Enhanced detection of neonatal invasive infection clusters in South Africa using epidemiological and genomic surveillance data

IntroductionInvasive bacterial infections, particularly those caused by the ESKAPE group, account for a substantial proportion of neonatal deaths in low- and middle-income countries, yet the contribution of healthcare-associated transmission of such infections remains poorly defined. Pathogen whole genome sequencing (WGS) offers superior resolution to delineate transmission events compared to conventional methods. We aimed to determine the proportion of neonatal infection cases attributable to healthcare-associated transmission by comparing classical epidemiological and enhanced genomic approaches. MethodsWe retrospectively analyzed infants with culture-confirmed bloodstream infection or meningitis from six lower-tier hospitals in South Africa from September 2019 through October 2020. Isolates including Klebsiella pneumoniae, Acinetobacter baumannii, Staphylococcus aureus, Enterococcus faecium, Enterococcus faecalis, and Escherichia coli underwent antimicrobial susceptibility testing (AST) and WGS. Clusters were delineated using: 1) SNP-EPI (WGS-based reference standard), defined as 2 or more cases in the same hospital within 14 days and 11 or less single nucleotide polymorphisms (SNPs) (10 or less for A. baumannii); and 2) AST-EPI (conventional approach), defined as 2 or more cases with identical AST patterns within 14 days in the same hospital. The healthcare-associated transmission rate was calculated as the proportion of all cases allocated to SNP-EPI clusters. ResultsWe included 428 cases of infection with both AST and WGS data, including K. pneumoniae (40%, n=172), A. baumannii (23%, n=98), and S. aureus (21%, n=90). The overall healthcare-associated transmission rate was 31% (131/428 cases); this was highest for A. baumannii (50%, 47/98) followed by K. pneumoniae (42%, 72/172) and S. aureus (11%, 10/90), with substantial variation by facility (range, 5%-95%). We identified 39 SNP-EPI clusters across all hospitals, mostly involving multi-drug-resistant (MDR) strains, including large outbreaks of carbapenem-resistant K. pneumoniae ST152 and high-risk A. baumannii ST1 and ST2 clones. The AST-EPI criteria detected only 28 clusters and yielded a lower overall healthcare-associated transmission rate of 18%. Compared to the SNP-EPI approach, the AST-EPI method particularly underestimated transmission for K. pneumoniae (23% vs. 42%) and A. baumannii (29% vs. 48%), often fragmenting larger, persistent SNP-EPI clusters. ConclusionsHealthcare-associated transmission contributed substantially to the burden of invasive bacterial infections among South African infants, primarily driven by MDR K. pneumoniae and A. baumannii. WGS provided greater resolution than conventional methods for cluster detection. Genomic surveillance should be integrated into routine practice for effective outbreak detection and informed infection prevention and control responses in low-resource settings.