Imputation of missing data did not materially change the estimated associations (eTable). Further adjustment for maternal age and firstborn status did not change any of the associations for other early-life exposures, and mutually accounting for both of these exposures in the same model did not affect their associations with age at menarche (data not shown). The association between pregnancy-related hypertensive disorder and early menarche was unchanged by adjustment for multiple birth and maternal diabetes (data not shown). After further adjustment for multiple birth, maternal age, and maternal factors during the index pregnancy (DES, smoking, diabetes, and hypertensive disorder), the association between preterm birth and very early menarche was attenuated (1.10 [0.82-1.47]), although some association with very late menarche remained (1.20 [0.94-1.54]). Adjustment for preterm birth, multiple birth, and maternal factors during the index pregnancy did not affect the association between low birth weight and very early menarche (1.33 [1.08-1.63]). Adjustment for factors indicative of higher-risk pregnancy (multiple birth, maternal diabetes, and maternal hypertensive disorder) did not affect the DES association with very early menarche (1.49 [1.20-1.86]).
