OBJECTIVE To assess pregnancy outcomes at 40, 41, and 42 weeks' gestation when labor induction is done routinely at 42 but not 41 weeks. METHODS We reviewed all singleton pregnancies delivered at 40 or more weeks' gestation between 1988 and 1998 at Parkland Memorial Hospital, Dallas, Texas. We excluded women with hypertension, prior cesarean, diabetes, malformations, breech presentation, and placenta previa. Labor characteristics and neonatal outcomes of pregnancies at 41 and 42 weeks' gestation were compared with pregnancies that ended at 40 weeks. Women with certain dating criteria had induction of labor at 42 weeks. Gestational age was calculated from the last menstrual period (LMP), sonography when available, and clinical examination. If the fundal height between 18 and 30 weeks was within 2 cm of gestational age, the reported LMP was accepted as correct. Sonogram was used to calculate gestational age if a discrepancy was identified. Statistical analysis consisted of chi(2) and analysis of variance. RESULTS We studied 56,317 pregnancies: 29,136 at 40 weeks, 16,386 at 41 weeks, and 10,795 at 42 weeks. Labor complications increased from 40 to 42 weeks, including oxytocin induction (2% versus 35%, P <.001), length of labor (5.5 +/- 4.9 versus 8.8 +/- 6. 5 hours, P <.001), prolonged second stage of labor (2% versus 4%, P <.001), forceps use (6% versus 9%, P <.001), and cesarean delivery (7% versus 14%, P <.001). Neonatal outcomes were similar in the three groups, including 5-minute Apgar score less than 4, admission to the neonatal intensive care unit (NICU), umbilical artery pH less than 7, seizures, and perinatal mortality. Sepsis was more frequent in the 42-week group than the other groups (0.1 versus 0.3%, P =. 001), as was admission to the NICU (0.4 versus 0.6%, P =.008). CONCLUSIONS Routine labor induction at 41 weeks likely increases labor complications and operative delivery without significantly improving neonatal outcomes.