Instrumental delivery is the main risk factor reported for Obstetric Anal Sphincter Injury (OASI) occurrence. Low level of evidence data in the literature is likely to explain why there are no clear guidelines about the indication of mediolateral episiotomy to prevent from OASI during instrumental delivery. Our main objective was to assess the impact of the French 2005 guidelines to reduce episiotomy during instrumental delivery within the AUDIPOG (Pediatric Electronic Records Users Association, Obstetrics and Gynecology) French database [1]. These guidelines aimed to reduce the rate of episiotomy to less than 30% among vaginal deliveries. Our secondary objectives were to assess the impact of such guidelines on the rate of OASI during instrumental delivery and the episiotomy rate according to the type of instrument used for the delivery.

This historical cohort study concerned deliveries included in the AUDIPOG sentinel network database (https://www.audipog.net) from 256 French participating maternities. We selected in the database deliveries that occurred from 2000 to 2016 by excluding pre-term birth (less than 34 weeks of gestation), cesarean sections, non-singleton pregnancy, breech presentations, operative maneuver for shoulder dystocia and spontaneous deliveries leading to include 96,035 deliveries. The main outcome was the episiotomy rate. The secondary outcome was OASI occurrence defined as a third- or fourth degree perineal tears according the RCOG classification [2].
First, social, demographic, obstetrical and delivery characteristics of women with an instrumental delivery had been compared according to the selected periods (2000-2005 vs. 2006-2011 vs. 2012-2016). Secondly, we compared OASI occurrence according to the selected study periods, and the episiotomy and OASI rate according to the instrument required for the delivery (vacuum, forceps, spatulas). Categorical variables were compared using χ2 tests (or Fischer’s exact test, if applicable) and continuous variables by the Student’s test. Crude relative risks (RR) of episiotomy and OASI for delivery periods (2000-2005 vs. 2006-2011 vs. 2012-2016) were calculated, with their 95% confidence intervals (95%CI). A log-binominal model was used to adjust for confounding factors or pertinent clinical prognostic factors according to each studied outcome in our work (level of maternity units, and/or parity, and/or head circumference, and/or percentile of birthweight according to gestational age and sex, and/or geographical origin, and/or BMI, and/or type of instrument, and episiotomy for the OASI outcome). To assess the global evolution of the episiotomy and OASI rates, we conducted time series analyses which were performed with a Prais-Winsten regression based on generalized least-squares method. Secondly, to compare evolution of the episiotomy and OASI rates according to the studied periods, we performed an autoregressive integrated moving-average (ARIMA) model for time-series. The results were expressed as adjusted odd-ratios (aOR) and 95%CI and were plotted using a forest-plot. The level of statistical significance was p<0.05.

We reported during the study period an increasing proportion of women of 35 years or more, with a BMI of 25Kg.m-2 or more and women with an history of cesarean section. About the characteristics of the instrumental delivery, there was an increase in the number of women who delivered under epidural analgesia across the time periods. Through the considered periods, vacuum became the main instrument with an increase from 28.5% to 44.3% whereas the use of forceps decreased from 48% to 28.9% (p<10-4). The proportion of neonates of more than 4000g decrease across the time but without any significant change about cranial perimeter (p<10-4). The overall rate of episiotomy during instrumental delivery decreased between 2000 (83.9%) and 2016 (56.4%) (Fig 1). The range of such a decrease was dependent from the parity and the type of considered instrument. Compared to 2000-2005, the aRR of having an episiotomy for 2006-2011 was 0.93 [95% CI: 0.92-0.95] and for 2012-2016 was 0.89 [95%CI: 0.87-0.90] (Table 1). Adjustment was done for geographical origin, parity, type of instrument, birthweight and head circumference. All the aRR for the selected subgroups according to parity, type of instrumental delivery, stressed a statistically significant risk reduction for episiotomy excepted for the subgroup of multiparous and forceps deliveries about the 2006-2011 period compared to the 2000-2005 period. The aRR of having an OASI, compared to 2000-2005, was 1.30 [95% CI: 1.10-1.53] in 2006-2011 and 1.57 [95%CI:1.33-1.85] in 2012-2016 (Table 1). Adjustment was done for the same factors than those reported above plus episiotomy. All the aRR for OASI occurrence stressed a significant risk reduction excepted for spatula, vacuum and forceps groups for whom the difference was statistically significant only when comparing 2011-2016 to 2000-2005 (Table 1). A significant variations of episiotomy rates were highlighted through the studied years (from 2000 to 2016). For the OASI rate, there was also a significant decrease from 2000 to 2016 (p=0.038) but the multivariate analysis failed to report statistically significant difference (p=0.10) (Figure 1).

A first strength of this study is that we report data from a national multicentric cohort about 96,035 deliveries in a country (France) having a strong culture in favor of vaginal delivery. Another strength is that we aim to report an analysis with a dynamic approach by investigating the effect of our national guidelines on the rate of episiotomy and OASI. Most of available studies provides analysis from national registers with an assessment of the association between episiotomy and OASI occurrence without considering the potential changes across the time in episiotomy use. The effect of decreasing the use of episiotomy during instrumental delivery on the OASI occurrence might have been attenuated by a change in the modalities of instrumental deliveries. Indeed, we reported in our data a significant decrease of the use of forceps (from 48 to 28.9%) in benefits of vacuum delivery (from 28.5 to 44.3%) who became the most frequently used instrument. There is high level of evidence data reporting that the risk of OASI is more important when the delivery is performed using a forceps compared to a vacuum, meaning when using an instrument that increase the diameter of cephalic presentation. Another point is that we observed a change in the neonate’s characteristics with a significant decrease of neonates large for gestational age at birth (from 9.3 to 7.6%) which is an important risk factor for OASI.  Another factor that may have attenuated the effect of restricting the use of episiotomy on OASI occurrence is the reduction of the proportion of nulliparous women which is the most important risk factor reported in the literature. Finally, regarding these changes during the whole period considered in this study, even if we did not report a significant increase in OASI occurrence, this does not reject the hypothesis of an association between episiotomy and OASI during instrumental delivery. This highlights the limitations of considering available international retrospective studies for assessing the effect of episiotomy in this indication and explain the low level of evidence of international guidelines [3].

After the publication of the 2005 French guidelines in favor of a restrictive use of episiotomy we observed, globally, a significant decrease in episiotomy uses without a significant increase in OASI occurrence. Several changes in women and neonate’s characteristics as well as in instrumental delivery modalities might have attenuated the effect of reducing episiotomy use. Further prospective population-based studies are required to efficiently assess the role of episiotomy in OASI prevention during instrumental delivery.

CNGOF. Episiotomy: recommendations ogf the CNGOF for clinical practice. J Gynecol Obstet Biol Reprod. 2006;34:275-9.
Royal College of Gynaecologists and Obstetricians (RCOG). Green-top Guideline No. 29. The management of third- and fourth-degree perineal tears. 2015. Available at: https://www.rcog.org.uk/en/guidelines-research-services/guidelines/gtg29/ (last consultation on the 30/03/2021).
van Bavel J, Hukkelhoven C, de Vries C, Papatsonis DNM, de Vogel J, Roovers JWR et al. The effectiveness of mediolateral episiotomy in preventing obstetric anal sphincter injuries during operative vaginal delivery: a ten-year analysis of a national registry. Int Urogynecol J. 2018;29:407-13.