Episiotomy, an incision through the perineum to facilitate delivery, is a common obstetric procedure.1 For decades, evidence has shown that routine episiotomy has unclear benefit but substantial risk of third- and fourth-degree extension and perineal trauma.2–4 In 2006, the American College of Obstetricians and Gynecologists recommended against routine episiotomy,5 and, in 2008, the National Quality Forum recognized limiting routine episiotomy as an important measure of quality and patient safety.6 It is unclear to what degree episiotomy has decreased in the setting of evidence-based recommendations. Additionally, little is known regarding how hospital and patient factors are related to likelihood of episiotomy. We used a large administrative database to investigate contemporary use of episiotomy.
We conducted a repeated cross-sectional analysis of nonoperative vaginal delivery hospitalizations in the National Inpatient Sample from 2000 to 2018. The National Inpatient Sample is one of the largest publicly available, all-payer inpatient databases in the United States and approximates a 20% stratified sample of all hospitalizations nationally.7 Delivery hospitalizations for females aged 15–54 years were included.8,9 Cesarean and operative (vacuum and forceps) vaginal deliveries were excluded. We characterized temporal trends in use of episiotomy reporting the average annual percent change with 95% CIs using the National Cancer Institute’s Joinpoint Regression Program 220.127.116.11.10,11 We performed unadjusted and adjusted log-linear regression models to evaluate the association between demographic, hospital, and clinical factors and likelihood of episiotomy.12 Given known differentials in episiotomy by race, we included this variable in the analysis. Measures of association are reported as unadjusted and adjusted risk ratios (aRRs) with 95% CIs. Given that the data are de-identified, the analysis was deemed exempt by the Columbia University institutional review board (AAAE8144). Two sensitivity analyses were additionally performed. First, we excluded deliveries with the following possible indications for episiotomy: shoulder dystocia, fetal heart rate abnormalities, and concerning fetal status. Second, we repeated the analysis using National Inpatient Sample population weights that can be applied to the 20% sample to create national estimates.13
After applying study criteria, 9,654,749 deliveries were identified and included in the analysis. The episiotomy rate during nonoperative vaginal deliveries fell from 26.4% in 2000 to 4.9% in 2018. The average annual percent change in episiotomy was −8.9% (95% CI −9.8% to −7.9%), with marked reductions across geographic regions (Fig. 1). Non-Hispanic White race (aRR 1.61 95% CI 1.59–1.62), private insurance (aRR 1.46, 95% CI 1.46–1.47), highest ZIP code income quartile (aRR 1.19 95% CI 1.19–1.20), and deliveries at urban, nonteaching hospitals (aRR 1.36 95% CI 1.36–1.37) had higher risk of episiotomy (Table 1). Trends and adjusted models were similar in the sensitivity analyses of 1) 8,618,242 deliveries excluding shoulder dystocia, fetal heart rate abnormalities, and concerning fetal status; and 2) an estimated 46,404,921 deliveries derived from applying population weights (data not shown). Episiotomy was higher in the presence compared with absence of shoulder dystocia (20.3% vs 14.1%, P<.01).
Reference: doi: 10.1097/AOG.0000000000004631