Een mediolaterale episiotomie tijdens vaginale kunstverlossing bleek uit een relatief klein observationeel onderzoek (Macleod, 2008) geen beschermend effect te hebben op ontstaan van ‘obstetric anal sphincter injuries’ (OASIS), terwijl drie andere grote cohortstudies wel een beschermend effect lieten zien op het ontstaan van OASIS (Baghestan 2010; de Leeuw 2008; Raisanen 2009). Eerder advies was om een adequate primaire mediolaterale episiotomie sterk te overwegen bij vaginale kunstverlossing, tenzij het risico op een totaalruptuur als laag wordt ingeschat. Gezien de mogelijke complicaties en nadelen van een episiotomie wordt het eerdere advies opnieuw geanalyseerd aan de hand van nieuwe literatuur.

Routine versus restrictive episiotomy use

Episiotomy versus no episiotomy

Description of studies

No studies were included in the systematic literature analysis.

Results

No studies were included in the systematic literature analysis.

Conclusions

No studies were included in the systematic literature analysis.

Narrative literature analysis

Here is a narrative overview of all studies that were assessed in full text, comparing ‘routine episiotomy policy’ versus ‘restrictive episiotomy policy’ or 'episiotomy’ with ‘no episiotomy’ in women with operative vaginal deliveries. The process of literature selection, and the studies that were initially selected full text, will be described below. The full-text review process ended with the inclusion of two large-scale observational studies.

1. Studies that directly compared routine and restrictive episiotomy policy

For this comparison, only one multicenter pilot study by Murphy and Macleod (2008) was found, conducted in England. 200 women were randomized to routine use (99 women) or restrictive use (101 women) of episiotomy. However, in addition to the fact that this was an underpowered pilot study with an insufficient number of participants to draw conclusions, the study also did not clearly define which women were invited to participate in the trial or parallel-performed observational study. In the parallel-performed observational study, it was not clearly determined whether the routine use of episiotomy was encouraged or discouraged by practicing clinicians. Another reason for exclusion was the high number of forceps deliveries performed (up to 55% of routine deliveries and 62% of restrictive deliveries) which resulted in fewer women delivered by vacuum. In the Netherlands, vacuum delivery is almost exclusively performed in case of instrumental vaginal delivery. Forceps delivery results in a higher chance of perineal ruptures and obstetric anal sphincter injuries, and therefore the results of this study cannot be extrapolated to the Dutch situation.

The Cochrane systematic review by Jiang (2017) investigated the same comparison (section B of the review is about anticipated instrumental deliveries) and also only included the trial of Murphy and Macleod (2008).

2. Studies that compared episiotomy versus no episiotomy, and clearly described whether their policy towards episiotomy use was routine or restrictive

                2.1 Studies found in 2013

For this comparison, in the earlier synthesis of 2013, 4 observational studies were described (in the justifications-section): Baghestan (2010), de Leeuw (2008), Macleod (2008) (the parallel observational study to the pilot RCT of Murphy 2008) and Räisänen (2009). The 4 studies had the following limitations:

The studies by de Leeuw (2008) and Räisänen (2009) assessed whether mediolateral episiotomy could be identified as a protective factor in a multivariable model in a case control design. The study by de Leeuw (2008) included besides mediolateral episiotomy also median and midline episiotomy. The study by Räisänen (2009) included both women with operative and spontaneous vaginal deliveries. Both studies (n=21 254 and n=514 741) concluded that mediolateral episiotomy was protective in singleton vacuum-assisted deliveries. However, this was a risk factor assessment and the size of the effect of episiotomy could not be determined with this design.

The study by Baghestan (2010) assessed the change in OASIS incidence over time in Norway and the effect of risk factors. The study included all vaginal deliveries as well, and therefore results for the group requiring vacuum-assisted deliveries could not be assessed. It was also unknown if the policy was routinely or restrictive towards episiotomy and whether episiotomy was performed mediolaterally.

The study by Macleod (2008) was a parallel observational study to the Murphy and Macleod (2008) pilot described earlier. It had the same limitations as the pilot study described before: the use of the forceps rather than vacuum-assisted deliveries in the majority of the included participants. Episiotomy was performed more often in the forceps group, where no protective effect of episiotomy was found.

                2.2 Studies found in 2021

In the current search of 2021, 3 systematic reviews were found: the Cochrane systematic review by Jiang (2017), the systematic review by Lund (2014) and the systematic review by Sagi-Dain (2015). During the process of this module, an additional systematic review appeared by Okeahialam (2022).

The Cochrane systematic review by Jiang (2017) was described earlier in comparison 1 and excluded due to limitations described before.

The systematic reviews by Lund (2016) and Sagi-Dain (2015) were comparable, both comparing episiotomy with no-episiotomy in women with vacuum-assisted deliveries, both including cohort studies and case control studies. The reviews were not selected, and it was decided to assess the original studies for the following reasons:

• Both reviews missed a large number of studies that were possibly eligible for inclusion, which are described below and in the exclusion table.
• Both reviews included studies that did not report clearly on parity (primiparous or multiparous), type of operative vaginal delivery and the episiotomy approach (mediolateral, midline, lateral). This, even though Lund was planning only to include studies with vacuum-assisted primiparous deliveries and mediolateral/lateral episiotomies. And Sagi-Dain was planning to perform subgroup analyses based on parity and approach, which they did, but in this analysis they included studies that did not report parity or approach, and it was also unclear which studies were included in this sub group analysis.
• The midline and lateral episiotomies that were included in the systematic reviews, would not have been included in our analysis, because our search was limited to mediolateral episiotomies.
• Both reviews included studies that did not adjust for (a large number of) prognostic factors/confounders. These studies had serious limitations and were not assessed as such.
• Some observational studies were performed after Lund (2016) and Sagi-Dain (2015) appeared.

The newly found systematic review by Okeahialam (2022) (published during the writing process of this module) included more or less the same studies for the subgroup of nulliparous women delivered with vacuum, as we assessed before the systematic review appeared (as so with the same limitations). The review can be used to validate the results that are found in this module.

The working group initially found 8 observational studies matching the inclusion criteria (Ankarcrona, 2021; Bodner-Adler, 2018; Boujenah, 2019; De Vogel, 2012; Gachon, 2019; Jangö, 2014; Schreiber, 2020 and van Bavel, 2018). After reading full text, five small observational studies were found with very high risk of bias, due to unbalanced prognostic factors for which no appropriate adjustment was performed (Bodner-Adler, 2019; Boujenah, 2019 de Vogel, 2012; Gachon, 2019, and Schreiber, 2020). One study (Jangö, 2014) did a risk factor analysis including only women with a previous OASI, in which both vacuum extraction and episiotomy were regarded as risk factor and protective factors for OASIS, but from this analysis (study design) the size of the effect could not be estimated in women who have vacuum delivery. Therefore, the study was excluded.

This left only two new studies for the narrative section of this review, Ankarcrona (2021) and van Bavel (2018).

                Study description

The large-scale study by Ankarcrona (2021) was performed in Sweden, and policy was restrictive towards episiotomy use (31.5% of nulliparous vacuum deliveries). Ankarcrona (2021) described a population-based cohort study, using the Swedish Medical Birth Register from 2000 to 2011. Nulliparous women with a vacuum-assisted vaginal delivery at gestational age of at least 34 0/7 weeks. The study compared lateral or mediolateral episiotomy with no episiotomy. In total, 19 801/63 654 (31,5%) of nulliparous women were delivered with an episiotomy during vacuum delivery. Groups were significantly different at baseline, with more risk factors for perineal tears (OASI) in the episiotomy group described above, for which the authors used a propensity score to adjust for in the analysis. A total of 19 801 women were included in the episiotomy group and 43 853 women in the no-episiotomy group. The study used a propensity score and found the episiotomy group was significantly different from the no-episiotomy group in maternal age, maternal continent of birth, maternal height (<1,60m), cohabitation, female genital mutilation, gestational age, neonatal sex, epidural, intrapartum foetal distress, foetal head station, head position, head circumference, birth weight, Apgar score at 1 minute, shoulder dystocia, year of delivery and hospital of delivery.

The large-scale study by van Bavel (2018) was performed in The Netherlands, and policy was routinely towards episiotomy use (87% of primiparous vacuum deliveries). Van Bavel et al (2018) described a population-based cohort study deriving from a National database in The Netherlands, called the Dutch Perinatal Registry. Data from all vacuum (n = 159.340) and forceps (n = 11.629) assisted deliveries in hospitals in The Netherlands were registered, encompassing all data from hospital-based deliveries from 2000 to 2010. Mediolateral episiotomy is assumed one of the interventions that could decrease the risk of obstetric anal sphincter injuries. Both primiparous and multiparous women were included who delivered a singleton live infant in cephalic presentation with the help of an operative vaginal delivery (vacuum or forceps) at term (>37 weeks). A total of 113 619 women were included in the episiotomy group and 16 538 in the no episiotomy group. Van Bavel (2018) adjusted for maternal age, maternal ethnicity, foetal position, indication for intervention, duration second stage and birth weight in a multivariable analysis.

Study results

Ankarcrona (2021), included 63 654 women and the episiotomy rate was 32% among operative vaginal deliveries. They reported an adjusted absolute risk reduction¹ for the outcome measure OASIS of -3.66% (95%CI -4.31 to -3.01) favouring episiotomy. The unadjusted odds ratio was 0.79 (95%CI 0.76 to 0.84) favouring episiotomy and the number needed to treat was 27. In the episiotomy group 12.3% (2240/19 801) developed an OASIS and in the no episiotomy group 15.0% (6588/43 853).

Van Bavel (2018), with a routine episiotomy policy and episiotomy rate of 87%, reported an adjusted odds ratio of 0.09 (95%CI 0.07 to 0.11) favouring episiotomy and the number needed to treat was 9. Among the women that received episiotomy, 2.4% (2830/113 619) developed an OASIS, while in the no episiotomy group 14.0% (2317/16 538) developed an OASIS. The effect size was large, indicating a strongly protective effect for episiotomy.

Table 1 Episiotomy versus no episiotomy, outcome measure OASIS with studies divided towards the policy they used towards episiotomy in primiparous/nulliparous women with vacuum-assisted delivery

*Adjusted for maternal age, maternal ethnicity, position of foetus, indication for intervention, duration of second stage of labour and birth weight

**Adjusted using propensity scores, adjusting for all baseline variables with p-value <0.20 described above

There is a large difference in OASI-rate in the episiotomy group between van Bavel (2018) (routine use) and Ankarcrona (2021) (restrictive use) (2.4% versus 12.3%), while the OASI-rate in the no-episiotomy group was comparable (14.0% versus 15.0%). This could be due to the restrictive policy in Sweden, causing a smaller effect of episiotomy (Table 1).

Level of evidence of the literature

The level of evidence (GRADE method) is determined per comparison and outcome measure and is based on results from observational studies and therefore starts at level “low”. Subsequently, the level of evidence was downgraded if there were relevant shortcomings in one of the several GRADE domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. The level of evidence can also be upgraded if a) level of evidence was not downgraded before, b) no serious limitations are present, c) the magnitude of the effect is large, d) the confidence interval of the effect does not substantially overlap with the chosen clinical threshold, e) there is high confidence in the estimated effect from the observational study (GRADE handbook).

1. Routine versus restrictive episiotomy use

For the comparison routine versus restrictive episiotomy use, no studies were found that directly compared routine and restrictive policy in the described patient group, besides a pilot trial, that was excluded for reasons described before. However, the studies of van Bavel (2018) and Ankarcrona (2021) could be used for an indirect comparison, because policy in The Netherlands is routinely (87% of all primiparous vacuum deliveries received episiotomy) and policy in Sweden is restrictive (32% of all nulliparous vacuum deliveries received episiotomy). The GRADE handbook describes that if evidence comes from an indirect comparison (there is no comparison A versus B, but A versus comparison and B versus comparison), the evidence has to be downgraded for indirectness. The level of evidence for this comparison starts at a low GRADE (for cohort studies) and was downgraded to a very low GRADE. However, the working group does not expect a randomized controlled trial to be realized in the near future, because it is very difficult to find women willing to participate in a trial who would voluntarily allow themselves to be randomized either to a routine episiotomy policy or restrictive episiotomy policy, among women with an anticipated operative vaginal delivery. This is seen in the pilot studies of Murphy and Macleod (2008) that did not proceed with a trial to include the number of women needed for a powered analysis. Currently a randomized controlled trial is registered in Sweden (https://clinicaltrials.gov/ct2/show/NCT02643108)  in which women will be randomized to lateral episiotomy or no episiotomy when operative vaginal delivery by vacuum extraction is indicated. However with an estimated enrolment of 710 participants, the trial is estimated to be completed in 2030 and will compare a relatively restricted use of episiotomy with no episiotomy, giving no answer to the comparison in this paragraph.

2. Episiotomy versus no episiotomy

For the direct comparison of episiotomy versus no episiotomy in primiparous women with vacuum-assisted deliveries more studies were found. However, we excluded those studies because of the risk of bias. The studies of Ankarcrona (2021) and Van Bavel (2021) were of high quality and the effect of episiotomy on the prevention of OASI was large and could be confirmed by a recent meta-analysis that also contained other studies (Okeahialam, 2022).

The study of Van Bavel (2018) used Dutch registry data in which 96% of the total population at risk was included from 2000 to 2010 (n= 1,534,850) This article was seen as leading in the GRADE classification here, as it covered the Dutch situation which the working group is primarily interested in. However, the study used observational registry data which was not designed to answer the study question and therefore residual confounding issues could be present. Even so, we consider it unlikely that the main result of van Bavel 2018, use of a mediolateral episiotomy during both vacuum delivery and forceps delivery is associated with a five- to tenfold reduction in the rate of OASIS, would have found contrary results after full adjustments. Because of the large magnitude of the effect, the narrow confidence intervals, and because the dose-response effect (larger magnitude with a more routine episiotomy use), it was decided to upgrade it to a moderate GRADE.

A systematic review of the literature was performed to answer the following question: What is the effect of a routine mediolateral episiotomy for the prevention of OASIS grade 3 or 4 in women during an operative vaginal delivery?

Direct comparison

Because no suitable articles were found (as explained below), and no systematic literature analysis could be made, a narrative literature analysis was made about the comparison episiotomy versus no episiotomy in women with an operative vaginal delivery, subdividing studies with either a ‘routine’ policy towards episiotomy (in most cases) versus studies with a ‘restrictive’ episiotomy policy (only when strictly necessary). This is an indirect comparison, not suitable to answer the beforementioned search question directly. The process of literature selection and studies that were found and assessed is thoroughly explained in the narrative literature analysis.

Indirect comparison

Relevant outcome measures

The guideline development group considered OASIS as a critical outcome measure for decision making; and patient satisfaction as an important outcome measure for decision making.

The working group defined the outcome measures OASIS as follows: a third- or fourth-degree perineal tear as according to the Royal College of Obstetricians and Gynaecologists (RCOG) classification system. The working group did not define the outcome patient satisfaction but used the definition that was used in the studies.

The working group defined RR ≤ 0.7 as a minimal clinically (patient) important difference for both OASIS and patient satisfaction.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until February 24^th, 2021. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 186 hits. Studies were selected based on the following criteria: women with an operative vaginal delivery (vacuum or forceps), comparing routine episiotomy policy (in most cases) versus restrictive episiotomy policy (only when strictly necessary), investigating OASIS as one of the outcomes. Clinical trials and observational studies with a comparative design were included. Case-control studies and other studies with a non-comparative design were excluded. Studies that did not report specifically operative vaginal deliveries were excluded. Totally, 45 studies (since the last search in 2013) were initially selected based on title and abstract screening. After reading full texts, no studies could be included that directly compared routine episiotomy with restrictive episiotomy policy.

Results

No studies were included in the systematic literature analysis.