Evaluating the Safety of Antiemetic Combination Use on Pregnancy Loss

Background: Antiemetic combination therapy is recommended to treat severe nausea and vomiting during pregnancy (NVP). Although several pharmacokinetic and -dynamic interactions are plausible (e.g., stronger synergistic effects on QT prolongation for promethazine-ondansetron combination (POC)), no prior study has assessed the safety of such combinations.

Objectives: To evaluate the risk for pregnancy loss (spontaneous abortion or stillbirth) following prenatal exposure to POC versus metoclopramide-ondansetron combination (MOC).

Methods: This was a retrospective cohort study using the MarketScan® Commercial Claims 2005-2019 and Medicaid Analytic eXtract data 2005-2015. We excluded women with ectopic pregnancies, induced abortions or unknown pregnancy outcomes, malignant neoplasms, using ≥2 of any antiemetic combinations (e.g., POC and MOC; POC and promethazine-metoclopramide combination), or with diagnoses for congenital anomalies or exposure to a known teratogenic drug. To be considered a concomitant user, women had to fill a prescription for a 2nd antiemetic during an active days' supply of the 1st antiemetic (index date) and then refill the 1st antiemetic during the active days' supply of the 2nd antiemetic (as index date for gestational age adjustment) during the first trimester. We adjusted for confounders via stabilized inverse probability treatment weighting. We used a random-effect model to combine the adjusted absolute risk difference (aARD) and relative risk (aRR) from the two databases. In sensitivity analyses, we restricted to patients with severe NVP to reduce confounding by severity, to those without any hospitalization before the index date, and expanded the exposure window to the first 20 weeks.

Results: The proportion of pregnancy loss among POC and MOC users was 3.5% (164/4622) and 2.2% (56/2571) in the MarketScan and 4.8% (127/2644) and 3.6% (64/1798) in the Medicaid population, respectively. After adjustment, the pooled aARD was 7 per 1000 POC users (95%CI 0, 14) and aRR was 1.21 (95%CI 0.97, 1.52). When restricting to patients with severe NVP, we observed a slight increase in risk estimates (pooled aARD=10, 95%CI 1, 20 and aRR= 1.36, 95%CI 1.00, 1.86). Other sensitivity analyses had results consistent with the main analysis.

Conclusions: While not statistically significant and potentially affected by residual confounding, we could not exclude an excess risk of pregnancy loss attributable to POC. Our findings may assist clinical decision-making in choosing antiemetic combinations for pregnant women.