LAUSR

Objectives Neonatal mortality is generally 20% higher in boys than girls due to biological phenomena. Only a few studies have examined more finely categorised age patterns of neonatal mortality by sex, especially in the first few days of life. The objective of this study is to examine sex differentials in neonatal mortality by detailed ages in a low-income setting. Design This is a secondary observational analysis of data. Setting Rural Sarlahi district, Nepal. Participants Neonates born between 1999 and 2017 in three randomised controlled trials. Outcome measures We calculated study-specific and pooled mortality rates for boys and girls by ages (0–1, 1–3, 3–7, 7–14, 14–21 and 21–28 days) and estimated HR using Cox proportional hazards models for male versus female mortality for treatment and control groups together (n=59 729). Results Neonatal mortality was higher in boys than girls in individual studies: 44.2 vs 39.7 in boys and girls in 1999–2000; 30.0 vs 29.6 in 2002–2006; 33.4 vs 29.4 in 2010–2017; and 33.0 vs 30.2 in the pooled data analysis. Pooled data found that early neonatal mortality (HR=1.17; 95% CI: 1.06 to 1.30) was significantly higher in boys than girls. All individual datasets showed a reversal in mortality by sex after the third week of life. In the fourth week, a reversal was observed, with mortality in girls 2.43 times higher than boys (HR=0.41; 95% CI: 0.31 to 0.79). Conclusions Boys had higher mortality in the first week followed by no sex difference in weeks 2 and 3 and a reversal in risk in week 4, with girls dying at more than twice the rate of boys. This may be a result of gender discrimination and social norms in this setting. Interventions to reduce gender discrimination at the household level may reduce female neonatal mortality. Trial registration number NCT00115271, NCT00109616, NCT01177111.