Laboratory-confirmed bloodstream infections at two large neonatal units in Sub-Saharan Africa.

INTRODUCTION Epidemiologic data on neonatal bloodstream infections (BSI) in sub-Saharan Africa is extremely limited. METHODS Comparative analysis of laboratory-confirmed neonatal BSI (LC-BSI) episodes was conducted retrospectively at two large neonatal units in Botswana and South Africa (1 January-31 December 2017). Routine laboratory and ward register data was used to determine BSI rates, pathogen spectrum and BSI outcome. RESULTS In 2017, the Princess Marina Hospital (PMH) and Tygerberg Hospital (TBH) neonatal units admitted 1187 and 2826 neonates respectively.BSI incidence rates were 12.1/1000 patient-days (95%CI 10.2-14.3) at PMH and 3.5/1000 patient-days (95%CI 2.9-4.1) at TBH (p < 0.0001). Most BSI episodes were hospital-acquired (260/284; 91.6%). Blood culture contamination rates were substantially higher at PMH than TBH (152/1116 [13.6%] versus 122/2559 [4.8%]; p < 0.001). The crude mortality rate in neonates with BSI was 21.2% (53/250) and significantly higher at TBH than PMH (38/128 [29.7%] versus 15/122 [12.3%], p = 0.001). Factors independently associated with death were birth weight <1500 grams, adjusted odds ratio (aOR) 2.8 (1.3-6.4) and male sex, aOR 2.1 (1.1-3.7) (p = 0.02 and p = 0.01 respectively). Klebsiella pneumoniae was the dominant BSI pathogen at both units, accounting for two-thirds of BSI and associated with a large infection outbreak at PMH. Antibiotic resistance rates were substantial at both neonatal units, particularly for K. pneumoniae (98/122 [80.3%] extended-spectrum β-lactamase producers, ESBL) and Staphylococcus aureus (22/33 [66.7%] methicillin-resistant). CONCLUSION BSI rates and associated mortality was substantial at these 2 Sub-Saharan African neonatal units. K. pneumoniae ESBL remains a leading BSI and outbreak pathogen.