Background
Neonatal sepsis is a leading cause of child mortality, and increasing antimicrobial resistance threatens progress towards the Sustainable Development Goals. Evidence to guide antibiotic treatment for sepsis in neonates and young infants from randomized controlled trials or observational studies in low- and middle-income countries (LMICs) is scarce. We aimed to describe patterns of antibiotic use, pathogens and outcomes in LMIC hospital settings globally to inform future clinical trials on the management of neonatal sepsis.
Methods & Findings
Hospitalised infants aged <60 days with clinical sepsis were enrolled during 2018-2020 by 19 sites in 11 countries (mainly Asia and Africa). Prospective daily data was collected on clinical signs, supportive care, antibiotic treatment, microbiology and clinical outcome at 28 days. The study was observational, with no changes to routine clinical practice. 3204 infants were enrolled, with median birth weight 2500g (IQR 1400-3000) and postnatal age 5 days (IQR 2-15). Of 309 enrolled aged 28-60 days, 58.6% (n=181) were ex-preterm and/or a neonate at admission. 2215 (69%) infants had been in hospital since birth.
206 different empiric antibiotic combinations were used, which were structured into 5 groups that were developed from the World Health Organisation (WHO) AWaRe classification. 25.9% (n=814) of infants started a WHO first line regimen (Group 1 Access, penicillin-based regimen) and 13.8% (n=432) started WHO second-line cephalosporins (cefotaxime/ceftriaxone) (Group 2 Low Watch). The largest group (34.0%, n=1068) started a regimen providing partial extended-spectrum beta-lactamase (ESBL)/pseudomonal coverage (piperacillin-tazobactam, ceftazidime, or fluoroquinolone-based) (Group 3 Medium Watch), 18.0% (n=566) started a carbapenem (Group 4 High Watch), and 1.8% (n=57) started a Reserve antibiotic (Group 5, largely colistin-based). Predictors of starting non-WHO recommended regimens included lower birth weight, longer in-hospital stay, central vascular catheter use, previous culture positive sepsis or antibiotic exposure, previous surgery and greater sepsis severity. 728/2880 (25.3%) of initial regimens in Group 1-4 were escalated, mainly to carbapenems, and usually for clinical indications (n=480; 65.9%).
564 infants (17.6%) isolated a pathogen from their baseline blood culture, of which 62.9% (n=355) had a Gram-negative organism, predominantly Klebsiella pneumoniae (n=132) and Acinetobacter spp. (n=72). These leading Gram-negatives were both mostly resistant to WHO-recommended regimens, and also resistant to carbapenems in 32.6% and 71.4% of cases respectively. MRSA accounted for 61.1% of Staphylococcus aureus (n=54) isolates.
Overall, 350/3204 infants died (11.3%; 95%CI 10.2-12.5%), with 17.7% case fatality rate among infants with a pathogen in baseline culture (95%CI 14.7-20.1%, n=99/564). Gram-negative infections accounted for 75/99 (75.8%) of pathogen-positive deaths, especially Klebsiella pneumoniae (n=28; 28.3%), and Acinetobacter spp. (n=24; 24.2%).
Conclusion
A very wide range of antibiotic regimens are now used to treat neonatal sepsis globally. There is common use of higher-level Watch antibiotics, frequent early switching and very infrequent de-escalation of therapy. Future hospital based neonatal sepsis trials will ideally need to account for the multiple regimens used as standard of care globally and include both empiric first line regimens and subsequent switching in the trial design.
