Eriksen 1997.

Methods	Randomized controlled trial, unblinded.
Participants	1024 term pregnant women presenting to labour and delivery. Included: 947 women (481 in chlorhexidine group (intervention) and 466 in sterile water group (control). Excluded after randomization: 77 women (71 incomplete records, 3 discharged home, 2 vaginal wash was not given, 1 anencephalic).
Interventions	Vaginal wash with 20 mL of 0.4% chlorhexidine solution versus sterile water.
Outcomes	Neonatal infection (pneumonia or sepsis) and use of antibiotics in neonates. Peripartal infection (intra‐amniotic infection and endometritis).
Notes	The report by Eriksen 1997 provided details of the neonatal outcomes. The report by Sweeten 1997 provided details of the maternal outcomes. This study was conducted in Houston, Texas, USA.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A computer software program (True Epistat, Richardson, TX) was used to generate a random block allocation sequence to assign women to either group.
Allocation concealment (selection bias)	Low risk	The randomization assignments were contained in sequentially numbered, opaque, sealed packets that were made up independent of the physicians managing the women. The randomization was known only after the woman had been enrolled in the study.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The authors chose not to blind the study because the syringe containing the chlorhexidine solution was pink and "we could not reproduce the colour artificially in the syringe containing the sterile water".
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Additionally, the physicians managing infants in the nursery were unaware to which arm of the study each participant was randomized.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1024 women were eligible for the study. Of these, 77 were excluded from the analysis for the following reasons: incomplete records (71), the patient was enrolled and subsequently discharged home (3), vaginal wash was not given (2), and a patient whose infant had anencephaly was inadvertently enrolled. Of the remaining 947 women, 481 were randomized to the study arm and 466 served as controls. 24 neonatal charts in the chlorhexidine group and 13 in the control group were unavailable for review. This left 457 neonates in the study group and 453 in a control group for analysis.
Selective reporting (reporting bias)	Unclear risk	Protocol was not available for review.
Other bias	Low risk	No other obvious biases.