kg per day or approximately 71 g protein per day starting in the second trimester, which is approximately 25 g more than what is recommended for nonpregnant women. Meats, poultry, seafood, eggs, milk and dairy products, beans, lentils, nuts, and seeds are rich sources of protein. Protein, in smaller amounts, is also found in vegetables and grains. Fat Fats are part of a healthy diet during pregnancy. Fat is primarily used as an energy source by the body, but is also used to transport the fat-soluble vitamins A, D, E, and K and to provide essential fatty acids that cannot be synthesized by the body. Although there is no DRI for fat specific to pregnancy, it seems reasonable for pregnant women to follow the current guidelines for Americans, which suggest keeping total dietary fat intake to between 25% and 35% of total calories.40,41 The quality of the fat is also an important consideration when looking at overall fat intake. Recent scientific investigation allows for greater understanding of specific fatty acid functions and is helpful to clarify the fatty acids best for health. Based on recent research and consensus, increased intake of the omega-3 (n-3) polyunsaturated fatty acids (PUFAs) should be emphasized, and intake of saturated fatty acids should be limited to no more than 7% to 10% of total calorie intake; trans-fatty acid intake should be avoided whenever possible.42

best for health. Based on recent research and consensus, increased intake of the omega-3 (n-3) polyunsaturated fatty acids (PUFAs) should be emphasized, and intake of saturated fatty acids should be limited to no more than 7% to 10% of total calorie intake; trans-fatty acid intake should be avoided whenever possible.42 High-fat diets in pregnancy have been found to increase insulin resistance. A high saturated fat intake is associated with development of glucose abnormalities in pregnancy and an increased risk of GDM. Higher intakes of animal fat and cholesterol before pregnancy are also associated with an increased risk of GDM, as is a higher intake of saturated fat during pregnancy.43–45 Docosahexaenoic acid (DHA) and n-3 fatty acids are needed for brain and retinal development throughout the third trimester of gestation and the first year of life. The fetus needs 200 to 300 mg/d of DHA and 500 mg of DHA plus eicosapentaenoic acid (EPA) per day. Good sources of n-3 fatty acids include fatty fish and seafood (recommendation of 12 oz per week), walnuts, and dietary supplements.

inal development throughout the third trimester of gestation and the first year of life. The fetus needs 200 to 300 mg/d of DHA and 500 mg of DHA plus eicosapentaenoic acid (EPA) per day. Good sources of n-3 fatty acids include fatty fish and seafood (recommendation of 12 oz per week), walnuts, and dietary supplements. Polyunsaturated Fatty Acids For women who do not eat fish, n-3 PUFA supplementation appears to be associated with a small decrease in preterm birth and LBW infants,46,47 and fish consumption is associated with increased birth weight and gestation length.48 Docosahexaenoic acid and EPA are important in visual and cognitive development.49 Some studies in women who do eat some fish fail to show effects,50 however, and other studies both support and refute a correlation between the ratio of n-6:n-3 fatty acids, where a ratio of less than 9 is considered healthful.51,52 However, there is also concern about consuming too much fish during pregnancy as some fish contain significant amounts of organic mercury compounds relative to the amount of DHA they provide. Studies done in the Faroe Islands53,54 showed an effect of fish consumption on measures of neurologic function at ages 7 and 14 years that were correlated with cord blood mercury concentrations. The population living on these islands eats both fish and whale blubber, and maternal blood levels of mercury have been shown to peak at 10 times higher levels after eating blubber than after eating fish. In addition, there is a higher intake of polychlorinated biphenylbromine compounds in the Faroe Island diet, although controlling for this contaminant did not affect the relationship with mercury.

rnal blood levels of mercury have been shown to peak at 10 times higher levels after eating blubber than after eating fish. In addition, there is a higher intake of polychlorinated biphenylbromine compounds in the Faroe Island diet, although controlling for this contaminant did not affect the relationship with mercury. A study done in the Seychelles islands, where fish consumption is high (with the offspring now 14 years old), did not show these findings,55–58 but a recent study out of Hong Kong59 confirmed the Faroe Island findings in a population that did not consume whale blubber. These results appear to show that a moderate amount of fish consumption in pregnant women is beneficial, but that fish with a low ratio of DHA to mercury (albacore, halibut, red snapper, for example) should be eaten in limited quantities and that shark, swordfish, king mackerel, and tilefish, fish known to have higher mercury concentrations, should be avoided. Also of note are several studies showing a decrease in atopic syndrome, eczema, and wheezing in the face of small-particle air pollution in infants born after prenatal supplementation with DHA in the third trimester60–62 and studies showing that the incidence of eczema and allergies can be increased by n-6 fatty acid intake in the last trimester and decreased by n-3 fatty acid intake in the same period.63–67 A recent report shows that pregnant women have avoided fish consumption altogether as a result of concerns about mercury, while it is clear that moderate consumption (8–12 oz/wk) of fish with low levels of mercury is beneficial for all pregnant women and needs to be reinforced.68

ty acid intake in the same period.63–67 A recent report shows that pregnant women have avoided fish consumption altogether as a result of concerns about mercury, while it is clear that moderate consumption (8–12 oz/wk) of fish with low levels of mercury is beneficial for all pregnant women and needs to be reinforced.68 Iron Maternal iron stores at conception are a strong predictor of maternal iron status and risk of iron deficiency and anemia in later pregnancy. There are some data that show that in populations with a high prevalence of anemia there is an increase in birth weight in infants following in utero iron supplementation (usually provided along with folate).69 The incidence of anemia and iron deficiency in the United States is about 12% in nonpregnant women and 18% in pregnant women. These conditions are found in 43% and up to 75% of nonpregnant and pregnant women, respectively, in developing countries. This is due to poor diet, limited consumption of iron-rich foods, intestinal parasites, and chronic blood loss.

ency in the United States is about 12% in nonpregnant women and 18% in pregnant women. These conditions are found in 43% and up to 75% of nonpregnant and pregnant women, respectively, in developing countries. This is due to poor diet, limited consumption of iron-rich foods, intestinal parasites, and chronic blood loss. There is no doubt that adequate iron stores are crucial for maternal safety. While there is no evidence of fetal growth restriction with a hemoglobin level greater than 6 g/dL, hemorrhage in the presence of anemia is the cause of at least 20% of maternal deaths worldwide.70 Mild anemia reflected by a hemoglobin concentration of 8 to 10.9 g/dL may be physiologic. There is a normal drop in hemoglobin concentration during pregnancy because of the increase in the circulating blood volume, and this may actually improve blood flow across the placental bed by decreasing viscosity,71 thus providing more efficient oxygen transfer. A recent review of current literature by the US Preventive Services Task Force72 failed to show any benefit to routine iron supplementation in pregnancy for women in the developed world. However, as noted previously, any woman with an existing iron deficiency or a major risk of severe bleeding (placenta previa, abruption, clotting disorder, etc) should of course receive treatment.

ices Task Force72 failed to show any benefit to routine iron supplementation in pregnancy for women in the developed world. However, as noted previously, any woman with an existing iron deficiency or a major risk of severe bleeding (placenta previa, abruption, clotting disorder, etc) should of course receive treatment. Folate Folate, derived from food, and folic acid, the synthetic form of folate found in vitamin supplements, is a water-soluble B vitamin. It is a methyl donor that is required for DNA synthesis and cell division. Folate/folic acid is required for neural tube development occurring within 28 days of conception. The value of folate in the prevention of neural tube defects (NTDs) is well established,73,74 and recent studies showing hypomethylation of neural tissue in cases of NTD support this observation.75 Folate is needed for the formation of the brain and spinal cord. The neural tube closes by day 28 of gestation. If it does not close completely, an opening at the lower end of the spine causes spina bifida, whereas a larger defect can lead to anencephaly (complete failure of the brain to develop).

Folate Folate, derived from food, and folic acid, the synthetic form of folate found in vitamin supplements, is a water-soluble B vitamin. It is a methyl donor that is required for DNA synthesis and cell division. Folate/folic acid is required for neural tube development occurring within 28 days of conception. The value of folate in the prevention of neural tube defects (NTDs) is well established,73,74 and recent studies showing hypomethylation of neural tissue in cases of NTD support this observation.75 Folate is needed for the formation of the brain and spinal cord. The neural tube closes by day 28 of gestation. If it does not close completely, an opening at the lower end of the spine causes spina bifida, whereas a larger defect can lead to anencephaly (complete failure of the brain to develop). Since 1998, the US Food and Drug Administration mandated increased fortification of cereal grains to 140 μg (0.14 mg) folic acid/100 g grain. Since then, the incidence of NTDs has decreased by approximately 30% to 40% (rates of spina bifida declined ~31%; anencephaly declined 16%) in the United States, but this differs by racial/ethnic group. It has been proposed that only folate in food, rather than more in supplements, will lead to further improvement.76 Data have shown, however, that the current supplementation policy has only dropped the incidence of NTDs by 20%, and calculations predict that if supplementation were increased to 4 mg/d, the incidence of NTDs would drop by 82%. Observational studies also suggest that folate given in pregnancy may drop the incidence of preterm labor between 20 and 32 weeks,77–80 while controlled trials do not.81 If the rest of the world adopted similar food fortification and folate supplementation policies, the worldwide prevalence of 300,000 NTDs could drop by 150,000 to 210,000 cases per year.

hat folate given in pregnancy may drop the incidence of preterm labor between 20 and 32 weeks,77–80 while controlled trials do not.81 If the rest of the world adopted similar food fortification and folate supplementation policies, the worldwide prevalence of 300,000 NTDs could drop by 150,000 to 210,000 cases per year. The US recommendations are for women of childbearing age to consume 0.4 mg of folate in vitamin form or fortified foods daily, in addition to a diet high in folate-rich foods, because of the lower bioavailability in foods. This amount should increase to 0.6 mg/d in pregnancy and 0.5 mg/d during lactation. Women with a history of pregnancy affected by a neural tube disorder (eg, spina bifida) should consume 4 mg/d of supplemental folic acid (a 10-fold increase, which requires a prescription), starting at least 1 month before conception. Foods rich in folate include beans, peas, orange juice, and green leafy vegetables. Prenatal vitamins contain folic acid, with up to 0.8 to 1 mg/tablet depending on the brand. Low folate intake (<.15 mg/d) has been associated with an increased risk of cancer, and very high intakes of folic acid (>4 times the normal dose) have been possibly associated with an increased risk of cancer in animal studies.82 Unfortunately, while benefit of folic acid supplementation during pregnancy shows up quickly, cancer risks must be studied over decades, so a small amount of uncertainty concerning high doses of folic acid remains.74 There is a biological basis for the concern, as folic acid is metabolized by a liver enzyme that may be unable to handle the higher amounts of this vitamin, and circulating folic acid is associated with reduced natural killer cell cytotoxicity.83

mall amount of uncertainty concerning high doses of folic acid remains.74 There is a biological basis for the concern, as folic acid is metabolized by a liver enzyme that may be unable to handle the higher amounts of this vitamin, and circulating folic acid is associated with reduced natural killer cell cytotoxicity.83 Hispanic women have higher rates of NTDs than non-Hispanic women: 10.34 versus 7.9 per 10,000 live births; this shifted to 7.92 versus 5.35 after the folic acid fortification program was established. Hispanic women consume less folate from food and less from supplements, 20% versus 37%, and 60% consume corn masa flour, which is not fortified with folic acid. Fortification of corn masa flour in the United States would increase folate intake by approximately 20% among 15- to 44-year-old Mexican American women and possibly lead to fewer NTD-affected pregnancies in this population.84

ents, 20% versus 37%, and 60% consume corn masa flour, which is not fortified with folic acid. Fortification of corn masa flour in the United States would increase folate intake by approximately 20% among 15- to 44-year-old Mexican American women and possibly lead to fewer NTD-affected pregnancies in this population.84 High supplement levels of folic acid after the first trimester have also been associated with an increase in childhood asthma and eczema,85 whereas low folate levels have been linked to language delay, emotional problems, and schizophrenia in the offspring.85–87 An epidemiological study in Pune, India, found that a combination of low blood B12 levels and high folate levels in women (measured in late pregnancy) are associated with small, relatively obese offspring with a high level of insulin resistance.88 Vitamin B12 deficiency blocks the metabolism of folate and leads to a buildup of 5-methyltetrahydrofolate, which may be the source of the problem. This population differs from a US cohort, as it had almost no women with folate deficiency, and a large number of the pregnant women were prescribed a dose of 5 mg folate daily from 12 weeks’ gestation or later until term.

olate and leads to a buildup of 5-methyltetrahydrofolate, which may be the source of the problem. This population differs from a US cohort, as it had almost no women with folate deficiency, and a large number of the pregnant women were prescribed a dose of 5 mg folate daily from 12 weeks’ gestation or later until term. Iodine Iodine requirements during pregnancy are increased because of a 50% increase in maternal thyroid (T4) hormone production. Fetal thyroid-stimulating hormone is not synthesized until the 10th to 12th week of gestation, approximately the same time that the fetal thyroid is capable of concentrating iodine and synthesizing iodothyronine. Little fetal hormone synthesis occurs, however, until the 18th to 20th week of gestation. There is also an increased loss of iodine in the urine during pregnancy. Therefore, through the first half of pregnancy, the fetus relies on the mother for thyroid hormone. When maternal iodine intake is low, maternal thyroid hormone production is low. Thyroid hormone is required for normal neuronal migration, myelination, and synaptic transmission and plasticity during fetal and early postnatal life. Poor neuronal development in the fetus, due to iodine deficiency during critical time points, causes irreversible brain damage, is the leading cause of preventable mental retardation worldwide, and can result in up to a 20-point drop in IQ. Iodine deficiency during fetal development is also associated with fetal goiter (potentially causing obstruction at delivery), hypothyroidism, and cretinism.

al time points, causes irreversible brain damage, is the leading cause of preventable mental retardation worldwide, and can result in up to a 20-point drop in IQ. Iodine deficiency during fetal development is also associated with fetal goiter (potentially causing obstruction at delivery), hypothyroidism, and cretinism. Major sources of iodine include iodized salt, seafood, seaweed, kelp, and dairy products. Iodized salt provides 77 μg iodine/g of salt (~220 μg iodine in ½ teaspoon of salt). The US Recommended Daily Allowance for iodine is 220 μg/d. The World Health Organization/International Council for Control of Iodine Deficiency Disorders recommends 250 μg/d.

de iodized salt, seafood, seaweed, kelp, and dairy products. Iodized salt provides 77 μg iodine/g of salt (~220 μg iodine in ½ teaspoon of salt). The US Recommended Daily Allowance for iodine is 220 μg/d. The World Health Organization/International Council for Control of Iodine Deficiency Disorders recommends 250 μg/d. Iodine intake was not felt to be a problem in the United States for many decades but is again an issue. The push to limit salt intake, coupled with the rise in popularity of sea salt and kosher salt, a fall in the number of bakeries using iodized dough, the popularity of organic milk (which has approximately 40% less iodine than regular commercial milk), soy and other “nonmilk” milks, and the rapid loss of iodine from iodinated salt in warm humid climates, has led to iodine intake dropping 50% between 1970 and 1990.89 Soy intake inhibits iodine absorption and interferes with thyroid hormone production, but can be circumvented by adequate iodine intake. Currently, as many as 30% of pregnant women in this country are estimated to have low levels (<100 μg/L), and as recently as 2012, just under 50% of prenatal vitamins contained no iodine. At least 1 study90 showed that supplementation given after 6 to 8 weeks of gestation was ineffective at preventing a drop in IQ levels (8–12 points) in the offspring. The current recommendation is for all pregnant women to take prenatal vitamins containing 150 to 250 μg daily.91 Older women and multiparas should be taking the higher doses. The iodine should come from potassium iodide (KI), not from kelp, as kelp has been shown to contain highly variable amounts of iodine, as well as (in some cases) large concentrations of arsenic and other heavy metals.

tamins containing 150 to 250 μg daily.91 Older women and multiparas should be taking the higher doses. The iodine should come from potassium iodide (KI), not from kelp, as kelp has been shown to contain highly variable amounts of iodine, as well as (in some cases) large concentrations of arsenic and other heavy metals. In addition, Connelly et al92 recently described 3 cases of congenital hypothyroidism caused by excess maternal iodine ingestion (12.5 mg/d from nutritional supplements, Iodoral [Optimox Corp, Torrance, CA]) Infants were identified through the Oregon newborn screening program. Concentrations of whole-blood iodine were 10 times above mean control levels. One infant required chronic levothyroxine treatment. Maternal breast milk iodine content was also significantly elevated in the mother. The other 2 infants were dizygotic twins who were treated with levothyroxine for 3 weeks, which was then discontinued. None of the infants presented with neonatal goiter; the long-term consequences are not known. These cases emphasize the need for health care providers to evaluate prenatal vitamin/mineral use among pregnant women and to recommend appropriate supplements to prevent unintentional toxicity.

s, which was then discontinued. None of the infants presented with neonatal goiter; the long-term consequences are not known. These cases emphasize the need for health care providers to evaluate prenatal vitamin/mineral use among pregnant women and to recommend appropriate supplements to prevent unintentional toxicity. Vitamin A Deficiency as a cause of night blindness is a major problem worldwide, but not a significant one in the United States. Supplementation does prevent and cure this condition.93–95 All studies to date fail to show any significant effects of vitamin A supplementation on pregnancy outcomes for mother or fetus.96 More than 10,000 IU of vitamin A supplementation per day (4 times the Recommended Daily Allowance) increases the risk of cleft lip or palate, hydrocephalus, and heart defects. β-Carotene, found in food, does not pose a risk. Vitamins B6, B12, and C A low level of B12 has been associated with high levels of homocysteine, which has been associated with preeclampsia97 and LBW.98 Deficiency may occur in vegans who do not take supplements. By serum testing, 17% to 39% of pregnant women are deficient in B12, but this was not shown to correlate with pregnancy complications or outcome. Vitamin B6 supplementation has been shown in 3 small studies to give a modest gain in birth weight.99 Vitamin C supplementation has not been shown to be of value in improving pregnancy outcomes,99 but has shown value for prevention of wheezing and asthma in the children of smokers.100

regnancy complications or outcome. Vitamin B6 supplementation has been shown in 3 small studies to give a modest gain in birth weight.99 Vitamin C supplementation has not been shown to be of value in improving pregnancy outcomes,99 but has shown value for prevention of wheezing and asthma in the children of smokers.100 Vitamin E One study correlated risk of wheeze inversely to the level of vitamin E as well as a linear relationship between the level and the forced expiratory volume at 1 minute.101 No other studies appear to support or refute this. Deficiency during pregnancy may cause miscarriage, preterm birth, preeclampsia, and intrauterine growth retardation.102 There does not seem to be a need for supplements. Vitamin D Adequate vitamin D status during pregnancy is necessary to ensure appropriate maternal responses to the calcium demands of the fetus for bone mineral accretion. Approximately 25 to 30 g of calcium are transferred to the fetal skeleton during pregnancy—approximately 250 mg/d in the third trimester. Major dietary sources of vitamin D are fatty fish (salmon), egg yolks, fortified milk, margarine, yogurt, and orange juice.

e calcium demands of the fetus for bone mineral accretion. Approximately 25 to 30 g of calcium are transferred to the fetal skeleton during pregnancy—approximately 250 mg/d in the third trimester. Major dietary sources of vitamin D are fatty fish (salmon), egg yolks, fortified milk, margarine, yogurt, and orange juice. Vitamin D screening and supplementation during pregnancy is a topic of great interest and controversy. American College of Obstetricians and Gynecologists103 suggests that current evidence does not support routine screening of all pregnant women for vitamin D deficiency. Serum 25-OH vitamin D concentrations of 20 ng/mL (50 nmol/L) or greater are associated with bone health; values of 32 ng/mL (80 nmol/L) or greater are associated with appropriate biomarker concentrations. The recommended intake of Vitamin D set by the Institute of Medicine in 2010 is 600 IU/d. However, when vitamin D deficiency is identified during pregnancy, supplementation with up to 1000 to 2000 IU/d is considered safe by most experts. Others feel that supplementation in doses up to 4000 IU/d is safe during pregnancy and lactation. It has been identified as a nutrient of public health concern by the US Department of Health and Human Services.104

ntified during pregnancy, supplementation with up to 1000 to 2000 IU/d is considered safe by most experts. Others feel that supplementation in doses up to 4000 IU/d is safe during pregnancy and lactation. It has been identified as a nutrient of public health concern by the US Department of Health and Human Services.104 Numerous studies have shown a prevalence of vitamin D levels of less than 50 nmol/L in 30% to 96% of pregnant women, increasing with latitude and with nonwhite race.105–111 A study done using archived samples of cord blood in Denmark showed an increased risk of developing schizophrenia with either elevated or low levels of Vitamin D. It was hypothesized that if all the infants could have been born with normal levels there would have been a reduction of 43% in the incidence of schizophrenia among the adults.112 There is also an association of vitamin D deficiency and insulin resistance,113 which was reversed with a single injection of the vitamin114 (not approved in the United States). Some studies report higher long-bone density in fetuses of women with adequate amounts of vitamin D,115,116 but the numbers are small.117 The general recommendation is 1000 to 2000 mg/d of supplementation, but there are no agreed-upon levels for sufficiency or insufficiency in pregnancy. There are no studies clearly showing differences in maternal outcomes or fetal survival, birth weight, or gestational length relative to vitamin D,118 but studies on bone health (and lung function and asthma at age 6 years119) support a level of at least 50 nmol/L, which implies a need for some supplementation for all pregnant women.120–123

learly showing differences in maternal outcomes or fetal survival, birth weight, or gestational length relative to vitamin D,118 but studies on bone health (and lung function and asthma at age 6 years119) support a level of at least 50 nmol/L, which implies a need for some supplementation for all pregnant women.120–123 Calcium Women lose 3% to 5% of their bone mass while lactating but rapidly regain it within 6 months after weaning. Calcium and/or vitamin D deficiency leads to porous, weak bones and to rickets. A Cochrane Review has concluded that there is sufficient evidence to show that calcium supplementation lowers the risk of preeclampsia and other hypertensive disease of pregnancy, especially in the face of a low-calcium diet.124 Food sources of calcium are milk, yogurt, cheese, fortified juices, tofu, cereals, Chinese cabbage, kale, broccoli, and fish with bones. Multivitamin supplements, including prenatal supplements, contain little calcium. Multiple Micronutrients (Multivitamins) When compared with iron and folate alone, multiple micronutrients can significantly lower the incidence of SGA infants, but in 5 studies, they increased the risk of neonatal death when started after the first trimester.125 Regular consumption of fortified cereal grains appears to be an effective source of all but B vitamins, iron, and folate. In the United States, prenatal vitamins should all contain 2 essential elements: 0.8 to 1.0 mg folate 120 to 150 μg iodine

Multiple Micronutrients (Multivitamins) When compared with iron and folate alone, multiple micronutrients can significantly lower the incidence of SGA infants, but in 5 studies, they increased the risk of neonatal death when started after the first trimester.125 Regular consumption of fortified cereal grains appears to be an effective source of all but B vitamins, iron, and folate. In the United States, prenatal vitamins should all contain 2 essential elements: 0.8 to 1.0 mg folate 120 to 150 μg iodine Calcium, in the amount of 1000 to 2000 mg, should be provided as a supplement. It is not clear that any other specific components are necessary at this time. The introduction of a nutrient known to be lacking in an isolated population can highlight the effects of a deficiency over a short time, but the recognition of epigenetic mechanisms means that patience is required to evaluate the full effect on subsequent generations.

er specific components are necessary at this time. The introduction of a nutrient known to be lacking in an isolated population can highlight the effects of a deficiency over a short time, but the recognition of epigenetic mechanisms means that patience is required to evaluate the full effect on subsequent generations. High Calorie Malnutrition Lack of adequate protein in a prenatal diet has been shown to cause lifelong damage to the developing fetus in humans and numerous animal studies. The goal should be to have a diet of 20% protein or more throughout the pregnancy. Specific damage has been shown at levels of 7.2% or less, but an extremely high protein intake has also been shown to be damaging.88,126 Damage from a low-protein diet includes decreased brain size,127 altered fat distribution,128 increased obesity,129 shorter gestation and decreased birth weight,130 increased stress sensitivity,131 decreased sperm quality,132 altered cardiac energy metabolism,133 and changes in muscular tone.134 A diet filled with carbohydrates and fats (soft drinks, chips, etc) can easily lead to satiation before an adequate amount of protein and other nutrients has been consumed. A diet skewed toward a high-meat, low-carbohydrate intake126 leads to a higher incidence of hypertension in the offspring as well as high cortisol levels.

filled with carbohydrates and fats (soft drinks, chips, etc) can easily lead to satiation before an adequate amount of protein and other nutrients has been consumed. A diet skewed toward a high-meat, low-carbohydrate intake126 leads to a higher incidence of hypertension in the offspring as well as high cortisol levels. DIETARY PATTERNS Evidence from international scientific research has identified various eating patterns that may provide short- and long-term health benefits, including a reduced risk of chronic disease.135 Analysis of overall food patterns takes into account the complex interactions and cumulative effects of multiple nutrients in the entire diet, therefore offering a more comprehensive and complementary approach to public health. The “Western” diet is a pattern of eating that is associated with adverse health outcomes. The typical Western diet is one low in fruits, vegetables, whole grains, fish/seafood, and low-fat dairy. It is often called the meat-sweet diet because it is high in refined sugars, refined grains (baked goods, desserts, chips), red meat, and saturated fat. It also typically contains high-sugar drinks, high-fat dairy, and higher intakes of processed meats. Other traditional eating patterns alternatively can provide health benefits. Their variety demonstrates that people can eat healthfully in a number of ways, which also likely applies to pregnancy.

The “Western” diet is a pattern of eating that is associated with adverse health outcomes. The typical Western diet is one low in fruits, vegetables, whole grains, fish/seafood, and low-fat dairy. It is often called the meat-sweet diet because it is high in refined sugars, refined grains (baked goods, desserts, chips), red meat, and saturated fat. It also typically contains high-sugar drinks, high-fat dairy, and higher intakes of processed meats. Other traditional eating patterns alternatively can provide health benefits. Their variety demonstrates that people can eat healthfully in a number of ways, which also likely applies to pregnancy. Several healthful dietary patterns have been inversely associated with the risk of type 2 diabetes mellitus, cardiovascular disease, and total mortality135 Examples of healthy dietary patterns include the aMED (alternative Mediterranean diet), DASH (Dietary Approaches to Stop Hypertension), and aHEI (alternative Healthy Eating Index). These healthy dietary patterns share common components, namely, emphasizing a high intake of vegetables and fruits, high-quality carbohydrates including whole grains, protein from beans and smaller amounts from lean meats, healthy fats from nuts and seeds, fish and seafood and liquid oils, high in fiber, low in added sugar, and low intake of red meat and processed meats.

s, namely, emphasizing a high intake of vegetables and fruits, high-quality carbohydrates including whole grains, protein from beans and smaller amounts from lean meats, healthy fats from nuts and seeds, fish and seafood and liquid oils, high in fiber, low in added sugar, and low intake of red meat and processed meats. The types of vegetarian diets consumed in the United States vary widely.135 Vegans do not consume any animal products, whereas lacto-ovo vegetarians consume milk and eggs. Vegan diets can be low in B12, riboflavin, vitamin D, calcium, and long-chain n-3 fats if not properly supplemented. Vegetarian diets can also potentially be low in certain nutrients depending on which food groups might be avoided such as dairy, eggs, and/or fish and seafood, so supplement recommendations should be individualized. Prepregnancy adherence to dietary patterns is now being investigated, with a few studies showing adherence to healthful dietary patterns being significantly associated with a lower risk of GDM,136 and a recent study showed that adherence to a Mediterranean diet pattern of eating during pregnancy was associated with lower incidence of GDM and better degree of glucose tolerance even in women without GDM.137 It has been speculated that these healthy dietary patterns may minimize the susceptibilities a pregnant woman has to β-cell dysfunction and insulin resistance.

erranean diet pattern of eating during pregnancy was associated with lower incidence of GDM and better degree of glucose tolerance even in women without GDM.137 It has been speculated that these healthy dietary patterns may minimize the susceptibilities a pregnant woman has to β-cell dysfunction and insulin resistance. These data suggest that efforts to encourage dietary patterns similar to the aMED, DASH, or aHEI might yield benefits for women of reproductive age and for pregnant women as well. My Pregnancy Plate is an education tool created using this emerging evidence of healthy eating patterns that can be advocated for in this population (Fig. 2). FIG. 2 My Pregnancy Plate. Dietary suggestions for healthy nutrition during pregnancy. © 2014 Oregon Health & Science University. All rights reserved. Reproduction in whole or in part without expressed written consent from OHSU is prohibited.

These data suggest that efforts to encourage dietary patterns similar to the aMED, DASH, or aHEI might yield benefits for women of reproductive age and for pregnant women as well. My Pregnancy Plate is an education tool created using this emerging evidence of healthy eating patterns that can be advocated for in this population (Fig. 2). FIG. 2 My Pregnancy Plate. Dietary suggestions for healthy nutrition during pregnancy. © 2014 Oregon Health & Science University. All rights reserved. Reproduction in whole or in part without expressed written consent from OHSU is prohibited. Glycemic Index The GI was proposed over 30 years ago as a classification of the blood glucose–raising potential of carbohydrate-containing foods,138 and the concept has stirred up controversy ever since. The GI uses a 100-point scale to measure how rapidly the carbohydrates in food cause blood sugars to rise. Foods with a high-GI value cause a big rise in blood glucose, and foods with a low-GI value cause a smaller rise. A competing measure, the glycemic load, adjusts this value by taking into account serving sizes and the amount of carbohydrates per serving. The glycemic load may be more practical than the GI because it accounts for quantity in addition to carbohydrate quality.139 There is evidence that low-GI diets improve glycemic control in people with diabetes140,141 and reduce risk of type 2 diabetes in men and women.142,143

nd the amount of carbohydrates per serving. The glycemic load may be more practical than the GI because it accounts for quantity in addition to carbohydrate quality.139 There is evidence that low-GI diets improve glycemic control in people with diabetes140,141 and reduce risk of type 2 diabetes in men and women.142,143 The application of the GI to pregnancy outcomes is more recent. Pregnancy is a condition in which the GI may be of particular relevance because maternal glucose is the main energy substrate for intrauterine growth, and elevated maternal blood glucose levels are well recognized to contribute to excessive fetal growth. A review article144 evaluating the evidence regarding the effect of GI on maternal and fetal nutrition concluded that there was insufficient evidence to recommend a low-GI diet during normal pregnancy, as 1 of 8 studies showed an increase in SGA babies. That study, however, had very small numbers. There is probably some benefit of low-GI diet advice in reducing a woman’s risk of having a large-for-gestational age infant.145 For pregnancy complicated by GDM, however, a low-GI diet may confer benefits. Current evidence, although limited, consistently supports the advantages of, and has demonstrated no disadvantages of, a low-GI diet.146 Pregnant women with GDM are likely to benefit from following a low-GI meal pattern, with no significant adverse effects, and consideration of the GI should be given when formulating a diet for GDM.146 Using a low-GI diet for women with GDM has been shown to halve the number needing to use insulin, with no compromise of obstetric or fetal outcomes.147

ly to benefit from following a low-GI meal pattern, with no significant adverse effects, and consideration of the GI should be given when formulating a diet for GDM.146 Using a low-GI diet for women with GDM has been shown to halve the number needing to use insulin, with no compromise of obstetric or fetal outcomes.147 Glycemic index is only one tool that can be used to determine carbohydrate quality. In another clinical trial, in intensively monitored women with GDM, a low-GI diet and a conventional high-fiber diet produced similar pregnancy outcomes.148 Excess glucose is also not the only fuel that can contribute to fetal overgrowth (see section on fat). Until further larger-scale intervention trials, preferably randomized controlled trials, are completed, a low-GI diet should not replace the current pregnancy recommendations from government and health agencies.144,145

xcess glucose is also not the only fuel that can contribute to fetal overgrowth (see section on fat). Until further larger-scale intervention trials, preferably randomized controlled trials, are completed, a low-GI diet should not replace the current pregnancy recommendations from government and health agencies.144,145 CONCLUSIONS There is good evidence to support a need for supplementation with folate, iodine, and calcium for all pregnancies. There is good evidence for supplementation with vitamin C in pregnant women who smoke. There is no good current evidence to show value in supplementation for iron and vitamins A, B6, B12, A, E, or D at this time, although there are many suggestive studies for vitamin D. All pregnant women should be encouraged to eat a balanced diet rich in fresh or frozen fruits and vegetables, high-quality carbohydrates including whole grains, and with a good mix of proteins from beans, lean meats, fish, and seafood. Their diet should be low in added sugar, red meat, and processed foods. Information beyond this simple prescription is simply not yet available for pregnant women or their offspring. All authors and staff in a position to control the content of this CME activity and their spouses/life partners (if any) have disclosed that they have no financial relationships with, or financial interests in, any commercial organizations pertaining to this educational activity.

Chief Editor's Note: This article is part of a series of continuing education activities in this Journal through which a total of 36 AMA PRA Category 1 Credits™ can be earned in 2016. Instructions for how CME credits can be earned appear on the last page of the Table of Contents. Adenomyosis is a benign uterine disorder characterized by the presence of heterotopic endometrial glands and stroma in the myometrium and reactive fibrosis of the surrounding smooth muscles cells of the myometrium. For the past 80 years, a number of theories have described how adenomyosis develops. Currently, the most widespread hypothesis is that adenomyosis originates from the invagination of the basalis of the endometrium into the myometrium. According to a second theory, this basalis invagination would proceed along the intramyometrial lymphatic system. A third theory suggests that a metaplastic process initiating from ectopic intramyometrial endometrial tissue is produced de novo.1–3 Reports show that approximately 20% of cases of adenomyosis involve women younger than 40 years, and 80% are 40 to 50 years old. The most severe symptoms are associated with the older group. Adenomyosis is completely asymptomatic in approximately one third of cases. The most frequent symptoms in the remaining two thirds are menorrhagia (50%), dysmenorrhea (30%), and metrorrhagia (20%). Dyspareunia may also be a complaint.4,5

40 to 50 years old. The most severe symptoms are associated with the older group. Adenomyosis is completely asymptomatic in approximately one third of cases. The most frequent symptoms in the remaining two thirds are menorrhagia (50%), dysmenorrhea (30%), and metrorrhagia (20%). Dyspareunia may also be a complaint.4,5 From the epidemiologic data, a large number of births, spontaneous and induced abortions, and endometrial hyperplasia are related to increased risk of adenomyosis. Other risk factors associated with adenomyosis are endometriosis, smoking, and surgical trauma, such as cesarean section or curettage.6 Since adenomyosis has been reported in 60% of postmenopausal women taking tamoxifen therapy for a long time, it seems to reactivate lesions of preexisting adenomyosis,7 implying that this condition is estrogen dependent. DIAGNOSIS OF ADENOMYOSIS Until recently, adenomyosis was thought to be found only in parous women. With few physical findings, diagnosis of adenomyosis relied on surgical resection and pathological examination. With improved imaging techniques, however, it is frequently encountered in infertile patients. Hysterectomy has been advised for women with severe symptoms from adenomyosis, although alternative conservative treatment can preserve the uterus. Therefore, the role of preoperative diagnostic tools is important to avoid unnecessary surgery. In patients with uterine masses and infertility, adenomyosis needs to be excluded before other treatment options are given.8

ere symptoms from adenomyosis, although alternative conservative treatment can preserve the uterus. Therefore, the role of preoperative diagnostic tools is important to avoid unnecessary surgery. In patients with uterine masses and infertility, adenomyosis needs to be excluded before other treatment options are given.8 Hysterosalpingography was the first imaging tool used to diagnose adenomyosis, but it is no longer used to evaluate patients because of its low overall accuracy.9 Uterine enlargement without any features of fibroids and asymmetric thickening of anterior and posterior myometrial walls are signs of adenomyosis that transabdominal ultrasonography (TAU) can reveal.10,11 In a retrospective review, Siedler et al.10 examined 80 patients using TAU and diagnosed adenomyosis with a sensitivity of 63%, a specificity of 97%, and a positive predictive value of 71%. However, because it is not possible to get sufficient image resolution to visualize the myometrium, TAU cannot reliably diagnose adenomyosis or differentiate it from leiomyoma.10,11

d 80 patients using TAU and diagnosed adenomyosis with a sensitivity of 63%, a specificity of 97%, and a positive predictive value of 71%. However, because it is not possible to get sufficient image resolution to visualize the myometrium, TAU cannot reliably diagnose adenomyosis or differentiate it from leiomyoma.10,11 In clinically suspected adenomyosis cases, transvaginal sonography (TVS) should be considered as the primary diagnostic tool. Fundamental TVS signs for adenomyosis are heterogeneous and hypoechogenic poorly described areas in the myometrium, those areas with or without anechoic lacunae or cysts of varying size, linear striation radiating out from the endometrium into the myometrium, poor definition of the junctional zone (JZ), and pseudo-widening of the endometrium (enlargement of uterus with asymmetric thickening of the anterior or posterior walls). Adenomyosis is most often diagnosed in the presence of 3 or more sonographic criteria.8,12 On 3D TVS, features linked to adenomyosis were JZmax 8 mm or greater, myometrial asymmetry, and hypoechoic striations.13 When at least 2 of the described ultrasound features were present, the accuracy was 90% (sensitivity, 92%; specificity, 83%; positive predictive value, 99%; and negative predictive value, 71%).

In clinically suspected adenomyosis cases, transvaginal sonography (TVS) should be considered as the primary diagnostic tool. Fundamental TVS signs for adenomyosis are heterogeneous and hypoechogenic poorly described areas in the myometrium, those areas with or without anechoic lacunae or cysts of varying size, linear striation radiating out from the endometrium into the myometrium, poor definition of the junctional zone (JZ), and pseudo-widening of the endometrium (enlargement of uterus with asymmetric thickening of the anterior or posterior walls). Adenomyosis is most often diagnosed in the presence of 3 or more sonographic criteria.8,12 On 3D TVS, features linked to adenomyosis were JZmax 8 mm or greater, myometrial asymmetry, and hypoechoic striations.13 When at least 2 of the described ultrasound features were present, the accuracy was 90% (sensitivity, 92%; specificity, 83%; positive predictive value, 99%; and negative predictive value, 71%). According to Dueholm et al. (2001),12 magnetic resonance imaging (MRI) and TVS were equally good at identifying patients with adenomyosis, but MRI was superior to TVS to exclude the diagnosis of adenomyosis, with equal sensitivity but a higher specificity (sensitivity: MRI, 0.70 (0.46–0.87); and TVS, 0.68 (0.44–0.86) (P = 0.66); specificity: MRI, 0.86 (0.76–0.93); and TVS 0.65 (0.50–0.77) (P = 0.03). The combination of TVS and MRI had the highest sensitivity, but, surprisingly, it also had the lowest specificity. In addition, measuring the difference in junctional zone thickness may optimize the MRI diagnosis. Gordts et al.14 recently suggested the following adenomyosis classification: simple JZ hyperplasia (zone thickness >8 mm but <12 mm on T2-weighed images, in women aged 35 years or younger); partial or diffuse adenomyosis (thickness >12 mm; high signal intensity myometrial foci; involvement of the outer myometrium <1/3, <2/3, and >2/3), adenomyoma (myometrial mass with indistinct margins with primarily low signal intensity on all MRI sequences). However, this classification still remains to be proved. The consensus today is that adenomyosis can be strongly considered when JZ thickness is greater than 12 mm, although there is no definable JZ on MRI in approximately 20% of premenopausal women.15 A diagnosis can also be suspected when the thickness of the JZ is between 8 and 12 mm; if other signs, such as a relative thickening of JZ in a localized area, poor definition of the JZ borders, or high signal foci on T2- or T1-weighed sequences, are present.9

le JZ on MRI in approximately 20% of premenopausal women.15 A diagnosis can also be suspected when the thickness of the JZ is between 8 and 12 mm; if other signs, such as a relative thickening of JZ in a localized area, poor definition of the JZ borders, or high signal foci on T2- or T1-weighed sequences, are present.9 Uterine Junctional Zone (JZ) In 1983, Hricak et al.16 first described the functional uterine zone, which is the junction between the endometrium and the inner myometrium. Today, in healthy women of reproductive age, through T2-weighted images, 3 distinct layers were noted: (i) a high signal intensity corresponding to the endometrial stripe, (ii) an inner low signal intensity that is adjacent to the basal endometrium, the JZ or subendometrial layer, and (iii) an outer medium signal intensity subserosal zone, or outer myometrium.17 However, diffuse thickening of the JZ should be carefully distinguished from normal physiological change, since the thickness of JZ varies considerably during the menstrual cycle.18 de Souza et al.19 reported an incidence of 54% myometrial JZ hyperplasia (a clear sign of adenomyosis) in subfertile patients complaining of menorrhagia or dysmenorrhea.

the JZ should be carefully distinguished from normal physiological change, since the thickness of JZ varies considerably during the menstrual cycle.18 de Souza et al.19 reported an incidence of 54% myometrial JZ hyperplasia (a clear sign of adenomyosis) in subfertile patients complaining of menorrhagia or dysmenorrhea. Association With Infertility Diagnosing adenomyosis was difficult until recently, and in the past, it was associated with multiparous women not with infertility. Indeed, women often delay their first pregnancy, and adenomyosis is typically observed in their late 30s and 40s. When nonsurgical diagnosis, such as TVS and MRI, became possible, the role of adenomyosis in infertility and early miscarriage was recognized.20 CLINICAL EVIDENCE OF CORRELATION BETWEEN ADENOMYOSIS AND INFERTILITY Dysfunctional Uterine JZ and Fertility Outcome Chiang et al.21 suggested the link between spontaneous abortion rate and uterine JZ dysfunction in infertile patients undergoing in vitro fertilization (IVF) and found that the spontaneous abortion rate was higher in women with a diffusely enlarged uterus on ultrasound imaging without distinct uterine masses compared with those with a normal uterus. However, their pregnancy rates were not statistically significant. Piver et al.22 also proposed that MRI evaluation of JZ thickness is the best negative predictive factor of implantation failure, and an increase in JZ diameter is inversely correlated to implantation rate. Implantation failure was found to be high when the average junctional zone was greater than 7 mm.23

y significant. Piver et al.22 also proposed that MRI evaluation of JZ thickness is the best negative predictive factor of implantation failure, and an increase in JZ diameter is inversely correlated to implantation rate. Implantation failure was found to be high when the average junctional zone was greater than 7 mm.23 Achieving Pregnancy After Medical or Surgical Therapy Infertile women reportedly achieve pregnancy after being treated for adenomyosis. Since 1993, published case reports or small series report adenomyosis treated with gonadotrophin-releasing hormone (GnRH) analog alone, conservative surgery, or combined therapy (Tables 1–3).24–42 A live birth rate after treatment with GnRH agonist (GnRH-a) for 5 months was first reported by Silva et al.25 In a case study by Nelson and Corson,24 the patient with histologically diagnosed adenomyosis who underwent a long-term course of GnRH-a conceived shortly after cessation of treatment. It was also reported that 2 cases of adenomyosis were conceived within 6 months after a short course of GnRH therapy with buserelin.26 A small Japanese study, in which 3 of 4 infertile patients successfully conceived after using a danazol IUD, is also additional evidence to link adenomyosis and infertility.42 TABLE 1 Successful Pregnancies After GnRH-a in Women With Adenomyosis TABLE 2 Successful Pregnancies After Conservative Surgery in Women With Adenomyosis TABLE 3 Successful Pregnancies After Combined Therapy in Women With Adenomyosis

Achieving Pregnancy After Medical or Surgical Therapy Infertile women reportedly achieve pregnancy after being treated for adenomyosis. Since 1993, published case reports or small series report adenomyosis treated with gonadotrophin-releasing hormone (GnRH) analog alone, conservative surgery, or combined therapy (Tables 1–3).24–42 A live birth rate after treatment with GnRH agonist (GnRH-a) for 5 months was first reported by Silva et al.25 In a case study by Nelson and Corson,24 the patient with histologically diagnosed adenomyosis who underwent a long-term course of GnRH-a conceived shortly after cessation of treatment. It was also reported that 2 cases of adenomyosis were conceived within 6 months after a short course of GnRH therapy with buserelin.26 A small Japanese study, in which 3 of 4 infertile patients successfully conceived after using a danazol IUD, is also additional evidence to link adenomyosis and infertility.42 TABLE 1 Successful Pregnancies After GnRH-a in Women With Adenomyosis TABLE 2 Successful Pregnancies After Conservative Surgery in Women With Adenomyosis TABLE 3 Successful Pregnancies After Combined Therapy in Women With Adenomyosis Furthermore, Fujishita et al.29 described the modified reduction surgical technique for adenomyosis; and in this study, one patient conceived spontaneously 4 months after operation by H-incision technique. One prospective study used a triple-flap method for reconstructing the uterine wall for 104 patients with severe adenomyosis. Of these, 4 of 26 women who wished to conceive became pregnant after conservative surgery.34 Uterus-sparing surgery for adenomyosis-associated subfertile women were also demonstrated by Kishi et al.35 In the group aged younger than 39 years, 60.8% of women with a history of IVF failure achieved pregnancy after surgery, although there was no clear benefit of surgery on fertility outcomes for patients older than 40 years. Wang et al.30 also suggested that laparoscopic cytoreductive surgery might be suitable for women with localized adenomyosis who failed the usual infertility treatments and assisted reproductive technology (ART). They reported one spontaneous pregnancy that occurred 21 months after surgery.

patients older than 40 years. Wang et al.30 also suggested that laparoscopic cytoreductive surgery might be suitable for women with localized adenomyosis who failed the usual infertility treatments and assisted reproductive technology (ART). They reported one spontaneous pregnancy that occurred 21 months after surgery. Conservative surgery or combination treatment in subfertile women with adenomyosis also had significant benefits for not only controlling symptoms but also for increasing the pregnancy rate compared with GnRH-a alone.27,28 The cumulative 3-year clinical pregnancy rate and final successful delivery rate were higher in adenomyotic women who underwent conservative surgery with or without GnRH-a compared with those who received GnRH-a alone for 6 months.27 In one large prospective study, 55 of 165 patients with adenomyosis became pregnant after surgery followed by 6-course GnRH treatment or surgery alone, with a clinical pregnancy rate of 77.5%, and 49 women (69.0%) had a successful delivery by the end of the 2-year follow-up period.33 The combination of microsurgical cytoreduction and GnRH-a treatment could be appropriate for patients who failed GnRH-a alone or would not tolerate long-term GnRH-a treatment for presumed severe adenomyosis.40,41

rate of 77.5%, and 49 women (69.0%) had a successful delivery by the end of the 2-year follow-up period.33 The combination of microsurgical cytoreduction and GnRH-a treatment could be appropriate for patients who failed GnRH-a alone or would not tolerate long-term GnRH-a treatment for presumed severe adenomyosis.40,41 Effect of Adenomyosis on Reproductive Outcome After Endometriosis Surgery The impact of adenomyosis on pregnancy rates following surgery for both rectovaginal and colorectal endometriosis was reviewed in five articles from 2005 to 2010.32,43–46　Diagnosis of uterine adenomyosis based on TVS,44 MRI,32 and both TVS and MRI.43,45,46 Among the five selected studies, the criteria for the presence of adenomyosis were described in only one.32 Although fertility was not their primary study end point, in Landi et al.32 in 2008, a significantly higher pregnancy rate was recorded in women with endometriosis, but without adenomyosis, compared with those with concurrent endometriosis and adenomyosis after laparoscopic excision of deep, infiltrated endometriosis (Table 2). One prospective study evaluated the cumulative pregnancy rate after intracytoplasmic sperm injection (ICSI)/IVF in patients with colorectal endometriosis. Cumulative pregnancy rates in patients with associated adenomyosis was 19%, and in those who had endometriosis only, it was 82%, revealing that adenomyosis was associated with decreased cumulative pregnancy rate (Fig. 1).47 FIG. 1 Clinical pregnancies in women with adenomyosis.

Although fertility was not their primary study end point, in Landi et al.32 in 2008, a significantly higher pregnancy rate was recorded in women with endometriosis, but without adenomyosis, compared with those with concurrent endometriosis and adenomyosis after laparoscopic excision of deep, infiltrated endometriosis (Table 2). One prospective study evaluated the cumulative pregnancy rate after intracytoplasmic sperm injection (ICSI)/IVF in patients with colorectal endometriosis. Cumulative pregnancy rates in patients with associated adenomyosis was 19%, and in those who had endometriosis only, it was 82%, revealing that adenomyosis was associated with decreased cumulative pregnancy rate (Fig. 1).47 FIG. 1 Clinical pregnancies in women with adenomyosis. Presurgical hormone treatment (GnRH agonists for 3 months) was used in 2 studies.45,46 Surgery was performed in 3 studies with laparoscopy32,44,45 and with laparotomy or laparoscopy in 2 studies.44,46 In 3 of 5 studies,43–45 no pregnancy occurred in women older than 35 years, and in the study by Darai et al.,46 only 1 of 24 women who conceived was older than 35 years. Moreover, the percentage of women with adenomyosis seeking pregnancy was 4 of 18 in Darai et al.,45 8 of 17 in Ferrero et al.,43 11 of 20 in Darai et al.,46 and 10 of 20 in Stepniewska et al.44

h their menses. They do not have the space to dispose of soiled cloths or other materials47; neither do they have sufficient privacy to wash and dry used cloths as they would be able to do in rural areas.50 In some cases, girls even report exchanging sex for money so they can purchase commercial sanitary products.19,23 INTERVENTIONS TO IMPROVE MENSTRUAL HYGIENE MANAGEMENT IN RESOURCE-POOR COUNTRIES We found only 11 studies evaluating interventions that tried to improve menstrual hygiene management or change menstrual hygiene practices. All of the interventions evaluated have been published since 2000. Seven of these studies were conducted in South Asia,53–59 whereas the other studies were done in Ghana,60 Tanzania,61 Iran,62 and Saudi Arabia63 (Table 2). TABLE 2 Summary Results of Intervention Articles Included in the Review

ed in women older than 35 years, and in the study by Darai et al.,46 only 1 of 24 women who conceived was older than 35 years. Moreover, the percentage of women with adenomyosis seeking pregnancy was 4 of 18 in Darai et al.,45 8 of 17 in Ferrero et al.,43 11 of 20 in Darai et al.,46 and 10 of 20 in Stepniewska et al.44 ART (IVF/ICSI) Outcome in Patients With Adenomyosis Many studies reported reproductive outcome after ART in women with adenomyosis.21,48,49 Most concluded that adenomyosis causes infertility, but more prospective studies with a large population should be performed to further evaluate this causal interaction and unravel the mechanisms responsible for this negative effect. One of the first studies that connect the thickness of JZ in women with adenomyosis with embryo implantation failure came from Maubon et al.23 Two recent large population studies confirmed lower pregnancy rates in women with adenomyosis who underwent IVF.48,49 There are significantly lower clinical pregnancy rates in women with adenomyosis diagnosed by ultrasound (23.6%) compared with the nonadenomyosis group (44.6%) after stimulation with a GnRH antagonist protocol for IVF. Possibly, the age factor in older women with adenomyosis may be related to a lower pregnancy rate. This difference was still significant (P = 0.031) even after normalization of groups for age using regression analysis.49 In a study by Tremellen et al.,50 4 patients with adenomyosis who previously had multiple unsuccessful IVF cycles promptly resulted in successful pregnancy with ART after prolonged down-regulation with GnRH-a.

ence was still significant (P = 0.031) even after normalization of groups for age using regression analysis.49 In a study by Tremellen et al.,50 4 patients with adenomyosis who previously had multiple unsuccessful IVF cycles promptly resulted in successful pregnancy with ART after prolonged down-regulation with GnRH-a. However, Costello et al.51 investigated the effect of TVS-diagnosed adenomyosis on subsequent IVF/ICSI outcome. After a single cycle of IVF/ICSI, reproductive outcome was compared in women with and women without adenomyosis excluding patients with severe endometriosis. There was no difference in live birth rate per patient (cycle) between the 2 groups. Mijatovic et al.52 also showed reproductive outcome in infertile patients with surgically proven endometriosis who were pretreated with more than 3 months of long-term pituitary down-regulation (GnRH-a) before IVF/ICSI. No significant differences in IVF/ICSI outcome were observed between women with and women without adenomyosis. However, in this study, the authors point out that endometriosis was an important confounder, and 90.4% of the patients with endometriosis were surgically staged with moderate to severe endometriosis (revised American Society of Reproductive Medicine stages III–IV; Figs. 1, 2). Having a study population of asymptomatic women with adenomyosis undergoing IVF attempts, Benaglia et al.53 failed to show a detrimental effect of the disease in the pregnancy rate of these women. The authors proposed that GnRH-a pretreatment might have a therapeutic effect on adenomyosis.

e stages III–IV; Figs. 1, 2). Having a study population of asymptomatic women with adenomyosis undergoing IVF attempts, Benaglia et al.53 failed to show a detrimental effect of the disease in the pregnancy rate of these women. The authors proposed that GnRH-a pretreatment might have a therapeutic effect on adenomyosis. FIG. 2 Ongoing pregnancies in women with adenomyosis. Association With Early Miscarriage Two recent prospective studies concluded that adenomyosis reduces implantation and number of embryos transferred, clinical pregnancy rate, and ongoing pregnancy rate in women undergoing IVF. The first trimester miscarriage rate was also higher in women with adenomyosis compared with the control group.48,54 Martínez-Conejero et al.55 examined the endometrium of women with adenomyosis undergoing oocyte donation and found a similar endometrial gene expression pattern in both the adenomyosis and the control group. The implantation and clinical pregnancy rates were not different in all groups, but the miscarriage rate was significantly higher in the adenomyosis group (13.1%) than adenomyosis + endometriosis group (6.1%) and the controls (7.2%). The term pregnancy rate was also lower in the adenomyosis group (26.8%) than in the adenomyosis + endometriosis (38%) and the control group (37.1%), showing that implantation is not affected by adenomyosis. However, the higher rate of miscarriages associated with this condition led to lower term pregnancy rates, indicating a negative effect on the final outcome of oocyte donation.

(26.8%) than in the adenomyosis + endometriosis (38%) and the control group (37.1%), showing that implantation is not affected by adenomyosis. However, the higher rate of miscarriages associated with this condition led to lower term pregnancy rates, indicating a negative effect on the final outcome of oocyte donation. Lifelong Primary Infertility in Baboons After assessing their necropsy records, 37 baboons diagnosed with adenomyosis and 38 baboons with normal uteri histology were compared in one case-control study. The authors analyzed the association between adenomyosis, primary infertility, and the presence of coexisting endometriosis. They found that adenomyosis is strongly associated with lifelong infertility (P < 0.001) and was maintained even after excluding coexisting endometriosis cases (N = 17). The weakness in that study lies in the selection of the negative controls. Unless uteri are exhaustively sectioned, adenomyosis cannot be excluded. However, this study showed that adenomyosis is strongly associated with the presence of endometriosis and lifelong infertility.56

g coexisting endometriosis cases (N = 17). The weakness in that study lies in the selection of the negative controls. Unless uteri are exhaustively sectioned, adenomyosis cannot be excluded. However, this study showed that adenomyosis is strongly associated with the presence of endometriosis and lifelong infertility.56 PROPOSED MECHANISM OF INFERTILITY IN PATIENTS WITH ADENOMYOSIS Abnormal Uterotubal Transport Intrauterine Abnormalities Anatomical distortion of the uterine cavity may be one important factor leading to infertility, although the mechanism by which uterine adenomyoma has a detrimental effect on reproductive function remains unknown. Adenomyoma that distorts the uterine cavity may obstruct the tubal ostia and interfere with sperm migration and embryo transport. Several studies have demonstrated that submucosal and intramural fibroids in the presence of endometrial cavity distortion are associated with reduced implantation and pregnancy rates.57,58

Adenomyoma that distorts the uterine cavity may obstruct the tubal ostia and interfere with sperm migration and embryo transport. Several studies have demonstrated that submucosal and intramural fibroids in the presence of endometrial cavity distortion are associated with reduced implantation and pregnancy rates.57,58 Disturbed Uterine Peristalsis and Sperm Transport Using transabdominal ultrasound, the presence of distinct contraction waves in the myometrium can be seen. This peristaltic activity originates solely from the JZ, whereas the outer myometrium remains inactive.59 Directed sperm transport toward the peritoneal opening of the tubes on the side of dominant follicle by uterine peristalsis is fundamental to the early reproductive process, and it depends on the architecture of the myometrial wall with circular muscular fibers.60 In women with adenomyosis, normal architecture of the “archimyometrium” (JZ myometrium) was destroyed owing to invagination of the endometrial glands and stroma. This gives rise to the development of hyperplastic muscular tissue that causes dysfunctional uterine hyperperistalsis with increased intrauterine pressure.61 It seems reasonable to suppose that those changes may affect fertility in patients with adenomyosis. Increased uterine JZ activity just before embryo transfer in IVF is also associated with a reduced pregnancy rate and increased frequency of ectopic pregnancy.62

hyperperistalsis with increased intrauterine pressure.61 It seems reasonable to suppose that those changes may affect fertility in patients with adenomyosis. Increased uterine JZ activity just before embryo transfer in IVF is also associated with a reduced pregnancy rate and increased frequency of ectopic pregnancy.62 Destruction of Normal Myometrial Architecture and Function Mehasseb et al.63 studied the ultrastructural features of myometrium in the presence or absence of adenomyosis. Myocytes of adenomyosis uteri are ultrastructurally different from those of normal uteri. In the presence of adenomyosis, JZ showed cellular and nuclear hypertrophy, abnormal nuclear and mitochondrial shape, abundant myelin bodies, and other abnormalities. This suggests that those ultrastructural abnormalities may cause a disturbance in normal calcium cycling in affected myocytes, with subsequent loss of normal rhythmic contraction, eventually affecting uterotubal transport.

rophy, abnormal nuclear and mitochondrial shape, abundant myelin bodies, and other abnormalities. This suggests that those ultrastructural abnormalities may cause a disturbance in normal calcium cycling in affected myocytes, with subsequent loss of normal rhythmic contraction, eventually affecting uterotubal transport. Altered Endometrial Function and Receptivity Aberrant Endometrial Metabolism Altered Endometrial Steroid Metabolism Kitawaki et al.64 demonstrated the expression of aromatase cytochrome P450 (P450arom) protein and mRNA only in adenomyotic tissues and the eutopic endometrium of patients with adenomyosis, but not in the normal endometrium of women without adenomyosis. P450arom is an enzyme that catalyzes the conversion of androgens to estrogens. Takahashi et al.65 found menstrual blood estradiol levels were highest in women with adenomyosis, whereas they were within normal levels in peripheral blood, suggesting that local estrogen production may increase estrogen concentration in the menstrual blood of women with adenomyosis. Lessey et al.66 also showed that overexpression of P450 aromatase increases local estrogen production within the endometrium. Clinical pregnancy rates were statistically lower in women with high endometrial P450arom mRNA levels,67 and they suggest that P450arom mRNA expression can identify women at increased risk of IVF failure. Moreover, treatment with GnRH agonist or danazol decreased expression of P450arom in the eutopic endometrium of women with adenomyosis.68

es were statistically lower in women with high endometrial P450arom mRNA levels,67 and they suggest that P450arom mRNA expression can identify women at increased risk of IVF failure. Moreover, treatment with GnRH agonist or danazol decreased expression of P450arom in the eutopic endometrium of women with adenomyosis.68 Abnormal Inflammatory Response Macrophages have the capacity to produce not only proinflammatory cytokines, such as TNF-α and IL-1, but also reactive oxygen species that can be toxic to embryos.69,70 Tremellan et al.70 reported women with severe adenomyosis who, after a failed implantation, have significantly greater stromal macrophage density than the nonadenomyosis controls. Additionally, IL-6 mRNA expression was increased in macrophage-cocultured endometriotic stromal cells in adenomyosis.71 Wang et al.72 have shown that IL-10 staining intensity in women with adenomyosis was higher in epithelial cells of both eutopic and ectopic endometrium than controls, suggesting that an abnormal inflammatory response may impair fertility (Fig. 3). FIG. 3 Implantation factors showing altered concentration in adenomyosis-associated infertility.

Abnormal Inflammatory Response Macrophages have the capacity to produce not only proinflammatory cytokines, such as TNF-α and IL-1, but also reactive oxygen species that can be toxic to embryos.69,70 Tremellan et al.70 reported women with severe adenomyosis who, after a failed implantation, have significantly greater stromal macrophage density than the nonadenomyosis controls. Additionally, IL-6 mRNA expression was increased in macrophage-cocultured endometriotic stromal cells in adenomyosis.71 Wang et al.72 have shown that IL-10 staining intensity in women with adenomyosis was higher in epithelial cells of both eutopic and ectopic endometrium than controls, suggesting that an abnormal inflammatory response may impair fertility (Fig. 3). FIG. 3 Implantation factors showing altered concentration in adenomyosis-associated infertility. Altered Expression of Estrogen and Progesterone Receptors Since IL-6 can activate the estrogen receptor in breast cancer cells,73 overexpressed IL-6 could lead to an increased estrogen receptor expression in adenomyosis women. Progesterone has antiproliferative activity through its receptors. Within the adenomyotic tissue, there was reduction in the expression of progesterone receptors (A and B) in all layers,74,75 resulting in up-regulation of ER-α, since down-regulation of estrogen receptor-α is one of the primary functions of progesterone. Overexpression of ER-α reduces β-3 integrin secretion and alters uterine receptivity.76

tissue, there was reduction in the expression of progesterone receptors (A and B) in all layers,74,75 resulting in up-regulation of ER-α, since down-regulation of estrogen receptor-α is one of the primary functions of progesterone. Overexpression of ER-α reduces β-3 integrin secretion and alters uterine receptivity.76 Altered Uterine Oxidative Stress Environment A low oxygen environment in the uterus needs to be maintained for implantation of fertilized eggs because an excessive free radical environment may damage fertilized eggs and interfere with embryo development.77 In a normal woman, concentrations of nitric oxide synthase, xanthine oxidase, superoxide dismutase are low during the proliferative phase and increase during the early and midsecretory phases.78,79 In women with adenomyosis, levels do not fluctuate during the menstrual cycle and are overexpressed.80 Abnormal levels of intrauterine free radicals seem to cause infertility in women with adenomyosis. Some studies with animal models demonstrated that excess or deficient levels of free radicals mediated by inflammatory factors in reproductive tissue may inhibit both embryo development in vitro and implantation in vivo, resulting in a low pregnancy rate (Fig. 3).81–83

Altered Uterine Oxidative Stress Environment A low oxygen environment in the uterus needs to be maintained for implantation of fertilized eggs because an excessive free radical environment may damage fertilized eggs and interfere with embryo development.77 In a normal woman, concentrations of nitric oxide synthase, xanthine oxidase, superoxide dismutase are low during the proliferative phase and increase during the early and midsecretory phases.78,79 In women with adenomyosis, levels do not fluctuate during the menstrual cycle and are overexpressed.80 Abnormal levels of intrauterine free radicals seem to cause infertility in women with adenomyosis. Some studies with animal models demonstrated that excess or deficient levels of free radicals mediated by inflammatory factors in reproductive tissue may inhibit both embryo development in vitro and implantation in vivo, resulting in a low pregnancy rate (Fig. 3).81–83 Impaired Implantation Lack of Expression of Adhesion Molecules Several cell adhesion molecules (integrin, selectin, and cadherin) expressed by the endometrium are necessary for embryo and endometrium interaction. Integrins are the best studied markers of endometrial receptivity.84 Abnormal endometrial expression of integrin subtype α-5 and β-3 takes place in patients with IVF failure despite good embryo quality.85 Osteopontin (OPN) is a small integrin-binding ligand, N-linked glycoprotein in the endometrium. It binds to integrin 3, giving rise to speculation that it may mediate trophoblast endometrial interaction during implantation.86 Integrin β-3 and OPN levels were significantly lower in patients with adenomyosis than in controls. Dysregulation of both integrin β-3 and OPN mRNA and protein in the endometrium during the implantation window suggests that adenomyosis is associated with impaired implantation.87

trial interaction during implantation.86 Integrin β-3 and OPN levels were significantly lower in patients with adenomyosis than in controls. Dysregulation of both integrin β-3 and OPN mRNA and protein in the endometrium during the implantation window suggests that adenomyosis is associated with impaired implantation.87 Reduced Expression of Implantation Markers Another factor associated with endometrial receptivity, leukemia inhibitory factor (LIF),88 is present during the implantation window. Leukemia inhibitory factor is an essential cytokine for successful egg implantation during human reproduction.89 In addition, adenomyotic endometrium shows abnormalities in the production of LIF, which may contribute to altering uterine receptivity.90 Leukemia inhibitory factor concentrations in uterine flushing are lower in women with infertility than in controls. Additionally, LIF expression is lower in the endometrium of women with adenomyosis during the midsecretory phase.89 Leukemia inhibitory factor and IL-6 expression is controlled in endometrial cells by nuclear factor kappa B (NF-kB) activation. Nuclear factor kappa B is a transcription factor and critical regulator of innate immune response and inflammation. Ponce et al.91 reported that NF-kB binding, phosphorylated NF-kB, and IL-6 expression were down-regulated in the late secretory phase in the eutopic endometrium of women with endometriosis. Nuclear factor kappa B activities in the eutopic endometrium of patients with adenomyosis are an intriguing target for future investigation.

l.91 reported that NF-kB binding, phosphorylated NF-kB, and IL-6 expression were down-regulated in the late secretory phase in the eutopic endometrium of women with endometriosis. Nuclear factor kappa B activities in the eutopic endometrium of patients with adenomyosis are an intriguing target for future investigation. Altered Function of the Gene for Embryonic Development HOXA 10 gene, essential for embryonic uterine development and proper adult endometrial growth during the menstrual cycle, may be involved in creating an impairment of implantation in women with adenomyosis.92 Its expression is necessary for endometrial receptivity, and it is significantly lower in the endometrial stroma of women with adenomyosis compared with fertile controls.93 PHARMACOLOGICAL AND SURGICAL TREATMENT OF ADENOMYOSIS Hormonal Treatments GnRH Agonists A lack of randomized, controlled trials exploring the impact of GnRH-a treatment on fertility hinders our understanding of adenomyosis. One case study reported that treatment of severe adenomyosis with GnRH-a for 16 weeks resulted in the live birth of a healthy male infant.39 Progestogens Women with adenomyosis are characterized by a lower expression of progesterone receptors (A and B) in endometrial lesions, but also in the outer and inner layers of myometrium. The treatment of adenomyosis with progesterone may be restricted owing to the abnormal expression of progesterone receptors.94

PHARMACOLOGICAL AND SURGICAL TREATMENT OF ADENOMYOSIS Hormonal Treatments GnRH Agonists A lack of randomized, controlled trials exploring the impact of GnRH-a treatment on fertility hinders our understanding of adenomyosis. One case study reported that treatment of severe adenomyosis with GnRH-a for 16 weeks resulted in the live birth of a healthy male infant.39 Progestogens Women with adenomyosis are characterized by a lower expression of progesterone receptors (A and B) in endometrial lesions, but also in the outer and inner layers of myometrium. The treatment of adenomyosis with progesterone may be restricted owing to the abnormal expression of progesterone receptors.94 Dienogest Dienogest (progestin) has been used to treat adenomyosis pharmacologically. Dienogest directly inhibited cellular proliferation and induced apoptosis in human adenomyotic stromal cells.95 Two nonrandomized studies on a small number of patients have been published, but neither refers to the patients' fertility. The first study compared 2 groups of approximately 20 subjects treated with danazol and dienogest for adenomyosis.96 That study did not clearly describe the effectiveness of the therapeutic protocols. Adenomyosis patients treated with dienogest are at higher risk of discontinuation owing to uterine bleeding. The second study presented a correlation of 3 factors: age younger than 38 years, lower hemoglobin levels before the starting point of the therapeutic protocol, and estradiol levels after 3 months of dienogest therapy.97

is patients treated with dienogest are at higher risk of discontinuation owing to uterine bleeding. The second study presented a correlation of 3 factors: age younger than 38 years, lower hemoglobin levels before the starting point of the therapeutic protocol, and estradiol levels after 3 months of dienogest therapy.97 Levonorgestrel Intrauterine System Levonorgestrel intrauterine system (LNG-IUS) is approved for treating women with adenomyosis who have completed their childbearing. Levonorgestrel intrauterine system treatment is accompanied by decreased pain and heavy uterine bleeding, which could be explained by the following mechanisms: (i) a progestogenic influence on adenomyosis foci; (ii) atrophy of the eutopic endometrium; and (iii) controlling of endometrial factors that changed during adenomyosis. Choi et al.98 described decreased expression of growth factor and the related receptors in women with heavy bleeding and adenomyosis after LNG-IUS treatment. In another randomized study, Maia et al.99 described a positive effect of LNG-IUS in approximately 100 women with adenomyosis experiencing heavy menstrual bleeding.

oi et al.98 described decreased expression of growth factor and the related receptors in women with heavy bleeding and adenomyosis after LNG-IUS treatment. In another randomized study, Maia et al.99 described a positive effect of LNG-IUS in approximately 100 women with adenomyosis experiencing heavy menstrual bleeding. Exploring Surgery Treatments and Pregnancy Rates Conservative Surgery Alone The results of conservative surgery, which consists of laparoscopy or laparotomy, are based on 3 studies in women with a diagnosis with adenomyosis. Two studies reported birth rate,31,100 and one reported pregnancy rate.38 The surgical techniques presented in all the studies consisted of excision of the adenomyoma and hysteroplasty using laparoscopy or laparotomy. The overall average birth rate reached 36.2% (21 of 58) in women who had this surgery.

th adenomyosis. Two studies reported birth rate,31,100 and one reported pregnancy rate.38 The surgical techniques presented in all the studies consisted of excision of the adenomyoma and hysteroplasty using laparoscopy or laparotomy. The overall average birth rate reached 36.2% (21 of 58) in women who had this surgery. Comparison of 2 Surgical Techniques A retrospective study of 104 patients undergoing conservative surgery compared the classical adenomyomectomy with a new technique, the H-incision surgery. The classical technique was performed in 5 women with adenomyosis who were selected retrospectively among 104 patients. The newer technique was used in 6 of 83 patients who desired to preserve fertility. The classical technique involves incision of the uterine wall and a stepwise resection of adenomyomatic tissue. The newer technique consists of an H-shape incision and excision of the adenomyomatic tissue. In this study, the newer technique resulted in one spontaneous pregnancy 4 months after operation compared with no pregnancy in women undergoing the classical technique. (Classical technique pregnancy rate, 0.14 [95% confidence interval].) Time between surgery and pregnancy was 4 to 6 months followed by a live birth, with one continuing pregnancy at the time the study finished.29

pregnancy 4 months after operation compared with no pregnancy in women undergoing the classical technique. (Classical technique pregnancy rate, 0.14 [95% confidence interval].) Time between surgery and pregnancy was 4 to 6 months followed by a live birth, with one continuing pregnancy at the time the study finished.29 Adenomyomectomy When women experience the severe symptoms of advanced adenomyosis, hysterectomy has been advised. However, the more conservative adenomyomectomy preserves the uterus. One study reported that from a pool of 103 patients, 55.34% presented with infertility, and of those, 16.50% had IVF failure, 8.74% had recurrent miscarriages, and 19.42% had abnormal uterine bleeding. Of the 103 patients, 70 attempted pregnancy, 21 naturally through intercourse, and 49 through IVF. Pregnancy outcomes were 30% pregnancy, with 23% live births. The symptoms of dysmenorrhea/hypermenorrhea lessened after surgery. Only one patient had a recurrence of adenomyosis.36

nd 19.42% had abnormal uterine bleeding. Of the 103 patients, 70 attempted pregnancy, 21 naturally through intercourse, and 49 through IVF. Pregnancy outcomes were 30% pregnancy, with 23% live births. The symptoms of dysmenorrhea/hypermenorrhea lessened after surgery. Only one patient had a recurrence of adenomyosis.36 CONCLUSIONS In summary, adenomyosis is a common gynecological disorder with unclear etiology. Several studies have demonstrated that the presence of adenomyosis may impair the fertility by affecting the uterotubal transport and altering endometrial function and receptivity. Some indirect proofs have shown that women with adenomyosis have poor reproductive outcomes compared with those without adenomyosis. Based on limited available evidence, it has been reported that infertile women who experience adenomyosis achieved pregnancy after being treated with different strategies, indirectly revealing poor reproductive outcomes in women with adenomyosis. Furthermore, surgery could be effective in women with adenomyosis with a history of IVF failure, although latter finding could be partly attributed to the higher rate of early miscarriage.

pregnancy after being treated with different strategies, indirectly revealing poor reproductive outcomes in women with adenomyosis. Furthermore, surgery could be effective in women with adenomyosis with a history of IVF failure, although latter finding could be partly attributed to the higher rate of early miscarriage. However, in the clinical situation, it is still difficult to determine whether adenomyosis is the cause of the infertility or not because, as previous studies have made us aware, unknown or as yet undiagnosed cases of endometriosis may be present in both cases and controls. Additionally, studies on treatment are limited to case series and a retrospective data without control groups. Better studies are needed to determine the molecular mechanism of implantation failure in women with adenomyosis and the impact of adenomyosis on infertile women with or without endometriosis. All authors and staff in a position to control the content of this CME activity and their spouses/life partners (if any) have disclosed that they have no financial relationships with, or financial interests in, any commercial organizations pertaining to this educational activity.

Menstruation is a normal biological process experienced by millions of women and girls around the world each month. Menarche signifies the start of a female’s reproductive years and often marks her transition to full adult female status within a society. Age at menarche tends to be slightly later in resource-poor countries than in higher-income countries,1 yet an apparent decline in age at menarche has been documented in both developed2,3 and developing countries4 over the past few decades. One challenge of menstruation that is taken for granted in affluent countries is the simple question of how to manage or contain the menstrual flow and what happens to a girl or woman who is not able to do this successfully. The United Nations defines adequate menstrual hygiene management as “women and adolescent girls using a clean menstrual management material to absorb or collect blood that can be changed in privacy as often as necessary for the duration of the menstruation period, using soap and water for washing the body as required, and having access to facilities to dispose of used menstrual management materials.”5 Particularly in poor countries, girls and women face substantial barriers to achieving adequate menstrual management.

ften as necessary for the duration of the menstruation period, using soap and water for washing the body as required, and having access to facilities to dispose of used menstrual management materials.”5 Particularly in poor countries, girls and women face substantial barriers to achieving adequate menstrual management. Gender equity in education has long been heralded as a cornerstone for social and economic development. The education of girls and women holds a prominent position in both the United Nations’ Millennium Development Goals and in the recently adopted Sustainable Development Goals.6 Although much progress has been made since 2000, in many countries (especially in South Asia and sub-Saharan Africa) a large number of girls either never attend school or attend only a few years of primary school before dropping out.7 Furthermore, the number of out-of-school girls is rising again after years of improvement.7 Earlier menarche and a greater emphasis on education mean that many adolescent girls are in school while menstruating. With a typical menstrual cycle lasting 25 to 30 days within which bleeding occurs for 4 to 6 days,8 postmenarcheal girls will experience menstrual bleeding on at least some school days every month. Menstrual hygiene management is therefore an increasingly important (yet often unrecognized) issue that is heavily intertwined with girls’ education, empowerment, and social development.

h bleeding occurs for 4 to 6 days,8 postmenarcheal girls will experience menstrual bleeding on at least some school days every month. Menstrual hygiene management is therefore an increasingly important (yet often unrecognized) issue that is heavily intertwined with girls’ education, empowerment, and social development. EVIDENCE ACQUISITION AND EXTRACTION As part of an ongoing program of menstrual hygiene management research and intervention in Ethiopia (www.dignityperiod.org), we carried out a systematic review of the published literature on menstrual hygiene management to understand the current state of knowledge and gaps in evidence surrounding these practices in resource-poor countries. We searched PubMed with the help of a reference librarian for English-language articles published through December 2015 using the MeSH search terms hygiene or menstrual hygiene products and menstruation. The PubMed search included the following: (("Hygiene"[MeSH] OR "Menstrual Hygiene Products"[MeSH]) AND English[lang]) AND ("Menstruation"[MeSH] AND English[lang]) AND (Journal Article[ptyp] AND English[lang]).

shed through December 2015 using the MeSH search terms hygiene or menstrual hygiene products and menstruation. The PubMed search included the following: (("Hygiene"[MeSH] OR "Menstrual Hygiene Products"[MeSH]) AND English[lang]) AND ("Menstruation"[MeSH] AND English[lang]) AND (Journal Article[ptyp] AND English[lang]). Our electronic search identified 199 unique citations that we then screened for relevance based on title and abstract (Fig. 1). Full-text screening was conducted of 58 articles, of which 27 were retained for inclusion and data extraction. We hand searched the reference lists of these articles plus key background articles to identify an additional 58 citations for screening. From the combined electronic and hand search, we identified 67 articles for inclusion that met the dual criteria of focusing on menstrual hygiene management in the setting of a low- or middle-income country.9 These 67 articles include 43 descriptive studies, 11 intervention evaluation studies, 11 commentaries/editorials, and 2 review articles. For each included article, we extracted essential information into a spreadsheet to facilitate analysis. Extracted information for all articles included bibliographic information, research question or purpose, and information on the setting or location in which it was carried out. For descriptive studies, intervention evaluations, and review articles, we also extracted information on study design, sample characteristics, results, and limitations. We also extracted a description of the intervention that was used in those evaluation studies.

on the setting or location in which it was carried out. For descriptive studies, intervention evaluations, and review articles, we also extracted information on study design, sample characteristics, results, and limitations. We also extracted a description of the intervention that was used in those evaluation studies. FIG. 1 Article identification flowchart. MHM, menstrual hygiene management.

on the setting or location in which it was carried out. For descriptive studies, intervention evaluations, and review articles, we also extracted information on study design, sample characteristics, results, and limitations. We also extracted a description of the intervention that was used in those evaluation studies. FIG. 1 Article identification flowchart. MHM, menstrual hygiene management. THE CURRENT STATUS OF MENSTRUAL HYGIENE MANAGEMENT IN RESOURCE-POOR COUNTRIES Most of the existing literature is descriptive in nature, explaining menstrual hygiene practices, knowledge, and attitudes—including beliefs and cultural taboos—and where girls get their information about menstruation. There are also observational studies that look at the associations between menstrual hygiene management practices and various sociodemographic and contextual characteristics (eg, lack of privacy, water, and/or proper sanitary disposal at school). Many of these studies are school based and often compare urban versus rural schools and/or private versus public schools. These articles frequently conclude that menstrual hygiene management is worse for girls in rural areas and for those who attend public schools (which tend to serve families of lower socioeconomic status). Studies are heavily concentrated in a handful of sub-Saharan African countries and the South Asia region. Very little is published in English from Latin America, North Africa and the Middle East, or Central Asia. The academic literature has recently been paying more attention to issues surrounding menstrual hygiene management. Only 10 of the 67 articles included in this review were published prior to 2000.

uth Asia region. Very little is published in English from Latin America, North Africa and the Middle East, or Central Asia. The academic literature has recently been paying more attention to issues surrounding menstrual hygiene management. Only 10 of the 67 articles included in this review were published prior to 2000. Girls in resource-poor countries around the world tend to use old cloths, tissue paper, cotton or wool pieces, or some combination of these items to manage their menstrual bleeding10–31 (Table 1). Egypt appears to be an exception to this pattern, where a large majority of girls report using commercially produced sanitary pads or napkins instead of homemade menstrual hygiene devices.48,52 Qualitative studies indicate that girls who know about commercial sanitary products may prefer these products because they are seen as more comfortable and less likely to leak, but for many girls such products are usually unavailable and/or unaffordable.13,19,24,40,44 TABLE 1 Summary Results of Descriptive Articles Included in the Review

Girls in resource-poor countries around the world tend to use old cloths, tissue paper, cotton or wool pieces, or some combination of these items to manage their menstrual bleeding10–31 (Table 1). Egypt appears to be an exception to this pattern, where a large majority of girls report using commercially produced sanitary pads or napkins instead of homemade menstrual hygiene devices.48,52 Qualitative studies indicate that girls who know about commercial sanitary products may prefer these products because they are seen as more comfortable and less likely to leak, but for many girls such products are usually unavailable and/or unaffordable.13,19,24,40,44 TABLE 1 Summary Results of Descriptive Articles Included in the Review Use of commercially produced sanitary pads is reported more commonly among girls in private schools, which typically serve wealthier families43; among those in urban areas41; and among girls who have received explicit training in how to use commercial sanitary products.46 The evidence is mixed, however, from urban slums where some studies report higher use of sanitary pads,28 whereas others report higher use of old cloths.42 Furthermore, knowledge about menstruation and menstrual hygiene tends to be higher in girls from urban areas compared with rural girls25,36,41 and in older as compared with younger adolescent girls.22,25,29

slums where some studies report higher use of sanitary pads,28 whereas others report higher use of old cloths.42 Furthermore, knowledge about menstruation and menstrual hygiene tends to be higher in girls from urban areas compared with rural girls25,36,41 and in older as compared with younger adolescent girls.22,25,29 Girls from resource-poor countries around the world attribute frequent school absences to difficulties managing their menses. In a Ugandan study of rural schoolgirls, nearly two-thirds said they miss school at least once per month because of menstruation.34 In India, only 54% of girls reported attending school while menstruating.21 In Egypt, more than one-third of girls in an urban secondary school reported staying home from school on the first day of menstruation.52 Similarly, in Amhara province, Ethiopia, more than half of girls in secondary and preparatory schools reported being absent during menstruation,36 and those who did not use sanitary pads were more than 5 times as likely to be absent.33 Even if girls are not absent and manage to attend school during menstruation, they report being distracted, unable to concentrate, and less willing to participate because, for example, standing to answer questions is the custom in many schools, and writing on a blackboard in front of the class may expose them revealing menstrual stains, leakage, or odors.19,23,24,33,34

chool during menstruation, they report being distracted, unable to concentrate, and less willing to participate because, for example, standing to answer questions is the custom in many schools, and writing on a blackboard in front of the class may expose them revealing menstrual stains, leakage, or odors.19,23,24,33,34 Absenteeism appears to be closely associated with lack of privacy and limited availability of water and sanitation facilities at schools. In Malawi, girls who reported that school toilets lacked privacy were more than twice as likely to be absent during their menstrual periods than girls at schools where more privacy was available.38 In Uganda, girls cited a lack of privacy and washing space, fear of leakage and stains, discomfort, and a lack of pads as reasons for school absences during menstruation.34,40 Given the opportunity to design their ideal toilet for school, girls emphasize the need for better lighting in latrines to be able to spot leaks and clean themselves34,40; more privacy including doors on latrines and functioning locks for the doors34,39,44; a water supply within the latrine in order to wash soiled hands, legs, or clothes34,39,44; lack of soap34,39,44 and toilet paper34; and no disinfectant or cleaning supplies to clean latrines after use.39,44

and clean themselves34,40; more privacy including doors on latrines and functioning locks for the doors34,39,44; a water supply within the latrine in order to wash soiled hands, legs, or clothes34,39,44; lack of soap34,39,44 and toilet paper34; and no disinfectant or cleaning supplies to clean latrines after use.39,44 While only a few studies have tested the relationship between infections and the type of material or product used to manage menstrual bleeding, those that have done so suggest that reusing old cloths may increase the risk of urogenital infections. In a case-control study from India, women with urogenital infections were twice as likely to have been using reusable cloths instead of disposable sanitary pads.32 In another study of schoolgirls in India, 65.7% of homemade menstrual cloth users reported urogenital infections compared with only 12.3% of those using sanitary pads.45 Qualitatively, women and girls recognize that the way they manage their menstrual blood may be unhygienic, but they do not have better alternatives. For example, women in Zimbabwe expressed concerns about reusing old cloths and know that ironing or drying the cloths in the sun would be best, but they often avoid doing this because of embarrassment, a desire for secrecy, and/or a lack of electricity or coal to heat an iron.13 Other girls feel they must store or hide cloths in places they know are unhygienic so that they are readily available when they need them.20 Some women also insert newspaper or tissue paper into their vaginas to reduce the chance of menstrual leakage despite their concerns that this might not be safe and that the newspaper ink might cause cancer.13

re or hide cloths in places they know are unhygienic so that they are readily available when they need them.20 Some women also insert newspaper or tissue paper into their vaginas to reduce the chance of menstrual leakage despite their concerns that this might not be safe and that the newspaper ink might cause cancer.13 The qualitative studies provide strikingly insightful information about women’s and girls’ perspectives on menstrual hygiene. These studies complement the evidence from quantitative studies. Girls realize that their menses may lead to school absences or even to their leaving school altogether.35,44 Water and sanitation facilities at school are often so inadequate for menstrual hygiene management that some girls report carrying plastic bags of drinking water to use in the school latrines.39,44 Nonetheless, girls are often highly resourceful at making sanitary “pads” out of whatever materials are available if menses start unexpectedly while they are at school, often resorting to the emergency use of leaves or grass.19,23 These homemade options are often uncomfortable, leak frequently, and cause distress.19 Especially in rural areas, women and girls may not even know about the existence of commercially manufactured sanitary products,17 do not know how to use or dispose of them,35 and/or perceive these products to be unaffordable.44 In India, female residents of urban slums report particular challenges in dealing with their menses. They do not have the space to dispose of soiled cloths or other materials47; neither do they have sufficient privacy to wash and dry used cloths as they would be able to do in rural areas.50 In some cases, girls even report exchanging sex for money so they can purchase commercial sanitary products.19,23

INTERVENTIONS TO IMPROVE MENSTRUAL HYGIENE MANAGEMENT IN RESOURCE-POOR COUNTRIES We found only 11 studies evaluating interventions that tried to improve menstrual hygiene management or change menstrual hygiene practices. All of the interventions evaluated have been published since 2000. Seven of these studies were conducted in South Asia,53–59 whereas the other studies were done in Ghana,60 Tanzania,61 Iran,62 and Saudi Arabia63 (Table 2). TABLE 2 Summary Results of Intervention Articles Included in the Review Nearly all of the interventions were purely educational in nature, most of them taking place in or through the school setting. The 2 quasi-experimental educational interventions both reported better menstrual health62 and hygiene practices63 among the intervention groups. The studies that assessed menstrual hygiene knowledge and practice on a pretest-posttest basis within a single intervention group all reported some improvements, ranging from an increase in knowledge54,57,58 and implementation of hygienic practices (such as drying cloths in the sun and washing with soap and water)53,55 to an increased use of sanitary pads (either disposable or reusable) instead of old homemade cloths.53,54 One study also reported that school absenteeism during menses decreased between the pretest and posttest evaluations.54 Overall, there is moderate evidence that education-based interventions can improve menstrual hygiene knowledge and practices among schoolgirls in resource-poor countries.

of old homemade cloths.53,54 One study also reported that school absenteeism during menses decreased between the pretest and posttest evaluations.54 Overall, there is moderate evidence that education-based interventions can improve menstrual hygiene knowledge and practices among schoolgirls in resource-poor countries. The 3 interventions that distributed sanitary products—either pads of various materials59,60 or menstrual cups56—to schoolgirls all evaluated these interventions using quasi-experimental designs. In Ghana, a 3-arm trial showed that both in the pad-with-underwear-distribution-plus-education arm and in the education-only arm attendance improved significantly over the control subjects who received neither pads nor education; however, attendance rose more quickly in the pad-plus-education arm than in the education-only arm of the study.60 In India, an intervention provided schoolgirls with falalin cloth—a short, absorbent, woven cloth—for 3 months, followed by sanitary pads for 3 months. Absenteeism was highest when girls were using old traditional cloths. Absenteeism decreased when falalin cloth was available, but there were no absences among girls when using commercially produced sanitary pads.59 Interestingly, however, a greater proportion of the girls preferred the falalin cloths to the commercial sanitary pads.59 In Nepal, an intervention that distributed menstrual cups to schoolgirls and their mothers in the treatment group and education booklets to all girls failed to find an impact on school attendance. The majority of the girls did not report that the cups were convenient or easy to use.56 There is limited (but mixed) evidence to suggest that distribution of sanitary products may reduce school absenteeism among girls.

e treatment group and education booklets to all girls failed to find an impact on school attendance. The majority of the girls did not report that the cups were convenient or easy to use.56 There is limited (but mixed) evidence to suggest that distribution of sanitary products may reduce school absenteeism among girls. The study from Nepal is the only one we found that evaluated outcomes beyond menstrual hygiene management practices and school attendance. This study did not find any effect of the menstrual cup intervention on test scores, cramps, premenstrual symptoms, or self-esteem/empowerment indicators; however, actual use of the menstrual cups within the intervention group was relatively low.56 ADDITIONAL PERSPECTIVES Eight of the 11 commentaries and editorials on menstrual hygiene management were penned in the last decade, demonstrating how this issue has started to gain prominence. One commentary summarizes a history of menstrual hygiene products,64 and another focuses on differences in menstrual hygiene management between Western nations and other countries, mostly in sub-Saharan Africa.65 Another commentary compares experiences across cultures, albeit between countries in different regions, noting that old cloths and naturally absorbent substances are the most common materials used in menstrual hygiene management, although a few cultures utilize menstrual huts or otherwise exclude women from their usual social interactions during menstruation.66

oss cultures, albeit between countries in different regions, noting that old cloths and naturally absorbent substances are the most common materials used in menstrual hygiene management, although a few cultures utilize menstrual huts or otherwise exclude women from their usual social interactions during menstruation.66 India and Kenya have received attention over the last few years as these countries have moved to subsidize commercial sanitary products for rural girls and to remove the value-added tax on menstrual hygiene products, respectively. In India, the subsidy plan faces potential pitfalls due to the lack of knowledge and awareness of family members, teachers, and health care providers; lack of water and sanitation facilities available in schools; and lack of sufficient solid-waste disposal in villages.50 Other authors have called attention to the disconnect between focusing on menstrual hygiene management when there is still a serious lack of access to water and sanitation facilities in schools, especially in the South Asia region.67

ities available in schools; and lack of sufficient solid-waste disposal in villages.50 Other authors have called attention to the disconnect between focusing on menstrual hygiene management when there is still a serious lack of access to water and sanitation facilities in schools, especially in the South Asia region.67 Sommer68 has written 4 of these commentaries and has become a voice of advocacy for improved menstrual hygiene management for girls in low-resource environments. She has advocated for adding menstrual hygiene management to the agenda of access to clean water and improved sanitation in schools68 and to including menstrual hygiene management as part of the response to humanitarian emergencies.69 She and her colleagues have also argued that menstrual hygiene management should be part of any educational agenda for school-aged girls5 and should be promoted more broadly as a general public health issue.70

schools68 and to including menstrual hygiene management as part of the response to humanitarian emergencies.69 She and her colleagues have also argued that menstrual hygiene management should be part of any educational agenda for school-aged girls5 and should be promoted more broadly as a general public health issue.70 The findings from our review are consistent with and expand upon the other review articles that we found on this topic. One review that focused on water and sanitation in schools found that the availability of water and sanitation facilities in schools is a key determinant of girls’ school attendance in general and that lack of such facilities increases the challenges girls face with respect to managing menstrual hygiene.71 Studies included in this review reported that girls often experience discomfort at school during menstruation, have special concerns about privacy when handling menstrual issues, fear teasing from peers (both male and female), lack mechanisms for the proper disposal of menstrual products, and have insufficient water for cleaning themselves while menstruating.71

at girls often experience discomfort at school during menstruation, have special concerns about privacy when handling menstrual issues, fear teasing from peers (both male and female), lack mechanisms for the proper disposal of menstrual products, and have insufficient water for cleaning themselves while menstruating.71 One other review focused more directly on menstrual hygiene management, but found it difficult to define this term given the variation in definitions used across the included articles.72 Education interventions were found to improve knowledge, awareness, and some menstrual hygiene practices, but documenting the effect of menstrual hygiene management on school attendance and dropout rates is much more difficult, given how poorly records are often kept and the often-ambiguous reasons for school absences.72

ucation interventions were found to improve knowledge, awareness, and some menstrual hygiene practices, but documenting the effect of menstrual hygiene management on school attendance and dropout rates is much more difficult, given how poorly records are often kept and the often-ambiguous reasons for school absences.72 We limited our search to peer-reviewed articles published in English. There may be additional interventions reported to improve menstrual hygiene management among adolescent girls in resource-poor countries that have not been published in the peer-reviewed English literature. This may account for the disproportionate number of articles originating from countries in sub-Saharan Africa and South Asia. In addition, the lack of consistency between studies in terms of how “good” menstrual hygiene management is defined and measured makes it challenging to compare studies and generate accurate summaries. Because menstrual hygiene management is a relatively new (but vitally important) topic for research and advocacy, we expect these definitions and measurements will become more standardized as the field evolves and our understanding of the issue improves.

t challenging to compare studies and generate accurate summaries. Because menstrual hygiene management is a relatively new (but vitally important) topic for research and advocacy, we expect these definitions and measurements will become more standardized as the field evolves and our understanding of the issue improves. FUTURE RESEARCH AND ADVOCACY DIRECTIONS The existing literature on menstrual hygiene management in resource-poor settings highlights common challenges experienced across different cultures. Major barriers to improved menstrual hygiene among girls include a lack of awareness and support from teachers, many of whom are male; lack of familial support; lack of cultural acceptance of certain menstrual hygiene products; limited economic resources to purchase commercially produced products; inadequate water and sanitation facilities at school with concurrent concerns about washing, privacy, and menstrual pad disposal; cramps, pain, and discomfort associated with menstruation irrespective of the menstrual hygiene products used; and travel difficulties to/from school, which can extend time away from home and increase the likelihood of leaks/stains, embarrassment, and discomfort. Fear of embarrassment from menstrual accidents and teasing by peers is a common thread in many qualitative studies.

rrespective of the menstrual hygiene products used; and travel difficulties to/from school, which can extend time away from home and increase the likelihood of leaks/stains, embarrassment, and discomfort. Fear of embarrassment from menstrual accidents and teasing by peers is a common thread in many qualitative studies. Our review identified several important gaps in the existing evidence base concerning menstrual hygiene management. First, there is a lack of knowledge around the household decision-making process for school enrollment. How is this decision made for postmenarcheal girls and by whom? What is the level of knowledge and awareness around menstrual hygiene management among those who are making these decisions? We found 1 intervention that targeted menstrual hygiene education at mothers in addition to schoolgirls,56 but we did not find any that also targeted fathers, in-laws, and/or community leaders who might be influential in valuing (or hindering) girls’ access to education. Considering how culturally and religiously embedded beliefs and practices about menstruation are, it is important to understand better the decision-making process for girls’ school enrollment and to extend intervention efforts beyond female relatives only to include all of those involved in the decision-making process. Increasing male understanding of menstruation and the importance of menstrual hygiene management is likely to emerge as a key consideration in improving both school attendance and the availability of suitable water and sanitation facilities at schools in poor countries.

of those involved in the decision-making process. Increasing male understanding of menstruation and the importance of menstrual hygiene management is likely to emerge as a key consideration in improving both school attendance and the availability of suitable water and sanitation facilities at schools in poor countries. Second, most intervention studies have been school based and focused on the absenteeism of girls who are already enrolled in school. Few studies have looked at girls who are out of school and whether education and awareness efforts with these girls and their families before menarche might reduce the number of girls who drop out around the time menstruation begins. Postmenarcheal girls who are still enrolled in school are likely to be from families who are more supportive of girls’ education. Once a girl drops out of school, getting her reenrolled may be difficult, even if the menstrual hygiene management situation improves.

girls who drop out around the time menstruation begins. Postmenarcheal girls who are still enrolled in school are likely to be from families who are more supportive of girls’ education. Once a girl drops out of school, getting her reenrolled may be difficult, even if the menstrual hygiene management situation improves. Third, except for the trial from Nepal that reported low use of provided menstrual cups and few significant findings,56 there are almost no studies looking at outcomes other than absenteeism in schoolgirls. No studies have looked at girls’ school performance or their absenteeism rates with respect to absenteeism among boys. A few qualitative studies identify fear, embarrassment, and pain as important barriers to attending school during menstruation, but these are measured quantitatively less frequently. Other outcomes such as school performance, self-confidence/self-esteem, empowerment, quality of life, and genitourinary infections are almost entirely absent from the reporting. While many of these are culturally constructed and can be difficult to measure, improvements in these areas have the potential to confer important benefits on girls.

school performance, self-confidence/self-esteem, empowerment, quality of life, and genitourinary infections are almost entirely absent from the reporting. While many of these are culturally constructed and can be difficult to measure, improvements in these areas have the potential to confer important benefits on girls. Finally, the literature on water and sanitation regarding schools in resource-poor countries is currently disconnected from the literature on menstrual hygiene management. Much of the menstrual hygiene literature mentions the importance of having good access to water and sanitation in schools, but few menstrual hygiene interventions actually attempt to address these issues. Conversely, menstrual hygiene management does not appear to be a significant outcome measure in studies in the literature on water and sanitation. We did not identify a single intervention that tried to improve menstrual hygiene management by focusing on what seems to be the important triad of (1) improved education regarding menstrual hygiene for girls, teachers, parents, and other decision makers; (2) adequate menstrual hygiene supplies; and (3) clean water and improved sanitation in schools. The ability of girls and women to manage their menses hygienically and without disruption of their daily activities is taken for granted in affluent countries. In these countries, access to menstrual hygiene supplies is generally easy and affordable. This is not the case for women and girls in resource-poor countries. We suspect that this is also true in impoverished parts of the United States and other seemingly resource-rich nations. We are learning that poor menstrual hygiene management may have serious health and developmental consequences for adolescent girls. This seems to be an important factor in hindering the education and empowerment of women in the world’s poor countries. Because these issues affect half of the world’s people, they merit increased attention by educators, policy makers, and government officials. Improved menstrual hygiene management needs to be on the health agendas of all resource-poor countries, particularly as those countries strive to train large cadres of teachers and community-based/primary-care health workers.

people, they merit increased attention by educators, policy makers, and government officials. Improved menstrual hygiene management needs to be on the health agendas of all resource-poor countries, particularly as those countries strive to train large cadres of teachers and community-based/primary-care health workers. In the United States and other high-income countries, we must also strive to increase the awareness of the importance of menstrual hygiene management among clinicians so that they can better serve their patients, particularly those who are recent immigrants from resource-poor countries where menstrual hygiene management may be a serious issue. Lack of adequate menstrual hygiene management may also be present in the poorest parts of affluent countries, where access to menstrual supplies is taken for granted and where these problems may easily be overlooked. Clinicians everywhere should be active partners in promoting better access to menstrual hygiene products, better water and sanitation in schools, and better knowledge within local communities concerning the basic biology of menstruation and how to manage the problems that may arise with it. Finally, we seek to make clinical researchers aware of the existing knowledge gaps concerning menstrual hygiene management among the world’s poor, especially gaps in our knowledge concerning the relationship between menstrual hygiene management and health outcomes such as genitourinary infections, healthy social development, and sound mental health.

al researchers aware of the existing knowledge gaps concerning menstrual hygiene management among the world’s poor, especially gaps in our knowledge concerning the relationship between menstrual hygiene management and health outcomes such as genitourinary infections, healthy social development, and sound mental health. ACKNOWLEDGMENTS The authors thank Donghua Tao, reference librarian at Saint Louis University, for her assistance in developing and implementing the search strategy in PubMed. Dr. Wall has disclosed that he is the President of Dignity Period, a not-for-profit charity in Tigray, Ethiopia. He and his spouse have no financial relationships with, or financial interests in, any commercial organizations pertaining to this educational activity. The remaining authors, faculty, and staff in a position to control the content of this CME activity and their spouses/life partners (if any) have disclosed that they have no financial relationships with, or financial interests in, any commercial organizations pertaining to this educational activity. Lippincott CME Institute has identified and resolved all conflicts of interest concerning this educational activity.