ObjectiveTo estimate the overall and infection-related neonatal mortality rate and the pathogens responsible using electronic death registrations.DesignRetrospective analysis of national electronic death registrations data.SettingEngland and Wales.PatientsNeonates aged <28 days.Main outcome measuresOverall and infection-related mortality rate per 1000 live births in term, preterm (28–36 weeks) and extremely preterm (<28 weeks) neonates; the contribution of infections and specific pathogens; comparison with mortality rates in 2003–2005.ResultsThe neonatal mortality rate during 2013–2015 (2.4/1000 live births; 5095 deaths) was 31% lower than in 2003–2005 (3.5/1000; 6700 deaths). Infection-related neonatal mortality rate in 2013–2015 (0.32/1000; n=669) was 20% lower compared with 2003–2015 (0.40/1000; n=768), respectively. Infections were responsible for 13.1% (669/5095) of neonatal deaths during 2013–2015 and 11.5% (768/6700) during 2003–2005. Of the infection-related deaths, 44.2% (296/669) were in term, 19.9% (133/669) preterm and 35.9% (240/669) extremely preterm neonates. Compared with term infants (0.15/1000 live births), infection-related mortality rate was 5.9-fold (95% CI 4.7 to 7.2) higher in preterm (0.90/1000) and 188-fold (95% CI 157 to 223) higher in extremely preterm infants (28.7/1000) during 2013–2015. A pathogen was recorded in 448 (67%) registrations: 400 (89.3%) were bacterial, 37 (8.3%) viral and 11 (2.4%) fungal. Group B streptococcus (GBS) was reported in 30.4% (49/161) of records that specified a bacterial infection and 7.3% (49/669) of infection-related deaths.ConclusionsOverall and infection-related neonatal mortality rates have declined, but the contribution of infection and of specific pathogens has not changed. Further preventive measures, including antenatal GBS vaccine may be required to prevent the single most common cause of infection-related deaths in neonates.