Clinical and laboratory factors associated with neonatal sepsis mortality at a major Vietnamese children’s hospital

Sepsis is a major cause of neonatal mortality and children born in low- and middle-income countries (LMICs) are at greater risk of severe neonatal infections than those in higher-income countries. Despite this disparity, there are limited contemporaneous data linking the clinical features of neonatal sepsis with outcome in LMICs. Here, we aimed to identify factors associated with mortality from neonatal sepsis in Vietnam. We conducted a prospective, observational study to describe the clinical features, laboratory characteristics, and mortality rate of neonatal sepsis at a major children’s hospital in Ho Chi Minh City. All in-patient neonates clinically diagnosed with probable or culture-confirmed sepsis meeting inclusion criteria from January 2017 to June 2018 were enrolled. We performed univariable analysis and logistic regression to identify factors independently associated with mortality. 524 neonates were recruited. Most cases were defined as late-onset neonatal sepsis and were hospital-acquired (91.4% and 73.3%, respectively). The median (IQR) duration of hospital stay was 23 (13–41) days, 344/524 (65.6%) had a positive blood culture (of which 393 non-contaminant organisms were isolated), and 69/524 (13.2%) patients died. Coagulase-negative staphylococci (232/405; 57.3%), Klebsiella spp. (28/405; 6.9%), and Escherichia coli (27/405; 6.7%) were the most isolated organisms. Sclerema (OR = 11.4), leukopenia <4,000/mm3 (OR = 7.8), thrombocytopenia <100,000/mm3 (OR = 3.7), base excess < –20 mEq/L (OR = 3.6), serum lactate >4 mmol/L (OR = 3.4), extremely low birth weight (OR = 3.2), and hyperglycaemia >180 mg/dL (OR = 2.6) were all significantly (p<0.05) associated with mortality. The identified risk factors can be adopted as prognostic factors for the diagnosis and treatment of neonatal sepsis and enable early risk stratification and interventions appropriate to reduce neonatal sepsis in LMIC settings.