Dannecker, C., Hillemanns, P., et al. (2004). "Episiotomy and perineal tears presumed to be imminent: randomized controlled trial." Acta Obstetricia et Gynecologica Scandinavica. 83 (4): 364-8.

Summary: This prospective randomized trial compared maternal and perinatal outcomes of a restricted policy where episiotomy was only performed for fetal indications (n=49) and a liberal policy where episiotomy was performed for fetal indications or when a perineal tear was judged by the birth attendant to be imminent (n=60). Primiparous women with uncomplicated singleton pregnancies at >34 weeks gestation were enrolled in the study. Episiotomies were mediolateral (cut diagonally toward the buttocks) and were performed by obstetricians or midwives . Tears were judged to be imminent when the perineal tissue was pale and extremely thin during crowning of the fetal head.
Restrictive use of episiotomy only for fetal indications resulted in a 36 percent lower episiotomy rate and a three-fold increase in the rates of intact perinea and minor perineal trauma. There were no statistically significant differences in rates of third-degree tears, anterior lacerations, maternal blood loss, Apgar scores, or umbilical artery pH values. This study did not measure long-term outcomes such as pelvic floor strength, urinary incontinence or sexual functioning. The study took place in an institution in Germany that formerly had a policy of routine episiotomy. This may account for the high episiotomy rate in both study groups (41 percent in the restrictive group and 77 percent in the liberal group).
The authors conclude that performing mediolateral episiotomies when tears are presumed to be imminent provides no benefit to the mother or newborn and causes increased perineal trauma and postpartum pain. Significance for Normal Birth: Birth is normal when there are no routine interventions. In the United States , the rate of episiotomy has decreased significantly since the evidence began to show that restricted use of this intervention resulted in better maternal outcomes than routine use, with no adverse effects for the newborn. However, “restricted use” is poorly defined, and many practitioners continue to cut episiotomies for maternal indications, primarily to avoid a tear that is presumed to be imminent. This study shows that episiotomy rates may be further reduced, with better maternal outcomes and no adverse effects for the newborn, when the intervention is only performed for fetal indications such as acute fetal distress. While this study took place in Germany where mediolateral episiotomy is performed, it is probable that similar results would be found in the United States , where midline episiotomy (toward the anus) is the norm.