Mackrodt 1998.

Methods	RCT. Factorial 2 x 2 design.
Participants	Setting ‐ Ipswich Hospital (NHS Trust), Ipswich, UK. 1780 women randomised. Inclusion criteria ‐ initially women who sustained an episiotomy or laceration (first or second degree) during a spontaneous vaginal delivery and had given their informed consent to participate were included. However, the trial was extended to include women who were delivered by a simple instrumental delivery (nonrotational forceps or vacuum extraction). Exclusion criteria ‐ not documented. Parity ‐ primigravida and multigravida included (split equally between groups). Mean age ‐ intervention group = 28.2; comparison group B = 28.4. Operator ‐ midwives and doctors.
Interventions	Method of repair ‐ each group had 50% of women randomly assigned for perineal repair using a 2‐stage (skin unsutured) technique and 50% assigned for perineal repair using the 3‐stage (skin sutured) method. Intervention group (n = 889) ‐ sutured with polyglactin 910 (Vicryl), gauge 2‐0 on 35 mm needle. Control group (n = 891) ‐ sutured with chromic catgut on 40 mm needle.
Outcomes	Short‐term pain ‐ day 2 and 10. Analgesia ‐ day 2 and 10 and 3 months. Tight stitches ‐ 2 and 10 days. Removal of sutures ‐ 10 days and 3 months. Resumption of sexual intercourse ‐ 3 months. Failure to achieve pain‐free intercourse ‐ 3 months. Suture dehiscence ‐ day 10 and 3 months (appearance of perineum, gaping, healing by first intention, healing by secondary intention, breaking down at 10 days and resuturing at 3 months).
Notes	The operator could 'choose' method of repair for perineal skin (subcutaneous or interrupted). In the group that had the perineal skin sutured ‐ 26% had subcuticular stitches inserted; 72% had interrupted transcutaneous stitches;1% had skin left unsutured and 1% had no sutures. 6 women who had a third degree laceration were recruited in error but were included in the analysis. Interim analysis carried out.
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Allocated randomly using balanced blocks varying in size between 4 and 12 ‐ stratified by type of delivery.
Allocation concealment?	Low risk	Concealed treatment allocation ‐ serially numbered; sealed opaque envelopes containing allocation details, suture material and data sheet. All envelopes accounted for.
Blinding? Women	Unclear risk	No details given.
Blinding? Clinical staff	High risk	Unable to blind operator due to obvious difference in suture methods and materials.
Blinding? Outcome assessors	High risk	Outcome assessment not fully blinded (unable to fully blind outcome assessment due to obvious difference in suture methods and materials).
Incomplete outcome data addressed? All outcomes	Low risk	99% completed questionnaires at 24‐48 hours and 93% at 3 months postpartum.
Free of other bias?	Low risk	No baseline imbalance apparent.