Sharma · American journal of obstetrics and gynecology 2025 · retrospective survey study · n=5414

Prepregnancy metformin use associated with lower risk of severe nausea and vomiting of pregnancy and hyperemesis gravidarum.

Cited 5 times in the scientific literature.

Level 4 - case-series / case-control

Retrospective self-reported survey cohort/case-control design

PubMed 40588059 · doi:10.1016/j.ajog.2025.06.055 · record verified 2026-08-28

What was done

A retrospective online structured survey was administered to visitors of the Hyperemesis Gravidarum Education and Research Foundation social media pages between January 2023 and September 2024. A total of 5,414 participants self-reported daily use of 32 medications and substances in the month prior to pregnancy alongside the severity of nausea and vomiting during pregnancy. Logistic regression was used to evaluate multivariate associations between substance exposures and severe nausea and vomiting of pregnancy or hyperemesis gravidarum, adjusting for maternal age and tobacco use.

What was found

Prepregnancy daily metformin use was associated with a >70% lower risk of hyperemesis gravidarum in the first pregnancy (adjusted relative risk [aRR], 0.29; 95% CI, 0.12-0.71; P=.007) and an 82% lower risk of severe nausea and vomiting of pregnancy and hyperemesis gravidarum in the second pregnancy (adjusted odds ratio [aOR], 0.18; 95% CI, 0.06-0.59; P=.005). Prepregnancy selective serotonin reuptake inhibitor use was associated with increased risk in both the first (aRR, 2.41; 95% CI, 1.33-4.38; P=.004) and second pregnancy (aOR, 1.84; 95% CI, 1.12-3.04; P=.016). Tobacco use before the first pregnancy was associated with lower risk (aRR, 0.51; 95% CI, 0.30-0.86; P=.011), while cannabis use before the second pregnancy was associated with increased risk (aOR, 3.48; 95% CI, 1.80-6.75; P<.001).

Why it matters

This study provides observational human data suggesting prepregnancy metformin exposure may desensitize patients to GDF15 and lower the risk of severe pregnancy sickness, providing a rationale for future randomized trials.

Limits

The study relies on self-reported survey data from a social media convenience sample, which introduces substantial selection and recall bias. Diagnoses, drug dosages, and adherence were not verified through medical records, and confounding by underlying indications (e.g., polycystic ovary syndrome, diabetes, or depression) was not fully controlled.

Cited by