Assessment of the Risk Evaluation and Mitigation Strategy (REMS) for Phentermine-Topiramate to Prevent Exposure During Pregnancy

Ann Intern Med. 2023 Apr;176(4):443-454. doi: 10.7326/M22-1743. Epub 2023 Mar 21.


Background: The U.S. Food and Drug Administration approved phentermine-topiramate for obesity in 2012 and required a Risk Evaluation and Mitigation Strategy (REMS) to prevent prenatal exposure. No such requirement was introduced for topiramate.

Objective: To evaluate the rate of prenatal exposure, contraceptive use, and pregnancy testing among patients with phentermine-topiramate compared with topiramate or other antiobesity medications (AOMs).

Design: Retrospective cohort study.

Setting: Nationwide health insurance claims database.

Participants: Females aged 12 to 55 years with no infertility diagnosis or sterilization procedure. Patients with other indications for topiramate were excluded to identify a cohort that was likely treated for obesity.

Measurements: Patients initiated use of phentermine-topiramate, topiramate, or an AOM (liraglutide, lorcaserin, or bupropion-naltrexone). Pregnancy at treatment initiation, conception during treatment, contraceptive use, and pregnancy testing outcomes were ascertained. Measurable confounders were adjusted for, and extensive sensitivity analyses were done.

Results: A total of 156 280 treatment episodes were observed. Adjusted prevalence of pregnancy at treatment initiation was 0.9 versus 1.6 per 1000 episodes (prevalence ratio, 0.54 [95% CI, 0.31 to 0.95]) for phentermine-topiramate versus topiramate. The incidence rate of conception during treatment was 9.1 versus 15.0 per 1000 person-years (rate ratio, 0.61 [CI, 0.40 to 0.91]) for phentermine-topiramate versus topiramate. Both outcomes were similarly lower for phentermine-topiramate compared with AOM. Prenatal exposure was marginally lower in topiramate users compared with AOM users. Approximately 20% of patients in all cohorts had at least 50% of treatment days covered by contraceptives. Few patients had pregnancy tests before treatment (≤5%), but this was more common among phentermine-topiramate users.

Limitations: Outcome misclassification; unmeasured confounding due to lack of prescriber data to account for possible clustering and spillover effects.

Conclusion: Prenatal exposure seemed to be significantly lower among phentermine-topiramate users under the REMS. Pregnancy testing and contraceptive use appeared to be inadequate for all groups, which deserves attention to prevent the remaining potential exposures.

Primary funding source: None.

MeSH terms

  • Anti-Obesity Agents* / adverse effects
  • Contraceptive Agents / therapeutic use
  • Female
  • Fructose / adverse effects
  • Humans
  • Obesity / chemically induced
  • Phentermine / adverse effects
  • Pregnancy
  • Prenatal Exposure Delayed Effects*
  • Retrospective Studies
  • Risk Evaluation and Mitigation
  • Topiramate / therapeutic use
  • Weight Loss


  • Topiramate
  • Phentermine
  • Anti-Obesity Agents
  • Contraceptive Agents
  • Fructose