Nausea and vomiting during pregnancy (sometimes termed morning sickness) are among the most common medical issue to arise during pregnancy. Most pregnant women experience some form of nausea and vomiting, and prevalence rates are as high as 50% to 80%.¹ Conservative measures, such as dietary modifications and avoidance of triggers, often do not control symptoms, so medications and other nondrug therapies are tried.² Nausea and vomiting is one of the most common indications for prescriptions during pregnancy.³ Because this is a condition mainly of the first trimester, pregnant women and clinicians have concerns about the potential effects these therapies might have on the developing fetus.

In this issue of JAMA, Huybrechts and colleagues⁴ attempt to clarify associations of a commonly used antiemetic, ondansetron, with fetal congenital malformations. Because these malformations are typically rare, establishing cause and effect of a single drug and an anomaly is extremely difficult. Association studies using robust data sets often provide the best available evidence for establishing risk or safety of a drug. The authors describe some of the limitations of prior association studies and detail how the Medicaid Analytic eXtract (MAX) database is ideally suited to overcome some of these prior limitations. Because the authors have used these data for other pregnancy exposure studies, they are well-suited for the current analysis.

In this study, the authors used data from more than 1.8 million pregnancies, allowing for a large number of analyses and adjustments. Focusing on cardiac malformations and oral clefts (the main congenital anomalies identified with any consistency in prior studies), the authors found no significant association between ondansetron and cardiac abnormalities (14 577 abnormalities among 1 727 947 unexposed infants and 835 abnormalities among 88 467 exposed infants; adjusted risk difference, −0.8; 95% CI, −7.3 to 5.7 per 10 000 births; adjusted relative risk, 0.99, 95% CI; 0.93 to 1.06). They also detected a small but statistically significant increased risk of oral clefts with first-trimester exposure to ondansetron (1912 abnormalities among unexposed infants and 124 abnormalities among exposed infants; adjusted risk difference, 2.7; 95% CI, 0.2 to 5.2 per 10 000 births; adjusted RR, 1.24; 95% CI, 1.03 to 1.48). After multiple adjustments, the authors also found no difference in the risk of cardiac or overall congenital anomalies for infants of women exposed to ondansetron.

According to the American College of Obstetricians and Gynecologists (ACOG)–endorsed treatment algorithm, ondansetron, a serotonin receptor antagonist, is not the first-line drug for treatment of nausea and vomiting.¹ However, it is one of the most commonly used agents in practice.^3,5 This may be because of its different formulations, including an oral dissolving tablet, and its perceived effectiveness. Thus, it is important that clinicians and pregnant women have accurate safety information about its use during pregnancy.

Treatment of nausea and vomiting can be difficult and must be tailored to the individual patient. Often, a single treatment or combination of treatments that work for one woman will be ineffective for another. In the population studied by Huybrecht et al, 139 932 pregnant women (7.7%) also filled a prescription for promethazine, 52 818 (2.9%) filled a prescription for metoclopramide, and 9036 filled a prescription for pyridoxine (vitamin B₆, 0.5%)—3 commonly used agents for nausea and vomiting. According to the treatment algorithm proposed by the ACOG Practice Bulletin, pyridoxine, alone or in combination with doxylamine, is listed as the first-line agent. Ondansetron appears as a third-line option in the algorithm.¹

Ondansetron is metabolized by cytochrome P450 (CYP) enzymes that are polymorphic, such as CYP 2D6. Understanding how genetic differences among individuals cause varied responses to the same drug may have implications for prescribing.⁶ Because there are often genetic-based differences in activity of common drug-metabolizing enzymes, transporters, and receptors among individuals, pharmacogenomics can help guide therapeutic choices in several disease states.⁷ Further research is needed for ondansetron and other drugs for nausea and vomiting to determine if pharmacogenomic or specific phenotype characteristics can help personalize treatment for improved outcomes.^8,9

In the evolving era of personalized medicine, rapidly determining the most effective therapy can have tangible benefits. Women with severe nausea and vomiting have decreased work productivity, more time away from personally fulfilling activities, and may have higher rates of adverse pregnancy outcomes.^10,11 Additionally, severe nausea and vomiting was cited as a reason for higher rates of pregnancy termination from the Motherisk project in Canada.¹² Thus, early and effective safe treatment of nausea and vomiting is crucial.

The article by Huybrechts and colleagues provides important and helpful information for physicians and other clinicians to use when counseling women about the safety of treatment options. The safety of some other options, such as pyridoxine, has been established.¹³ The current article documents clearly that while the adjusted relative risk of oral clefts was elevated, the absolute risk increase is very low. The multiple adjustments and comparison groups presented, including accounting for the potential baseline risk of nausea and vomiting, demonstrate the robustness of the authors’ conclusions.

In addition, the authors acknowledge the typical limitations of these types of analyses. Because of the low frequency of congenital anomaly outcomes, randomized trials would need to have enormous sample sizes to detect differences between drug exposures. Thus, observational data are best suited for these studies. One additional relevant limitation of the data set is that prescriptions for the current recommended first-line treatment, pryidoxine (with or without doxylamine), may not have been completely captured. The prescription combination of doxylamine and pyridoxine was approved by the US Food and Drug Administration in 2013, the final year the MAX data were queried. Many women may have obtained these components over-the-counter, and thus the true polypharmacy rates for nausea and vomiting treatment are not accounted for in the data set.

Utilizing a large data set, the study by Huybrechts et al provides some reassurance for obstetricians and other clinicians, as well as for pregnant women regarding the safety of a commonly used medication for nausea and vomiting. The potential risk and safety findings must be weighed with the effectiveness of ondansetron in treating nausea and vomiting and avoiding hyperemesis gravidarum. The analysis also provides additional reassurance of no increased risk of cardiac or total congenital malformations. As clinicians and pregnant women engage in informed, shared decision-making surrounding treatment for nausea and vomiting, the current information is important for contextualizing risks in light of the potential benefits.