Continuation of psychiatric medications during pregnancy

Abstract Background While medications for anxiety and depression are commonly used in the United States, it is unclear to what degree they are continued during pregnancy. Methods We used a large administrative database to determine whether psychiatric medications are continued during pregnancy and predictors of continued medication treatment. Results Of 2,672,656 women included in our analysis, 86,454 (3.1%) filled a pre-pregnancy prescription for an anxiolytic or antidepressant medication within 3 months of estimated conception. Of women who filled a pre-pregnancy prescription, 49.4%, 26.1%, and 20.1% filled subsequent prescriptions in the 1st, 2nd, and 3rd trimesters. Discontinuation rates ranged by pharmaceutical agent, from 16% for fluoxetine to 71% for alprazolam. White women and women over 25 were more likely to continue anxiolytic and antidepressant treatment during pregnancy. Conclusion Because untreated and under-treated mental health conditions are linked to adverse maternal outcomes, high discontinuation rates may have important implications for maternal health.


Introduction
Antidepressants and anxiolytics are among the most prescribed medications in reproductive-aged people in the United States [1,2]. There is no consensus on optimal use of psychiatric medications in pregnancy [3,4]: some women may discontinue antidepressant and anxiolytic medications when they become pregnant due to concern for birth defects and/or poor pediatric neurodevelopmental outcomes [5][6][7]. But untreated or under-treated mental health conditions in pregnancy also carry risk for adverse maternal outcomes, including relapse of psychiatric disease and pregnancyrelated death [8][9][10]. Nearly 40% of pregnancy-related deaths attributable to mental health conditions are associated with discontinuation of or change in medication treatment for a psychiatric or substance use disorder [10]. Large population data on continuation of antidepressants and anxiolytics may be of clinical and public health importance. We used a large administrative database to determine whether psychiatric medications are continued during pregnancy as well as predictors of continued medication treatment.

Methods
We analyzed a retrospective cohort of women hospitalized for delivery in the IBM MarketScan Research Databases from 2008 through 2014. In 2017, the MarketScan dataset captured administrative claims data, including pharmaceutical claims, from more than 50 million privately-insured patients and 6 million Medicaid enrollees per year across 12 states; due to its size, MarketScan is considered a nationally-representative sample of patients with employer-sponsored insurance [11]. We identified women aged 15 to 54 included in MarketScan who underwent a delivery hospitalization during 2008-2014 and had continuous insurance coverage, including pharmacy enrollment, for at least one year prior to delivery. To identify delivery hospitalizations, we applied an algorithm of International Classification of Diseases, 9th Edition, Clinical Modification (ICD-9-CM) codes that has previously been shown to capture more than 95% of deliveries [12]. For women with >1 delivery during the study period, we included only the index delivery hospitalization and index pregnancy. We queried pharmaceutical claims for filled prescriptions for commonly-used anxiolytics and antidepressantsincluding benzodiazepines, selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs), atypical antidepressants, and other medicationsamong included women, using National Drug Codes in the MarketScan RED BOOK [11]. We identified a cohort who received an anxiolytic or antidepressant medication within 3 months prior to pregnancy, applying a validated approach to estimate conception date [11,13]. Using population-level data, estimates that assigned 35 weeks of gestational age at birth to deliveries with the preterm-status indicator and 39 weeks to those without them were within 2 weeks of the clinical gestational age at birth in 75% of preterm and 99% of term deliveries. We then determined the likelihood of subsequent medication receipt during the 1st, 2nd, and 3rd trimesters. Using adjusted log-linear regression, we evaluated the association between demographic and hospital factors and likelihood of continuing psychiatric medications during each trimester of pregnancy. Demographic factors included year of delivery, maternal age, race, and payer. Hospital factors included hospital region and location. We report adjusted risk ratios (aRRs) with 95% confidence intervals (CIs) as measures of effect. This analysis was granted an exemption from Columbia University's Institutional Review Board, as MarketScan is a limited dataset with de-identified data.

Discussion
Nearly 80% of women who filled a prescription for an anxiolytic and/or antidepressant medication in the prepregnancy period discontinued psychiatric medication by the end of pregnancy. Discontinuation occurred differentially by pharmaceutical agent, with higher discontinuation rates among commonly-prescribed benzodiazepines as compared to commonly-prescribed SSRIs or venlafaxine. Among commonly-prescribed antidepressants, discontinuation rates ranged from 16% (fluoxetine) to 59% (venlafaxine), possibly reflecting variation in risk of birth defects between individual antidepressants which occurs even among antidepressants within the same class [6]. However, risks associated with venlafaxine and further updated analyses are indicated to further analyze factors to differentials in discontinuation of specific agents. We found that White women and women aged 25 and older were more likely to continue anxiolytics and antidepressants during pregnancy. Prior studies using administrative data have shown higher rates of mental health condition diagnoses among deliveries to non-Hispanic White women [14] though non-Hispanic Black pregnant women have higher screen-positive rates for these conditions [15]. Our findings may support racial differentials in the management of mental health conditions in pregnancy. This study also found that discontinuation in the third trimester increased in later years of the study suggesting potential trends, although these associations were modest. These findings add to prior analyses of psychopharmacoepidemiology during pregnancy by evaluating use of medications and discontinuation through pregnancy by trimester [5].
There are limitations that are important in interpreting this study. First, the study inclusion criteria included a disproportion number of patients with commercial insurance and the Medicaid sample was limited. Second, we analyzed only the first three months prior to pregnancy. It is possible that many women may have discontinued these medications more than 3 months prior to becoming pregnant. Subsequent analyses could analyze consistent use in the year prior to pregnancy, discontinuation before pregnancy, and discontinuation during pregnancy. Strengths include the large sample size. Third, it is possible that our measure of discontinuation could be affected by chronicity and women using benzodiazepines may use them intermittently. Further analyses are indicated to determine whether pregnancy results in changes of patterns of benzodiazepine refills compared to periods outside pregnancy for similar reproductive age women. Fourth, data was analyzed only through 2014 and practice patterns and trends may have changed since the end of the study period.
Given associations between psychiatric medication instability, relapse of psychiatric disease and pregnancy-related death [8][9][10], high medication discontinuation rates during pregnancy may have important implications for maternal health depending on the specific agent being used and the underlying mental health condition diagnosis. While discontinuation of anxiolytics such as benzodiazepines may provide maternal and neonatal benefit in appropriate circumstances, changes in medication for other conditions may impose risks. Further research is needed to better contextualize the maternal risks associated with antidepressant and anxiolytic discontinuation with any protective effects that discontinuation might have on the risk for birth defects. Further research is also indicated to evaluate polypharmacy, clinical outcomes, management of specific mental health condition diagnoses, changes in agent, the role of comorbid factors such as substance use disorder and domestic violence, and medication initiation in pregnancy.

Disclosure statement
Dr. D'Alton has had a leadership role in ACOG II's Safe Motherhood Initiative which has received unrestricted funding from Merck for Mothers. Dr. Wright has served as a consultant for Clovis Oncology and received research funding from Merck. Timothy Wen disclosed that money was paid to him from Delfina Care, Inc. The other authors did not report any potential conflicts of interest.

Funding
The author(s) reported there is no funding associated with the work featured in this article.