Forceps Compared With Vacuum: Rates of Neonatal and Maternal Morbidity

Abstract

This retrospective cohort study was carried out to compare perinatal outcomes in forceps- and vacuum-assisted deliveries. The authors tested the hypothesis that the force vectors operating in forceps deliveries will result in fewer cases of shoulder dystocia but more severe perineal lacerations. The retrospective cohort study enrolled 4120 term singleton pregnancies having nonrotational operative vaginal deliveries. The study included 2075 forceps-assisted deliveries and 2045 vacuum-assisted deliveries. Multivariate logistic regression analyses were carried out to control for potentially confounding factors such as maternal age, birth weight, ethnicity, parity, station at delivery, episiotomy, attending physician, anesthesia, and the length of labor. Shoulder dystocia was more frequent in women having vacuum-assisted delivery than in those having forceps delivery (3.5% vs 1.5%, P < 0.001). Low 5-minute Apgar scores, cephalohematoma, admission to a neonatal intensive care, and neonatal jaundice also were more frequent in women having vacuum-assisted delivery. On the other hand, women having forceps delivery had more third- or fourth-degree lacerations (36.9% vs 26.8%, P < 0.001). There were no significant group differences in umbilical artery blood gases, intracranial hemorrhage, or other neonatal injuries. Most of these findings persisted on multivariate analysis. The adjusted odds ratio (AOR) for shoulder dystocia in the forceps group compared with the vacuum group was 0.34 (95% confidence interval [CI], 0.20-0.57). The AOR for third- or fourth-degree lacerations when comparing forceps delivery with vacuum delivery was 1.79 (95% CI, 1.52-2.10). When parity was considered, forceps delivery was associated with a lower rate of shoulder dystocia in both nulliparous (1.4% vs 3.1%, P < 0.001) and multiparous (1.7% vs 4.5%, P = 0.012) women. However, only nulliparous women had higher rates of third- and fourth-degree lacerations (40.9% vs 30.7%, P < 0.001 in nulliparas and 19.6 vs 15.7%, P = 0.107 in multiparas).