Work

Associations between self-reported personal care products use and menstrual cycle length and regularity in a US digital cohort

Year: 2026

Type: article

Abstract: BACKGROUND: Personal care products (PCPs) may contain endocrine-disrupting chemicals (EDCs) that can impact menstrual health. Despite widespread usage, little is known about the associations of PCP usage, EDC avoidance, and menstrual cycle characteristics. METHODS: This analysis included female participants from a US-based digital cohort who enrolled between 11/2019-04/2025, provided consent, aged < 50 years, completed PCP questions (past year usage frequency, age at initiation, and ingredient avoidance when purchasing PCPs), and either reported their usual menstrual cycle length (<21, 21-39, ≥40 days/too irregular) (n = 4,155) or shared logged cycle data (n = 4,482 with 61,079 cycles). We used confounder-adjusted multinomial logistic regression or linear mixed effect models to examine associations between each exposure and cycle length or variability. RESULTS: PCP usage frequency varied considerably by product (e.g., 54%, 8%, 2% using body fragrance, dry shampoo, and self-tanner ≥ 3 times/week; 23% and 2% using permanent and semi-permanent hair dye ≥ 4 times/year and ≥ monthly). We found consistent associations between both frequent current usage and childhood initiation of the aforementioned PCPs with very long, very short, or irregular cycles. Participants avoiding harmful ingredients (petroleum propellants, formaldehyde, lead, mercury, or talc) were more likely to report regular cycles (lower cycle variability). Effect modifications were found, and individuals who self-reported fibroids or endometriosis diagnoses or of younger age (18-29 years) were more susceptible to some of these associations. DISCUSSION: These findings suggest that EDC-associated PCPs may impact reproductive health, thus suggesting PCP use as an add to modifiable environmental factors affecting menstrual health worthy of future research.

Source: Environment International

DOI: 10.1016/j.envint.2026.110260