Background
Amniotomy is a commonly used, inexpensive method of labour induction; however, the optimal timing of amniotomy during labour induction is debated.

Aims
To investigate whether artificial rupture of membranes <4 cm dilation is associated with caesarean, severe maternal and neonatal morbidity, and labour induction duration. Materials and Methods Retrospective cohort study of 228 438 deliveries at 19 US hospitals. Women with a viable, singleton gestation undergoing induction ≥37 weeks with cervical dilation <4 cm were included. Women were excluded if membranes spontaneously ruptured <4 cm. Women were compared by early amniotomy (<4 cm dilation) versus not early. The primary outcome was caesarean. Secondary outcomes included severe maternal and neonatal morbidity, and labour duration. Logistic and Cox proportional hazard regression estimated the association between early amniotomy and study outcomes. Results Of 15 525 eligible women, 10 421 (67%) had early amniotomy. Early amniotomy was associated with higher adjusted odds of caesarean and severe maternal morbidity, but not neonatal morbidity. After accounting for interaction, early amniotomy was associated with increasingly higher odds of caesarean as body mass index increased. Early amniotomy was associated with lower odds of severe maternal morbidity among multiparas with mechanical ripening. Median labour induction was 2.5 h shorter with early amniotomy, significant in hazard regression. Conclusions Early amniotomy was associated with increased odds of caesarean among obese women. The association between early amniotomy and severe maternal morbidity varied by maternal characteristics, but early amniotomy was not associated with neonatal morbidity. Early amniotomy in labour induction may be advantageous in certain populations, particularly non‐obese women requiring mechanical ripening.