Background:
Bloodstream infections (BSIs) are a major cause of morbidity and mortality in hospitalized neonates. Data on antibiotic resistance in neonatal BSIs and their impact on clinical outcomes in Africa are limited.

Methods:
We conducted a prospective cohort study at 2 tertiary level neonatal intensive care units (NICUs) in Ghana. All neonates admitted to the NICUs were included from October 2017 to September 2019. We monitored BSI rates and analyzed the effect of BSI and antibiotic resistance on mortality and duration of hospitalization.

Results:
Of 5433 neonates included, 3514 had at least one blood culture performed and 355 had growth of a total of 368 pathogenic microorganisms. Overall incidence of BSI was 1.0 (0.9–1.1) per 100 person days. The predominant organisms were Klebsiella pneumoniae 49.7% (183/368) and Streptococcus spp. 10.6% (39/368). In addition, 512 coagulase negative Staphylococci were isolated but considered probable contaminants. Among K. pneumoniae, resistance to gentamicin and amikacin was 91.8% and 16.4%, respectively, while carbapenem resistance was 4.4%. All-cause mortality among enrolled neonates was 19.7% (1066/5416). The mortality rate was significantly higher in neonates with BSI compared with culture-negative neonates in univariate analysis (27.9%, n = 99/355 vs. 16.5%, n = 520/3148; hazard ratio 1.4, 95% confidence interval 1.07–1.70) but not in multivariate analysis.

Conclusion:
The diversity of etiologic agents and the high-risk of antibiotic resistance suggest that standard empirical treatment is unlikely to improve the outcome of BSIs in low and middle income. Such improvements will depend on access to reliable clinical microbiologic services.