Combined Oral Contraceptive Use and Binge Eating

Importance: Previous studies suggest endogenous ovarian hormones significantly increase binge-eating (BE) risk in females. Approximately 85% of women use combined oral contraceptives (COCs) that mimic the riskiest hormonal milieu for BE (ie, elevated estradiol and progesterone postovulation). The effects of COCs on BE risk remain unknown. Objective: To examine the associations of COCs with BE. Design, Setting, and Participants: This population-based longitudinal survey study collected daily reports of COC active vs inactive pill use and BE across 49 consecutive days in women from the Michigan State University Twin Registry. Analyses examined within-person changes in a continuous measure of BE (ie, emotional eating [EE]) when women were using active hormone pills vs inactive pills. Data were collected from 2017 to 2024. Participants were women already using monophasic COCs. Analyses examined the full sample as well as women with clinically defined BE episodes. Data were analyzed from April 2024 to November 2025. Exposure: COC pill type (active vs inactive pills). Main Outcomes and Measures: The outcome of interest was within-person changes in EE between inactive vs active hormone pills, controlling for negative affect. Changes across 2 pill packs were examined for replication. Analyses also examined weight preoccupation (WP) as a control outcome, given its lack of past associations with ovarian hormones. Primary models focused on the full sample; sensitivity analyses examined women with clinically defined BE. Results: A total of 422 women (mean [SD] age, 21.95 [3.10] years) were included in the full sample. Significant within-person increases in EE were observed in the full sample during active hormone vs inactive pills in both cycles (cycle 1: β = 0.11 [95% CI, 0.06 to 0.16]; cycle 2: β = 0.07 [95% CI, 0.04 to 0.10]). Increases were not mediated by changes in negative affect and were observed in the subsample of 51 women (mean [SD] age, 22.44 [3.57] years) with clinically defined BE episodes (cycle 1: β = 0.13 [95% CI, -0.07 to 0.33]; cycle 2: β = 0.12 [95% CI, 0.02 to 0.23]. Importantly, no significant changes in WP were observed across pill type, and post hoc analyses of negative affect as the outcome showed more modest COC outcomes. Conclusions and Relevance: This intensive, daily survey study of COC use found a specific association of active COC pills with risk for EE. Future studies are needed to identify for whom COCs are most risky to inform personalized medicine and identify contraceptive options that may be less likely to impact BE or EE.