Search

To assess associations between influenza vaccination during etiologically-relevant windows and selected major structural non-cardiac birth defects.We analyzed data from the National Birth Defects Prevention Study, a multisite, population-based case-control study, for 8233 case children diagnosed with a birth defect and 4937 control children without a birth defect with delivery dates during 2006-2011. For all analyses except for neural tube defects (NTDs), we classified mothers who reported influenza vaccination 1 month before through the third pregnancy month as exposed; the exposure window for NTDs was 1 month before through the first pregnancy month. For defects with five or more exposed case children, we used logistic regression to estimate propensity score-adjusted odds ratios (aORs) and 95% confidence intervals (CIs), adjusting for estimated delivery year and season; plurality; maternal age, race/ethnicity, smoking and alcohol use, low folate intake; and, for NTDs, folate antagonist medications.There were 334 (4.1%) case and 197 (4.0%) control mothers who reported influenza vaccination from 1 month before through the third pregnancy month. Adjusted ORs ranged from 0.53 for omphalocele to 1.74 for duodenal atresia/stenosis. Most aORs (11 of 19) were ≤1 and all adjusted CIs included the null. The unadjusted CIs for two defects, hypospadias and craniosynostosis, excluded the null. These estimates were attenuated upon covariate adjustment (hypospadias aOR: 1.25 (95% CI 0.89, 1.76); craniosynostosis aOR: 1.23 (95% CI: 0.88, 1.74)).Results for several non-cardiac major birth defects add to the existing evidence supporting the safety of inactivated influenza vaccination during pregnancy. Under-reporting of vaccination may have biased estimates downward.

BACKGROUND: Trends in the prevalence of hepatitis C virus (HCV) infection among women delivering live births may differ in rural vs. urban areas of the United States, but estimation of trends based on observed counts may lead to unstable estimates in rural counties due to small numbers. OBJECTIVES: The objective of the study was to use small area estimation methods to provide updated county-level prevalence estimates and, for the first time, trends in maternal HCV infection among live births by county-level rurality. METHODS: Cross-sectional natality data from 2016 to 2020 were used to estimate maternal hepatitis C prevalence using hierarchical Bayesian models with spatiotemporal random effects to produce annual county-level estimates of maternal HCV infection and trends over time. Models included a 6-Ievel rural–urban county classification, year, maternal characteristics and county-specific covariates. Data were analysed in 2022. RESULTS: There were 90,764/18,905,314 live births (4.8 per 1000) with HCV infection reported on the birth certificate. Hepatitis C prevalence was higher among rural counties as compared to urban counties. Rural counties had the largest annual increases in maternal hepatitis C prevalence (per 1000 births) from 2016 to 2020 (micropolitan: 0.39; noncore: 0.40), with smaller increases among less densely populated urban counties (medium metro: 0.28; small metro: 0.28) and urban counties (large central metro:0.11; large fringe metro: 0.14). CONCLUSIONS: The prevalence of maternal HCV infection was the highest in rural counties, and rural counties saw the greatest average prevalence increase during 2016–2020. County-level data can help in monitoring rural–urban trends in maternal HCV infection to reduce geographic disparities.

Introduction The prevalence of hepatitis C virus infection among women delivering live births in the U.S. may be higher in rural areas where county-level estimates may be unreliable. The aim of this study is to model county-level maternal hepatitis C virus infection among deliveries in the U.S. Methods In 2020, U.S. natality files (2010–2018) with county-level maternal residence information were used from states that had adopted the 2003 revised U.S. birth certificate, which included a field for hepatitis C virus infection present during pregnancy. Hierarchical Bayesian spatial models with spatiotemporal random effects were applied to produce stable annual county-level estimates of maternal hepatitis C virus infection for years when all states had adopted the revised birth certificate (2016–2018). Models included a 6-Level Urban–Rural County Classification Scheme along with the birth year and county-specific covariates to improve posterior predictions. Results Among approximately 32 million live births, the overall prevalence of maternal hepatitis C virus infection was 3.5 per 1,000 births (increased from 2.0 in 2010 to 5.0 in 2018). During 2016–2018, posterior predicted median county-level maternal hepatitis C virus infection rates showed that nonurban counties had 3.5–3.8 times higher rates of hepatitis C virus than large central metropolitan counties. The counties in the top 10th percentile for maternal hepatitis C virus rates in 2018 were generally located in Appalachia, in Northern New England, along the northern border in the Upper Midwest, and in New Mexico. Conclusions Further implementation of community-level interventions that are effective in reducing maternal hepatitis C virus infection and its subsequent morbidity may help to reduce geographic and rural disparities.

In many perinatal pharmacoepidemiologic studies, exposure to a medication is classified as “ever exposed” versus “never exposed” within each trimester or even over the entire pregnancy. This approach is often far from real-world exposure patterns, may lead to exposure misclassification, and does not to incorporate important aspects such as dosage, timing of exposure, and treatment duration. Alternative exposure modeling methods can better summarize complex, individual-level medication use trajectories or time-varying exposures from information on medication dosage, gestational timing of use, and frequency of use. We provide an overview of commonly used methods for more refined definitions of real-world exposure to medication use during pregnancy, focusing on the major strengths and limitations of the techniques, including the potential for method-specific biases. Unsupervised clustering methods, including k-means clustering, group-based trajectory models, and hierarchical cluster analysis, are of interest because they enable visual examination of medication use trajectories over time in pregnancy and complex individual-level exposures, as well as providing insight into comedication and drug-switching patterns. Analytical techniques for time-varying exposure methods, such as extended Cox models and Robins’ generalized methods, are useful tools when medication exposure is not static during pregnancy. We propose that where appropriate, combining unsupervised clustering techniques with causal modeling approaches may be a powerful approach to understanding medication safety in pregnancy, and this framework can also be applied in other areas of epidemiology.

PURPOSE: There is unmet need for decision support regarding medication use during pregnancy. We aimed to inform the development of a decision aid on oral corticosteroid (OCS) use during pregnancy through focus groups. METHODS: We invited patients from one health system who had a recent live birth and a condition for which OCSs may be prescribed (i.e., asthma or other autoimmune disease) to participate in focus groups. We conducted conventional qualitative content analysis of verbatim transcripts of the focus groups using inductive coding. RESULTS: There were 30 participants across five focus groups from May-June 2019. Women endorsed the need for patient-provider discussions about OCS use during pregnancy in which the provider shares risks and benefits and the patient makes her decision. Furthermore, women generally expressed support for patient-centered handouts about OCS use during pregnancy that the provider discusses with the patient. When considering whether to take OCSs in pregnancy, women had concerns about: the medication’s impact on their baby (e.g., miscarriage, birth defects, long-term effects), themselves (e.g., effects on mood, sleep, weight gain), pregnancy complications (e.g., preterm birth, increased blood pressure), and lactation. Women wanted information on OCSs (e.g., indications, length of treatment, and cost), alternative treatments, and risks of not taking OCSs. CONCLUSIONS: We established patient need for a decision aid on OCS use during pregnancy that providers can discuss with patients. To address patient concerns, the aid should at a minimum describe the medication’s impact on baby, including long-term effects, maternal health, pregnancy complications, and lactation.

BACKGROUND: Antidepressant use later in pregnancy has been associated with preeclampsia and postpartum hemorrhage (PPH) in some studies. OBJECTIVE: To evaluate the association between patterns of prenatal antidepressant dose across gestation and risk of precclampsia and PPH. METHODS: We utilized OptumLabs® Data Warehouse (2012–2016) administrative health care claims, identifying 226,932 singleton live-born deliveries for this retrospective cohort study. Antidepressant dispensing doses were converted to fluoxetine equivalents. Using k-means longitudinal, we identified women with similar patterns of antidepressant exposure, i.e., trajectory groups, during the first 20 and 35 gestational weeks. We estimated risk ratios (RR) and 95% confidence intervals (CI) for the association between trajectory groups and preeclampisa (20-week groups) and PPH (35-week groups), adjusting for demographics, comorbidities, and other psychotropic medications. Linear trend tests assessing increasing risk of the outcomes across groups were performed. RESULTS: Among 15,041 (6.6%) pregnancies exposed to an antidepressant, the following trajectory groups were identified: A-low exposure, starting pregnancy at ~10mg/day, with 1(st) trimester reduction/discontinuation, B-low sustained exposure of ~20 mg/day, C-moderate exposure (~40mg/day) with 1(st) trimester reduction/discontinuation, D-moderate sustained exposure of ~40 mg/day, and E-high sustained exposure of ~75mg/day. In the low exposure with reduction/discontinuation trajectory, risks were 8.2% for preeclampsia and 2.7% for PPH. Compared with this group, low, moderate, and high sustained trajectories were associated with preeclampsia [adjusted (a)RR: 1.17 (95% CI: 1.01, 1.34), aRR: 1.31 (95% CI: 1.12, 1.54), aRR: 1.41 (95% CI: 1.05, 1.90), respectively] and PPH [aRR: 1.32 (1.05, 1.66), aRR: 1.35 (95% CI: 1.03,1.78), aRR: 2.51 (95% CI: 1.69, 3.71), respectively]; p

Nhung TH Trinh,1 Hedvig ME Nordeng,1,2 Gretchen Bandoli,3,4 Kristin Palmsten,5 Malin Eberhard-Gran,6,7 Angela Lupattelli1 1PharmacoEpidemiology and Drug Safety Research Group, Department of Pharmacy, PharmaTox Strategic Research Initiative, Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway; 2Department of Child Health and Development, Norwegian Institute of Public Health, Oslo, Norway; 3Department of Pediatrics, University of California San Diego, La Jolla, CA, USA; 4Department of Family Medicine and Public Health, University of California, San Diego, La Jolla, CA, USA; 5HealthPartners Institute, Minneapolis, MN, USA; 6Norwegian Research Centre for Women’s Health, Women’s and Children’s Division, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 7Institute for Clinical Medicine, University of Oslo, Oslo, NorwayCorrespondence: Nhung TH Trinh, Department of Pharmacy, Faculty of Mathematics and Natural Sciences, University of Oslo, Post box 1068 Blindern, Oslo, 0316, Norway, Email nhung.trinh@farmasi.uio.noBackground: Few studies investigated longitudinal antidepressant exposure during pregnancy and included dosage in the assessment.Methods: We conducted a nationwide, registry-linkage study in Norway using data on antidepressant prescription fills in pregnancies lasting ≥ 32 weeks in women with a delivery between 2009 and 2018 who had a depression/anxiety diagnosis and antidepressant fills prior to pregnancy. Information on antidepressant exposure by week (measured by filled prescriptions) and prescribed average daily dose was used in longitudinal k-means trajectory modelling for a 108-week time window from six months prior to pregnancy to one year after delivery. Factors associated with trajectory group membership were examined using multinomial logistic regression models.Results: We included 8,460 pregnancies in 8,092 women. Four antidepressant fill trajectories were identified based on filled antidepressant prescriptions: two distinct discontinuing patterns, one at around the start of pregnancy (30.4%) and one around the end of pregnancy (33.8%); one continuing pattern (20.6%); and one interrupting pattern (15.2%). Using average usual daily dose, we identified low dose discontinuing (60.3%), medium dose reducing (20.6%) and high dose continuing (15.2%) patterns. The multinomial logistic regressions showed that the fill trajectory group membership was strongly associated with: antidepressant type and dose prior to pregnancy and co-medication prior to pregnancy, maternal age, marital status, parity, previous pregnancy loss, and pregnancy planning.Conclusion: Longitudinal trajectory modelling revealed distinct antidepressant fill and dosage patterns in the period around pregnancy. Knowledge about factors associated with utilization trajectories might be useful for health-care personnel counselling women about antidepressant use in pregnancy.Keywords: longitudinal k-means trajectory modelling, antidepressant fill trajectories, depression, anxiety, pregnancy, drug utilization

Purpose To evaluate the use of data from population-based surveys such as the National Health and Nutrition Examination Survey (NHANES) for external adjustment for confounders imperfectly measured in health care databases in the United States. Methods Our example study used Medicaid Analytic eXtract (MAX) data to estimate the relative risk (RR) for prenatal serotonin-norepinephrine reuptake inhibitors (SNRIs) exposure and cardiac defects. Smoking and obesity are known confounders poorly captured in databases. NHANES collects information on lifestyle factors, depression, and prescription medications. External adjustment requires information on the prevalence of confounders and their association with SNRI use; which was obtained from the NHANES. It also requires estimates of their association with the outcome, which were based on the literature and allowed us to correct the RR using sensitivity analyses. Results In MAX, the RR for the association between prenatal SNRI exposure and cardiac defects was 1.51 unadjusted and 1.20 adjusted for measured confounders and restricted to women with depression. In NHANES, among women of childbearing age with depression, the prevalence of smoking was 60.2% (95% Confidence Interval 43.2, 74.3) for SNRI users and 44.1% (39.6, 48.8) for nonusers of antidepressants. The corresponding estimates for obesity were 59.2% (43.2, 74.3) and 40.5% (35.9, 45.0), respectively. If the associations between smoking and obesity with cardiac defects are independent from each other and from other measured confounders, additional adjustment for smoking and obesity would move the RR from 1.20 to around 1.10. Conclusion National surveys like NHANES are readily available sources of information on potential confounders and they can be used to assess and improve the validity of RR estimates from observational studies missing data on known risk factors.

PURPOSE: The state-assigned Case ID number in the Medicaid Analytic eXtract (MAX) allows for potential linkage of mothers to infants. No validation of respective linkage algorithms is available. We established and validated an algorithm within MAX that links mothers to infants and to identify factors influencing successful mother-infant linkage. METHODS: We identified all mother-infant pairs in FL and TX birth certificates records (BCR) that could be linked individually to MAX records (1999–2005 for FL and 1999–2010 for TX) based on Social Security Number (gold standard pairs). Case ID linkage performance was evaluated as the proportion of gold standard mother-infant pairs that were identified by the algorithm (sensitivity) and the proportion of algorithm defined mother-infant pairs that were correctly linked. Generalized estimating equations were used to calculate the probability for successful Case ID algorithm linkage versus non-linkage using maternal and infant characteristics. RESULTS: We identified 323,160 gold standard pairs in FL BCR and MAX and 1,025,350 in TX BCR and MAX. Depending on Medicaid enrollment the algorithm sensitivity ranged from 85.51% to 87.96% in FL and 19.60% to 35.75% in TX. In both states, positive predictive value exceeded 99%, regardless of enrollment periods. Determinants for successful linkage varied across states, but suggested better results for younger mothers, minority women, and those with lower educational achievement. CONCLUSIONS: Our algorithm can correctly link liveborn infants to their mothers. The algorithm’s sensitivity in identifying pairs varied across states, but PPV was consistently high. Linkage performance was associated with certain characteristics that may affect representativeness of successfully linked pairs.

OBJECTIVE: To provide updated prevalence estimates of asthma and asthma medication use for women of childbearing age in the United States. METHODS: Using data from 11,383 women aged 18–44, including a subset of 1,245 pregnant women, enrolled in the National Health and Nutrition Examination Survey (2001–2016), we assessed the age-adjusted prevalence of self-reported diagnosed asthma. For women aged 18–44, we stratified by year, demographics, and other characteristics. Furthermore, we assessed asthma medication use among women aged 18–44 with asthma. RESULTS: After age-adjustment, 9.9% (95% confidence interval (CI) 9.2%, 10.7%) of women aged 18–44 and 10.9% (95% CI 7.2%, 14.6%) of pregnant women reported having asthma. Asthma prevalence was highest in 2015–2016 (12.0% 95% CI 9.8%, 14.3%) and lowest in 2003–2004 (8.6% 95% CI 6.4%, 10.8%). Women aged 18–44 with Medicaid or State Children’s Health Insurance Program insurance coverage (16.8% 95% CI 14.5%, 19.2%), obesity (14.4% 95% CI 12.9%, 15.8%), diabetes (18.7% 95% CI 12.1%, 25.2%), hypertension (16.6% 95% CI 14.2%, 19.0%), and current smokers (12.8% 95% CI 11.4%, 14.2%) had the highest asthma prevalence. Of women with asthma, 38.3% (95% CI 34.5%, 42.1%) reported using asthma medications in the past 30 days. CONCLUSIONS: Among women of childbearing ages, asthma burden varies across demographic and clinical characteristics and has increased in recent years.

Background Early-onset preeclampsia, traditionally defined as presenting before 34 gestational weeks, is associated with even higher risks of perinatal death, placental abruption, and stroke, than late-onset preeclampsia. Objective We estimated the degree of misclassification in a high-risk population of lupus pregnancies and a general population comparator when gestational age at delivery defined preeclampsia phenotype compared to first preeclampsia diagnosis. Methods Patients with lupus and general population comparators from Sweden with ≥1 singleton pregnancy in the Medical Birth Register with a documented ICD code for preeclampsia were included (2002-2016). We used gestational age at delivery (

BackgroundPrevious studies of prenatal acetaminophen use have not addressed what indications and maternal co-factors describe acetaminophen use.ObjectiveThe objective of this study was to describe these parameters in a well-characterised, prospective birth cohort.MethodsData were drawn from the MotherToBaby study of pregnant women enrolled from 2004 to 2018. Daily acetaminophen diaries were calculated for all exposed women with complete dose and duration information. Descriptive statistics were used to assess maternal characteristics associated with acetaminophen use. Prevalence by 2-year interval was described, and linear regression was used to test for trend. Indication of use and dose per indication were summarised.ResultsOf 2441 subjects, 1515 (62%) reported use of acetaminophen. Over the 15-year period, there was a decline in use of 2.5% for each 2-year period (test for trend=0.001) with 58% reporting acetaminophen use in 2017-2018. Among women with acetaminophen use in pregnancy (n=1515), 58% reported

PurposeTo characterize prednisone use in pregnant women with rheumatoid arthritis using individual-level heat-maps and clustering individual trajectories of prednisone dose, and to evaluate the association between prednisone dose trajectory groups and gestational length.MethodsThis study included pregnant women with rheumatoid arthritis who enrolled in the MotherToBaby Autoimmune Diseases in Pregnancy Study (2003-2014) before gestational week 20 and reported prednisone use without another oral glucocorticoid during pregnancy (n=254). Information on medication use and pregnancy outcomes was collected by telephone interview plus by medical record review. Prednisone daily dose and cumulative dose were plotted by gestational day using a heat map for each individual. K-means clustering was used to cluster individual trajectories of prednisone dose into groups. The associations between trajectory group and demographics, disease severity measured by the Health Assessment Questionnaire at enrollment, and gestational length were evaluated.ResultsWomen used prednisone 3 to 292days during pregnancy, with daily doses ranging from

Background There are limited data regarding the comparability of medication exposure information during pregnancy from maternal report and medical records, including for rheumatoid arthritis and asthma-related medications. Methods This study included pregnant women with rheumatoid arthritis (n = 216) and asthma (n = 172) enrolled in the MothertoBaby Pregnancy Studies (2009-2014). Women reported types and dates of medications used through semi-structured telephone interviews up to three times during pregnancy and once after delivery, and medical records were obtained. We calculated Cohen's kappa coefficients and 95% confidence intervals (CIs) and per cent agreement for agreement between report and records. Results For rheumatoid arthritis, prednisone was reported most frequently (53%). During pregnancy, kappa coefficients for rheumatoid arthritis medications ranged from 0.32 (95% CI 0.15, 0.50) for ibuprofen, with 84.3% agreement, to 0.90 (95% CI 0.84, 0.96) for etanercept with 95.4% agreement, and was 0.44 (95% CI 0.33, 0.55) for prednisone, with 71.3% agreement. For asthma, albuterol was reported most frequently (77.9%). During pregnancy, kappa coefficients for asthma medications ranged from 0.21 (95% CI 0.08, 0.35), with 64.5% agreement for albuterol to 0.84 (95% CI 0.71, 0.96) for budesonide/formoterol, with 96.5% agreement. Where kappas for any use during pregnancy were less than excellent (i.e. ≤0.80), medication use was more frequently captured by report than record. Conclusions Agreement was higher for medications typically used continuously than sporadically. Information on medication use from medical records alone may not be adequate when studying the impact of intermittently used medications during pregnancy on perinatal outcomes.

BACKGROUND: Pre-existing conditions are imperfectly recorded in healthcare databases. We assessed whether pre-existing neurologic conditions (epilepsy, multiple sclerosis [MS]) were differentially recorded in the presence of major obstetric outcomes (Caesarean delivery, preterm delivery, preeclampsia) in delivery records. We also evaluated the impact of differential recording on measures of frequency and association between the conditions and outcomes. METHODS: The 2011-2014 Truven Health MarketScan® Commercial Claims Dataset was used to identify deliveries. We calculated the relative recording of epilepsy and MS at delivery compared with a 270-day pre-delivery interval, both overall and by the presence of major obstetric outcomes. We estimated risk ratios of the association between epilepsy and MS with the outcomes for each ascertainment window. RESULTS: We identified 909,065 deliveries in women continuously enrolled from 270-days before the delivery date. Of women with epilepsy identified in the pre-delivery interval, 73% had the condition coded at delivery. For MS, the proportion was 60%. MS recording at delivery did not vary by obstetric outcomes, however delivery-coded epilepsy was less likely confirmed in the pre-delivery interval in the presence of preeclampsia. Generally, the period of ascertainment did not meaningfully impact risk ratios, however the risk ratio for preeclampsia associated with epilepsy was 1.67 [95% CI: 1.47, 1.90] when epilepsy was ascertained at delivery and 1.26 [95% CI: 1.07, 1.48] when epilepsy was ascertained in the pre-delivery interval (heterogeneity, p=0.007). CONCLUSIONS: Ascertainment of epilepsy and MS in delivery records underestimated prevalence. However the window of recording generally did not impact associations with obstetric outcomes.

Background A "Table Fallacy," as coined by Westreich and Greenland, reports multiple adjusted effect estimates from a single model. This practice, which remains common in published literature, can be problematic when different types of effect estimates are presented together in a single table. The purpose of this paper is to quantitatively illustrate this potential for misinterpretation with an example estimating the effects of preeclampsia on preterm birth. Methods We analysed a retrospective population-based cohort of 2 963 888 singleton births in California between 2007 and 2012. We performed a modified Poisson regression to calculate the total effect of preeclampsia on the risk of PTB, adjusting for previous preterm birth. pregnancy alcohol abuse, maternal education, and maternal socio-demographic factors (Model 1). In subsequent models, we report the total effects of previous preterm birth, alcohol abuse, and education on the risk of PTB, comparing and contrasting the controlled direct effects, total effects, and confounded effect estimates, resulting from Model 1. Results The effect estimate for previous preterm birth (a controlled direct effect in Model 1) increased 10% when estimated as a total effect. The risk ratio for alcohol abuse, biased due to an uncontrolled confounder in Model 1, was reduced by 23% when adjusted for drug abuse. The risk ratio for maternal education, solely a predictor of the outcome, was essentially unchanged. Conclusions Reporting multiple effect estimates from a single model may lead to misinterpretation and lack of reproducibility. This example highlights the need for careful consideration of the types of effects estimated in statistical models.

Purpose The decision whether to continue antidepressant use for depression during pregnancy requires weighing maternal and child risks and benefits. Little is known about the effectiveness of antidepressant therapy during pregnancy. The goal of this study is to evaluate whether standard administrative claims data can be used to evaluate the effectiveness of antidepressants. Methods Using prescription and healthcare visit Medicaid claims (2000–2007), we identified 28 493 women with a depression diagnosis and antidepressant fill in the 90 days before their last menstrual period. Antidepressant continuation was defined based on prescription fills during the first trimester. Depression hospitalizations and deliberate self-harm served as measures of the effectiveness of treatment continuation during pregnancy. Propensity score and instrumental variable analyses were used to attempt to account for confounding. Results Relative to women who discontinued antidepressant therapy, women who continued were more likely to have a depression inpatient stay (odds ratio [OR] = 2.2, 95% confidence interval [95%CI]: 2.0–2.4) and deliberate self-harm code (OR = 1.4, 95%CI: 0.7–2.7). Accounting for measured covariates in the propensity score analysis, including age, race, comorbidities, comedications, features of the depression diagnosis, and antidepressant class, led to slightly attenuated estimates (OR = 2.0, 95%CI: 1.8–2.2; OR = 1.1, 95%CI: 0.5–2.4). Similar associations were estimated in subgroups with different levels of baseline depression severity. Proposed preference-time, calendar-time-based, and geography-based instruments were unlikely to meet the required conditions for a valid analysis. Conclusions Our findings suggest that either antidepressant medications do not reduce the risk of depression relapse in pregnant women, or that administrative data alone could not be used to validly estimate the effectiveness of psychotropic medications during pregnancy. Copyright © 2015 John Wiley & Sons, Ltd.

Purpose To describe methods reported in the literature to estimate the beginning or duration of pregnancy in automated health care data, and to present results of validation exercises where available. Methods Papers reporting methods for determining the beginning or duration of pregnancy were identified based on Pubmed searches, by consulting investigators with expertise in the field and by reviewing conference abstracts and reference lists of relevant papers. From each paper or abstract, we extracted information to characterize the study population, data sources, and estimation algorithm. We then grouped these studies into categories reflecting their general methodological approach. Results Methods were classified into 5 categories: (i) methods that assign a uniform duration for all pregnancies, (ii) methods that assign pregnancy duration based on preterm-delivery or health care related codes, or codes for other pregnancy outcomes, (iii) methods based on the timing of prenatal care, (iv) methods based on birth weight, and (v) methods that combine elements from 2 and 3. Validation studies evaluating these methods used varied approaches, with results generally reporting on the mistiming of the start of pregnancy, incorrect estimation of the duration of pregnancy, or misclassification of drug exposure during pregnancy or early pregnancy. Conclusions In the absence of accurate information on the beginning or duration of pregnancy, several methods of varying complexity are available to estimate them. Validation studies have been performed for many of them and can serve as a guide for method selection for a particular study. Copyright © 2015 John Wiley & Sons, Ltd.

Purpose The aim of this study is to assess the validity of preeclampsia, congenital cardiac malformations, and persistent pulmonary hypertension of the newborn (PPHN) diagnoses in the US Medicaid Analytic eXtract (MAX), a nationwide healthcare utilization database that may be useful for perinatal research.

Background Systemic lupus erythematosus (SLE) is a chronic systemic autoimmune disease that occurs during childbearing years and has been associated with preeclampsia. However, little is known about preeclampsia of early onset, which is associated with severe adverse maternal and perinatal outcomes. Methods Using national population-based Swedish registers we identified women with SLE (≥2 visits with corresponding ICD codes) and a sample without SLE who gave birth to singleton infants 2001–12. Risk ratios (RR) and 95% confidence intervals (CI) for early-onset preeclampsia (defined by ICD codes corresponding to preeclampsia registered at

Background Covariate selection to reduce bias in observational data analysis has primarily relied upon statistical criteria to guide researchers. This approach may lead researchers to condition on variables that ultimately increase bias in the effect estimates. The use of directed acyclic graphs (DAGs) aids researchers in constructing thoughtful models based on hypothesised biologic mechanisms to produce the least biased effect estimates possible. Methods After providing an overview of different relations in DAGs and the prevailing mechanisms by which conditioning on variables increases or reduces bias in a model, we illustrate examples of DAGs for maternal antidepressants in pregnancy and four separate perinatal outcomes. Results By comparing and contrasting the diagrams for maternal antidepressant use in pregnancy and spontaneous abortion, major malformations, preterm birth, and postnatal growth, we illustrate the different conditioning sets required for each model. Moreover, we illustrate why it is not appropriate to condition on the same set of covariates for the same exposure and different perinatal outcomes. We further discuss potential selection biases, overadjustment of mediators on the causal path, and sufficient sets of conditioning variables. Conclusion In our efforts to construct parsimonious models that minimise confounding and selection biases, we must rely upon our scientific knowledge of the causal mechanism. By structuring data collection and analysis around hypothesised DAGs, we ultimately aim to validly estimate the causal effect of interest.

Hypertensive disorders occur in approximately 6% to 8% of all pregnancies and are a significant source of maternal and fetal morbidity. Little is known about the range of agents routinely used in practice. We used Medicaid claims from 2000 to 2007 to identify completed pregnancies. We included women who were Medicaid beneficiaries from at least 3 months prior to last menstrual period to 1 month postdelivery, and were successfully linked to infant records. Maternal exposure to antihypertensive medications was derived from Medicaid pharmacy claim files, and duration of exposure was assigned based on the days’ supply dispensed. We identified 1 106 757 Medicaid patients in our cohort, of whom 48 453 (4.4%) were exposed to antihypertensive medications during pregnancy. The prevalence of antihypertensive use increased from 3.5% to 4.9% during the study period. Antihypertensive medication users were older than nonusers, more likely to be white or black, and more likely to have comorbid diabetes mellitus and renal disease. Overall, 1.9% of pregnant women were exposed during the first trimester, 1.7% during the second trimester, and 3.2% during the third trimester. The range of antihypertensive medications to which patients were exposed was highly heterogeneous and frequently included agents other than methyldopa or labetalol. Angiotensin-converting enzyme inhibitor exposure, which is contraindicated in late pregnancy, occurred in 928 (4.9%) antihypertensive medication users in the second trimester and 383 (1.1%) in the third trimester. Antihypertensive use during pregnancy is relatively common and increasing. The wide range of agents used during pregnancy includes medications considered contraindicated during pregnancy.

A previous study suggested an increased risk of preeclampsia among women treated with selective serotonin reuptake inhibitors (SSRIs). Using population-based health-care utilization databases from British Columbia (1997–2006), the authors conducted a study of 69,448 pregnancies in women with depression. They compared risk of preeclampsia in women using SSRIs, serotonin-norepinephrine reuptake inhibitors (SNRIs), or tricyclic antidepressants (TCAs) between gestational weeks 10 and 20 with risk in depressed women not using antidepressants. Among prepregnancy antidepressant users, the authors compared the risk in women who continued antidepressants between gestational weeks 10 and 24 with the risk in those who discontinued. Relative risks and 95% confidence intervals were estimated. The risk of preeclampsia in depressed women not treated with antidepressants (2.4%) was similar to that in women without depression (2.3%). Compared with women with untreated depression, women treated with SSRI, SNRI, and TCA monotherapy had adjusted relative risks of 1.22 (95% confidence interval (CI): 0.97, 1.54), 1.95 (95% CI: 1.25, 3.03), and 3.23 (95% CI: 1.87, 5.59), respectively. Within prepregnancy antidepressant users, the relative risk for preeclampsia among continuers compared with discontinuers was 1.32 (95% CI: 0.95, 1.84) for SSRI, 3.43 (95% CI: 1.77, 6.65) for SNRI, and 3.26 (95% CI: 1.04, 10.24) for TCA monotherapy. Study results suggest that women who use antidepressants during pregnancy, especially SNRIs and TCAs, have an elevated risk of preeclampsia. These associations may reflect drug effects or more severe depression.

Between 7% and 13% of pregnant women in the United States use antidepressants.1-2 As the number of pregnant women taking antidepressants has increased over the last 15-20 years, so has the number of publications on the safety of antidepressant use during pregnancy. Although the evidence is inconsistent, antidepressant use has been associated with a wide variety of adverse outcomes for both infant and mother, including spontaneous abortion, specific congenital malformations, persistent pulmonary hypertension of the newborn, low birth weight, poor neonatal adaptation, and preeclampsia.3 Previous studies have also reported an association between antidepressant use and preterm delivery, but the study by Yonkers and colleagues4 in this issue of EPIDEMIOLOGY stands out because the authors address potential confounding by depression, depression severity, and psychiatric co-morbidities. Their study benefits from carefully collected clinical information on psychiatric illness history and severity. The reference group consisted of women who did not have depressive illness and did not use SRI (serotonin reuptake inhibitor) antidepressants during pregnancy. A comparison of women who used SRIs but did not have a major depressive episode with the reference group suggests that SRIs are associated with preterm delivery ([odds ratio] 1.5 [95% confidence interval (CI) = 0.9-2.4]). However, these results may reflect residual confounding because it is unclear whether the authors controlled for the SRI indications of women who used SRIs but did not have depression. Further, their results suggest that major depressive disorder is not associated with preterm delivery, based on the comparison between women with a major depressive episode but no SRI treatment and the reference group, although the estimates are imprecise (95% CI = 0.4 - 1.7). Even so, the association for women who had both a major depressive episode and SRI treatment was attenuated after adjustment for age of depression onset, number of prior depressive episodes, number of hospitalizations for depression, post-traumatic stress disorder, generalized anxiety disorder, panic disorder and suicidal ideation (from a partially adjusted OR of 2.1 to a fully adjusted OR of 1.5), suggesting that depression severity and psychiatric comorbidities may be associated with preterm delivery. Preterm delivery is a syndrome with many etiologies.5 By focusing on preterm delivery overall, an association between antidepressant use and preterm delivery could be missed if antidepressants affect only certain subtypes of the outcome. By understanding the type of preterm delivery that is affected, we may gain insights into the mechanism by which antidepressants are associated with prematurity. Yonkers et al. ask the appropriate question: which type of preterm delivery is affected by SRI use during pregnancy? However, the number of outcomes in their study was too small to produce stable estimates for subtypes of preterm delivery. The results from Yonkers et al,4 together with previously reported associations, raise a number of questions. Why do women who use antidepressants have an increased risk of such a wide range of perinatal outcomes? Does depression play a an etiologic role in the development of these outcomes? Do maternal behaviors that are caused by depression or anxiety increase the risk, or the diagnosis, of the outcomes? Do the associations arise from genetic or environmental factors that cause both psychiatric illness and adverse pregnancy outcomes? Or, are antidepressants so toxic that they actually cause the whole spectrum of adverse outcomes? There are challenges when studying antidepressant safety during pregnancy. Outcomes of interest are often very rare; specific malformations typically occur in as few as 1-30 per 10,000 live births.6 Because different antidepressants may have different adverse effects, it is critical to consider antidepressant classes, and even specific drugs; obtaining enough exposed pregnancies may become difficult. Moreover, findings suggest that the potential increases in risks associated with antidepressants are small to moderate. Consequently, the first challenge in studying rare outcomes, rare exposures, and modest associations is to find large data sources. However, large data sources typically lack detailed clinical information on antidepressant indication and indication severity. This limitation brings us to the second challenge, probably the Achilles’ heel of this matter: confounding. Confounding by depression or depression severity is a major concern because it could bias associations away from the null when the reference group contains women without depression or with less severe depression. This is assuming of course that depression or factors associated with it (e.g. use of other psychotropic medications), increases the risk of congenital malformations, persistent pulmonary hypertension, preeclampsia, etc. Although this belief is strong among some investigators, the evidence to support the independent association of depression with these outcomes is weak. To reduce the potential for confounding by indication, investigators have compared (1) women with depression who are treated with antidepressants during pregnancy with women with depression but no antidepressant treatment during pregnancy; (2) women who continue treatment late in pregnancy (the etiologically relevant period for certain outcomes) with those who discontinue treatment by the end of the first trimester; and (3) women with different types of antidepressant treatment, in a head-to-head, comparative-safety analysis.7 Although these analyses may seem to address confounding by indication, they cannot resolve confounding by indication severity because treated depression, antidepressant continuation and non-SRI antidepressants are likely associated with more severe depression. Other biases, such as misclassification of the exposure (e.g., discontinuation of antidepressant prescriptions) and outcome (e.g., non-validated health insurance claims) are also likely in many published studies but would tend to bias results toward the null. Interestingly, although false negatives are worrisome in safety studies, the discussion in this particular literature has focused primarily on finding non-causal, alternative explanations for the positive findings. Another challenge in this debate has been strong a priori beliefs either for or against antidepressant safety during pregnancy. Results from non-randomized studies, no matter how large or how valid, may be unable to shift such prior convictions one way or the other. For example, a 1.5-fold increase in cardiac malformations associated with first-trimester use of an SRI was interpreted as evidence of risk in one meta-analysis8; while a 1.5-fold increase for the same association in a recent study has been interpreted as evidence of safety because the confidence interval included one.9 Considering the wide range of studies on antidepressant safety, it is surprising that very few studies exist on the effectiveness of antidepressants during pregnancy. Cohen et al.10 found that women with a history of depression who discontinued antidepressant treatment had a higher risk of relapse of major depression during pregnancy compared with pregnant women who continued treatment. However, Yonkers et al.11 found no association between antidepressant use and risk of a major depressive episode among pregnant women with a history of depression. Studies are needed on the effectiveness of antidepressant continuation during pregnancy and withdrawal symptoms among discontinuers. Moreover, these studies should consider the broad range of indications for antidepressant prescription during pregnancy, including anxiety, smoking cessation, insomnia, and chronic pain. In summary, we have learned from Yonkers et al. that, when assessing the effect of antidepressants on preterm delivery, studies should (1) control not only for the presence of depression but also for its severity; (2) consider specific types of prematurity; and (3) be sufficiently large. To justify the potential risks of antidepressant treatment during pregnancy or to justify spending more resources to define boundaries around safety, we must also demonstrate the benefits of antidepressant continuation during pregnancy. This poses a new question: can non-randomized studies convincingly assess the effectiveness of antidepressants during pregnancy?

Preeclampsia can seriously compromise maternal and offspring health.1 It causes intrauterine growth restriction and is a major cause of medically indicated preterm delivery.1–2 Current evidence suggests an association between antidepressant use during pregnancy and preeclampsia,3–6 although it is unclear if pharmacotherapy affects the risk of preeclampsia independently of mood disorders.7–9 Previous studies of selective serotonin reuptake inhibitors (SSRI), the most commonly used antidepressants during pregnancy,10 and risk for preeclampsia have reported varying degrees of association. The first study reported a 3.2-fold increase in risk of preeclampsia among SSRI users (95% confidence interval 1.9–5.3),3 whereas in two subsequent studies, the increases in risks were more moderate (1.2 to 1.6-fold).5–6 The evidence is more limited for non-SSRI antidepressants, although serotonin-norepinephrine reuptake inhibitor (SNRI) and tricyclic antidepressants were associated with preeclampsia in one study.5 Findings from these studies were challenged on the basis of potential confounding by indication, insufficient size to provide precise estimates, assess non-SSRIs, or conduct subgroup analyses, and whether the results could be replicated in other populations. Using healthcare utilization data from the Medicaid Analytic eXtract (MAX), we investigated the association between specific antidepressants used during mid-pregnancy and preeclampsia. To reduce the potential for confounding by underlying mood disorders, only women with depression diagnoses were included in the study population. The large cohort of over 100,000 pregnancies produced stable estimates and permitted us to conduct novel analyses, such as estimating the comparative safety of specific antidepressants during pregnancy, and stratifying analyses within subgroups defined by age and race. Our evaluation of antidepressants and preeclampsia was conducted within a racially diverse and indigent population typically neglected in volunteer-based studies.

BACKGROUNDIn the absence of clinical trial data, large post-marketing observational studies are essential to evaluate the safety and effectiveness of medications during pregnancy. We identified a cohort of pregnancies ending in live birth within the 2000-2007 Medicaid Analytic eXtract (MAX). Herein, we provide a blueprint to guide investigators who wish to create similar cohorts from healthcare utilization data and we describe the limitations in detail.METHODSAmong females ages 12-55, we identified pregnancies using delivery-related codes from healthcare utilization claims. We linked women with pregnancies to their offspring by state, Medicaid Case Number (family identifier) and delivery/birth dates. Then we removed inaccurate linkages and duplicate records and implemented cohort eligibility criteria (i.e., continuous and appropriate enrollment type, no private insurance, no restricted benefits) for claim information completeness.RESULTSFrom 13,460,273 deliveries and 22,408,810 child observations, 6,107,572 pregnancies ending in live birth were available after linkage, cleaning, and removal of duplicate records. The percentage of linked deliveries varied greatly by state, from 0 to 96%. The cohort size was reduced to 1,248,875 pregnancies after requiring maternal eligibility criteria throughout pregnancy and to 1,173,280 pregnancies after further applying infant eligibility criteria. Ninety-one percent of women were dispensed at least one medication during pregnancy.CONCLUSIONSMother-infant linkage is feasible and yields a large pregnancy cohort, although the size decreases with increasing eligibility requirements. MAX is a useful resource for studying medications in pregnancy and a spectrum of maternal and infant outcomes within the indigent population of women and their infants enrolled in Medicaid. It may also be used to study maternal characteristics, the impact of Medicaid policy, and healthcare utilization during pregnancy. However, careful attention to the limitations of these data is necessary to reduce biases.