Objectives To determine the relation between place and skilled birth attendance at birth and early neonatal mortality. Design Retrospective analysis using data from Demographic and Health Surveys on obstetric complications. Setting Nine low and middle income countries between 2006 and 2013. Population 71 758 women aged 15–49 years. Methods A secondary analysis was carried out to investigate the occurrence and effect of obstetric complications on early neonatal mortality and association with place and attendance at birth. Obstetric complications studied were prolonged labour, puerperal infection and eclampsia. Main outcome measures Association between early neonatal mortality and place and attendance at birth, unadjusted and adjusted for presence of severe obstetric complications. Results Thirty-five percent of all births were at home: 70% of these were without skilled attendamts. Obstetric complications were reported in 17 079 women: 82% of these women gave birth in health facilities. Overall, no association was observed between place of birth or attendance at birth and early neonatal mortality. When adjusted for obstetric complications, the odds of early neonatal deaths for births at home without a skilled attendant were 1.3 (95% CI 1.1–1.5) compared with 1.2 (95% CI 1.0–1.5) with a skilled attendant and births in health facilities. Conclusions When adjusted for obstetric complications, births in health facilities were associated with reduced early neonatal mortality. However, reporting and referral bias account for at least part of the association. Tweetable abstract Births in health facilities are linked with fewer early newborn deaths when adjusted for obstetric complications.