Post-mortem characterisation of pathogen-specific causes of infection-related deaths in African and south Asian neonates: a prospective, observational, multicentre study

Background: Post-mortem biological investigation of causes of deaths is uncommon in low-income and middle-income countries (LMICs). The primary objective of Child Health and Mortality Prevention Surveillance (CHAMPS) is to determine the underlying and contributing causes of death for stillbirths and children younger than 5 years, including identification of specific pathogens and risk factors associated with mortality. We report on the role of infections, including pathogen-specific causes, in the overall causal pathway to death in neonates (aged <28 days) across CHAMPS sites in Ethiopia, Kenya, Mali, Mozambique, Sierra Leone, and South Africa in Africa and Bangladesh in south Asia. Methods: CHAMPS is a prospective, multicentre, multicountry, observational study that is undertaken across 11 sites with defined catchment areas and high mortality rates among children younger than 5 years. Data collected in CHAMPS include clinical record abstraction such as pregnancy history and clinical presentation, laboratory investigations, and management during hospitalisation. Furthermore, post-mortem information on decedents includes verbal autopsy, and collection of biological specimens (blood, cerebrospinal fluid, rectal swab, swabs of nasopharynx and oropharynx, and tissue biopsies of brain, lung, and liver) using minimally invasive tissue sampling. Specimens were tested using microbial culture, TaqMan Array Card-based PCR, and histopathology including immunohistochemistry. Final causes of death were adjudicated for each decedent by a multidisciplinary Determination of Cause of Death (DeCoDe) panel. Findings: CHAMPS investigated 2609 neonatal deaths from Dec 17, 2016, through to Dec 31, 2023. Infections were implicated anywhere in the causal pathway to death in 1147 (44·0%) cases, including as the underlying cause in 432 (16·6%) decedents. Polymicrobial infection was diagnosed in 360 (31·4%) of 1147 infection-related deaths. Gram-negative bacteria were attributed to 850 (74·1%) infection-related deaths. Overall, the most common pathogens attributed to causing infection-related deaths were Klebsiella pneumoniae (478 [41·7%] of 1147), Acinetobacter baumannii (295 [25·7%]), Escherichia coli (119 [10·4%]), and Group B Streptococcus (65 [5·7%]). The relative contribution of pathogens to infection-related deaths differed when stratified by age group, across the sites, and between presumed community-acquired compared with presumed hospital-acquired infections. Over 80% of neonatal deaths with infectious causes could have been averted under current or improved facility-based conditions. Interpretation: The repertoire of the dominant bacterial pathogens implicated in the causal pathway to death in our study indicate a need for review of empirical antibiotic treatment for management of severe neonatal infections, as well as improved interventions aimed at prevention of neonatal infections in LMICs. Funding: Gates Foundation. Translations: For the Afan Oromo, Amharic, Bangla, Krio, Portuguese, Tigrigna and Zulu translations of the abstract see Supplementary Materials section.