Birth length and perinatal exposures and rheumatoid arthritis risk in the E3N cohort

OBJECTIVE: To examine associations between participants' perinatal factors, including harmonised birth weight and length, prematurity, exposure to maternal smoking during pregnancy and breastfeeding during infancy, and incident rheumatoid arthritis (RA) in a large prospective cohort of French women. METHODS: The Etude Epidémiologique auprès des femmes de la Mutuelle générale de l'Éducation Nationale (E3N) cohort enrolled 98 995 women in 1990. Perinatal data were collected in 2002. Birth weight and length were reported continuously and/or categorically and harmonised according to term status. RA cases were validated using a specific questionnaire. Adjusted HRs (aHRs) and 95% CIs were estimated using Cox proportional hazards models informed by a directed acyclic graph. RESULTS: Among 78 473 women, 698 incident RA cases were identified. No evidence of an association was observed between perinatal factors and RA risk in the overall population. Point estimates for large birth length (aHR 1.32, 95% CI 0.97 to 1.78) and preterm birth (aHR 1.34, 95% CI 0.92 to 1.94) were above 1, but CIs included the null value of 1.In sensitivity analyses restricted to women with continuous birth data, a signal for large birth length was observed (aHR 1.98, 95% CI 1.23 to 3.19), whereas no such signal was detected when using categorised self-reported measures. CONCLUSIONS: In this large prospective cohort, most perinatal factors were not associated with RA risk in adulthood. A signal related to larger birth length emerged only in selected analyses based on more precise exposure assessment and should be interpreted cautiously. These findings are hypothesis-generating and warrant replication in independent cohorts using prospectively measured perinatal data. TRIAL REGISTRATION NUMBER: .