Cause of death in neonates in low- and middle-income settings is usually inferred rather than established. The CHAMPS network did it properly, using minimally invasive tissue sampling at post mortem — blood, cerebrospinal fluid, rectal, nasopharyngeal and oropharyngeal swabs, and biopsies of brain, lung and liver — with culture, PCR arrays and histopathology, and a multidisciplinary panel adjudicating each case. It covered 2,609 neonatal deaths between December 2016 and December 2023 across Ethiopia, Kenya, Mali, Mozambique, Sierra Leone, South Africa and Bangladesh.
Infection was somewhere in the causal pathway in 1,147 deaths (44.0%) and was the underlying cause in 432 (16.6%). Gram-negative organisms accounted for 850 (74.1%) of infection-related deaths. The pathogen list is the finding: Klebsiella pneumoniae in 478 (41.7%), Acinetobacter baumannii in 295 (25.7%), Escherichia coli in 119 (10.4%), and Group B Streptococcus in only 65 (5.7%). Nearly a third of infection deaths were polymicrobial. The panel judged that over 80% could have been averted under current or improved facility conditions.
The implication for empirical therapy is uncomfortable and direct. Ampicillin plus gentamicin, the standard first-line neonatal sepsis regimen worldwide, was designed for a pathogen mix dominated by Group B Streptococcus and E coli. It does not reliably cover Klebsiella and covers Acinetobacter poorly to not at all — and these two organisms together account for two-thirds of the deaths here. The regimen is not wrong because it is old; it is wrong because the organisms changed and the regimen did not.
For Indian neonatal units this is confirmation rather than news, but it is confirmation with autopsy-grade evidence behind it, which is what has been missing from the argument. The action is local: know your own unit's neonatal isolate and susceptibility pattern, and let that rather than an international default set the empirical regimen. The other half of the finding matters just as much — the split between presumed community-acquired and hospital-acquired infection differed, and 80% avertability points at infection prevention, hand hygiene and device care as much as at antibiotic choice.
- Review your unit's empirical neonatal sepsis regimen against your own isolate and susceptibility data, not against a textbook default.
- Assume Klebsiella and Acinetobacter, not Group B Streptococcus, dominate late and hospital-acquired neonatal sepsis in this region.
- Distinguish early community-acquired from hospital-acquired episodes; the pathogen mix and the right empirical choice differ.
- Nearly a third of fatal infections were polymicrobial, so a single narrow organism on culture may not be the whole story.
- Most of these deaths were judged avertable — audit line care, hand hygiene and cohorting alongside any antibiotic change.
The statistics, in plain English
This is an observational study, and its strength is the method rather than any comparison. Attributing cause of death by expert panel using tissue sampling, culture, PCR and histology is far more reliable than verbal autopsy or clinical coding, which is how most such figures are generated. Two cautions. Only deaths were studied, so these proportions describe what kills neonates, not what infects them — organisms that are common but rarely fatal are under-represented. And the sites were deliberately chosen for high child mortality, so the pathogen mix reflects those settings rather than every hospital in the region.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for infectious diseases, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free