Study design

Design: Community-based cluster-randomised trial.

Allocation: Clusters were computer randomised. There were 133 clusters, with a mean population of 4100 people.

Blinding: Interventions were not masked.

Study question

Setting: Rural area of Bangladesh.

Patients: All liveborn babies were eligible for enrolment. Of 35 908 eligible infants, 29 760 were enrolled. Over 85% received the intervention within 24 h of birth. Baseline characteristics were similar between the three groups.

Intervention: Clusters were randomised to three groups: topical 4% chlorhexidine once at birth, chlorhexidine at birth plus daily cord cleansing with chlorhexidine for 7 days, or dry cord care. Village health workers visited patients daily and applied the chlorhexidine, or provided cord care guidance.

Outcomes: Primary outcome was death within 28 days after birth. Cord redness and pus, as assessed by an unblinded health worker, were also analysed.

Follow-up period: Outcomes were assessed at days 3, 6, 9, 15 and 28–35.

Patient follow-up: An intention-to-treat analysis was performed. Final health status was available for 95% of patients enrolled.

Main results Newborns in …