Abstract

Background: Aetiological data for neonatal infections are essential to inform policies and programme strategies at various levels, but such data are scarce from sub-Saharan Africa. We therefore conducted a systematic review and meta-analysis of available data from the African continent since 1980, with a focus on regional differences in aetiology and antimicrobial resistance (AMR) in the last decade (2008 – 2018). Methods: We included data for microbiologically confirmed invasive bacterial infection including meningitis and AMR among neonates in sub-Saharan Africa and assessed the quality of scientific reporting according to Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection (STROBE-NI) checklist. We calculated pooled proportions for reported bacterial isolates and AMR. Findings: We included 151 studies comprising data from 84534 neonates from 26 countries, almost all of which were hospital-based. Of the 82 studies published between 2008 and 2018, insufficient details were reported regarding most STROBE-NI items. Regarding culture positive bacteraemia/sepsis, S aureus, Klebsiella spp and E coli accounted for 25% (95% CI 21 – 29%) 21%, (16 – 27%) and 10% (8 – 10%) respectively. For meningitis, the predominant identified causes were Group B streptococcus 25% (16 – 33%), S pneumoniae 17% (9 – 26%), and S aureus 12% (3 – 25%). Resistance to WHO recommended β-lactams was reported in >68% of 904 cases and to aminoglycosides in 27% of 1176 cases. Interpretation: Hospital-acquired neonatal infections and AMR are a major burden in Africa, and improved surveillance is required. More population-based neonatal infection studies are also needed, and all studies should be reported according to standardised reporting guidelines, such as STROBE-NI to aid comparability and reduce research wastage.