Providing Data to Empower Pregnant Patients to Make Evidence-Based Choices | Clinical Pharmacy and Pharmacology | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Views 1,462
Citations 0
Invited Commentary
Obstetrics and Gynecology
April 23, 2021

Providing Data to Empower Pregnant Patients to Make Evidence-Based Choices

Author Affiliations
  • 1Department of Obstetrics and Gynecology, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
JAMA Netw Open. 2021;4(4):e215359. doi:10.1001/jamanetworkopen.2021.5359

Medication use in pregnancy is nearly universal, and nausea and vomiting in early pregnancy (NVP) is a common indication. Approximately 30% of all pregnant individuals in the US take medication for NVP, and ondansetron, a 5-HT3 receptor antagonist, is the most commonly prescribed such medication (12% of all pregnancies).1 NVP is most severe during the first trimester of pregnancy, which overlaps with a critical period of organogenesis and placental development; thus, understanding the risks of congenital malformations and other adverse pregnancy outcomes in ondansetron-exposed pregnancies is a critically important question.

In a multicenter study, Dormuth et al2 address this question by examining the risk of spontaneous abortion, stillbirth, and major congenital malformations in association with ondansetron exposure compared with other commonly used antiemetics (doxylamine with pyridoxine, metoclopramide, or promethazine) in 163 810 pregnancies exposed to ondansetron and 306 766 pregnancies exposed to other antiemetics.2 Reassuringly, they find no significant association between ondansetron exposure and risk of fetal death (spontaneous abortion or stillbirth) or congenital malformations.

The study by Dormuth et al2 adds to the body of literature regarding ondansetron safety in pregnancy, which remains a controversial area of investigation. Although some studies have suggested a potential association of ondansetron use with congenital heart defects, more recent work provides reassurance.3 A remaining concern is a potential small increase in the risk of oral clefts in ondansetron-exposed pregnancies—a concern that the current study by Dormuth et al2 is not powered to address. These unresolved concerns have led the American College of Obstetricians and Gynecologists to recommend ondansetron as a second-line agent for NVP4 and have prompted the European Medicines Agency to release warnings about ondansetron use in the first trimester of pregnancy. Much of the debate stems from the imperfect data on which studies assessing medication risks (including ondansetron) are based. In the absence of adequate clinic trial data, one commonly used tool to evaluate the association of medication exposures with maternal, fetal, and obstetric outcomes is the use of large claims databases that link health care administrative data to patient data. The study by Dormuth et al2 combines data from 3 such databases from Canada, the US (IBM MarketScan), and the United Kingdom (UK Clinical Practice Research Datalink).

Although these data sets provide the power necessary to examine critical questions about the association of medication exposures with specific clinical outcomes, the use of administrative data for this purpose also poses many challenges. For example, to understand the association of medication exposure with pregnancy outcomes, the pregnancy episode must be accurately identified and dated, the timing of medication exposure must be defined, and the type of pregnancy outcome (ie, spontaneous abortion, stillbirth, or live birth) must be curated. For studies assessing congenital malformations as a result of medication exposure, accurate assignment of gestational age is critical because the window of exposure in relation to organogenesis is key for understanding potential adverse medication effects. In addition, unlike clinical trials where treatment compliance is carefully monitored, determination of medication use from administrative databases is imperfect because the data reflect prescriptions either written or dispensed (depending on the specific data set), not medication actually taken by individuals. Furthermore, the ability to control for confounders is dependent on important covariates being included in administrative data. Although certain covariates (eg, maternal diabetes and chronic hypertension) are generally well recorded, others (eg, obesity, substance use, and over-the-counter medication use) are limited. Given the sparse annotation of a subset of important covariates, the possibility of residual confounding should be strongly considered when evaluating study results. Taken together, these limitations suggest exercising caution when formulating clinical recommendations from studies based on administrative data.

The lack of safety data on medications commonly used in pregnancy is unfortunately common. Pregnant individuals are routinely excluded from industry-sponsored clinical trials, which means that many years must pass before enough observational data can be gathered to provide safety data to guide clinical decision-making.5 Even during the current pandemic, despite well-documented risks of severe COVID-19 in pregnancy, most clinical trials for COVID-19 (including all initial COVID-19 vaccine trials) have excluded pregnant individuals. The result of exclusion is that pregnant individuals experience delays in access to effective therapies. Even more concerning is that pregnancy dramatically changes physiology and drug pharmacokinetics; thus, data from nonpregnant individuals are not directly translatable to those who are pregnant.6 Separate trials in pregnant women may often be needed to generate sufficient relevant data and should be promoted to facilitate advances in knowledge about therapies used during gestation. Furthermore, conditions that occur only during pregnancy that remain suboptimally treated, including NVP, should be prioritized for additional research funding. Because 10% of all women in the US become pregnant in any given year,7 a sizeable proportion of the population is impacted by these knowledge gaps and lack of effective treatments.

Optimizing maternal health requires the prioritization of maternal well-being and autonomy in decision-making. During pregnancy, an individual must weigh their own maternal risks with potential fetal risks and make decisions in accordance with their values and goals.8 Increasing the available pregnancy-specific data is an essential step for facilitating evidence-based treatment recommendations. To truly empower women, we must prioritize studies that recognize health care across the life span, including the investigation of therapies during pregnancy and lactation, such that decisions can be based on evidence, not inference.

Back to top
Article Information

Published: April 23, 2021. doi:10.1001/jamanetworkopen.2021.5359

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Gray KJ. JAMA Network Open.

Corresponding Author: Kathryn J. Gray, MD, PhD, Department of Obstetrics and Gynecology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115 (

Conflict of Interest Disclosures: Dr Gray reported having served as a consultant for Illumina, BillionToOne, and Aetion outside the submitted work. No other disclosures were reported.

Additional Contributions: Krista Huybrechts, MS, PhD (Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA), provided thoughtful insights on this research area and was not compensated for this contribution.

Haas  DM, Marsh  DJ, Dang  DT,  et al.  Prescription and other medication use in pregnancy.   Obstet Gynecol. 2018;131(5):789-798. doi:10.1097/AOG.0000000000002579PubMedGoogle ScholarCrossref
Dormuth  CR, Winquist  B, Fisher  A,  et al; Canadian Network for Observational Drug Effect Studies (CNODES) Investigators.  Comparison of pregnancy outcomes of patients treated with ondansetron vs alternative antiemetic medications in a multinational, population-based cohort.   JAMA Netw Open. 2021;4(4):e215329. doi:10.1001/jamanetworkopen.2021.5329Google Scholar
Huybrechts  KF, Hernández-Díaz  S, Straub  L,  et al.  Association of maternal first-trimester ondansetron use with cardiac malformations and oral clefts in offspring.   JAMA. 2018;320(23):2429-2437. doi:10.1001/jama.2018.18307PubMedGoogle ScholarCrossref
Erick  M, Cox  JT, Mogensen  KM.  ACOG Practice Bulletin 189: nausea and vomiting of pregnancy.   Obstet Gynecol. 2018;131(5):935. doi:10.1097/AOG.0000000000002604PubMedGoogle ScholarCrossref
Briggs  GG, Polifka  JE, Wisner  KL,  et al.  Should pregnant women be included in phase IV clinical drug trials?   Am J Obstet Gynecol. 2015;213(6):810-815. doi:10.1016/j.ajog.2015.05.047PubMedGoogle ScholarCrossref
Ayad  M, Costantine  MM.  Epidemiology of medications use in pregnancy.   Semin Perinatol. 2015;39(7):508-511. doi:10.1053/j.semperi.2015.08.002PubMedGoogle ScholarCrossref
Curtin  SC, Abma  JC, Ventura  SJ, Henshaw  SK.  Pregnancy rates for U.S. women continue to drop.   NCHS Data Brief. 2013;136:1-8.PubMedGoogle Scholar
Minkoff  H, Ecker  J.  Balancing risks: making decisions for maternal treatment without data on fetal safety.   Am J Obstet Gynecol. Published online February 1, 2021. doi:10.1016/j.ajog.2021.01.025PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words