RESUMO
OBJECTIVE: To identify the variables associated with sepsis-associated mortality, as well as to develop a severity risk score to predict death in very low-birth-weight (VLBW) neonates affected by nosocomial sepsis. STUDY DESIGN: Retrospective cohort study. Infants weighing ≤ 1500 g with neonatal sepsis of nosocomial origin were included. Epidemiological, clinical, and laboratory variables were recorded at onset (0 hours), at 12 hours, and at 24 hours. Bivariate and multivariate analyses were performed. RESULT: The study population included 95 VLBW infants who presented 129 episodes of nosocomial sepsis. In the bivariate analysis, gram-negative bacilli as the etiology of sepsis, seizures, age, postconceptional age, weight, serum procalcitonin (24 hours), platelet count (24 hours), blood urea nitrogen (0 and 24 hours), creatinine (24 hours), diuresis (12 and 24 hours), mean blood pressure (12 and 24 hours), pH, base excess (0, 12, and 24 hours), and SpO(2) (pulse oximetric saturation):F(IO(2)) (fraction of inspired oxygen) ratio (12 and 24 hours) were significantly associated with mortality. In the multivariate analysis, weight at the onset of sepsis, base excess (0 hours), and SpO(2):F(IO(2)) ratio (12 hours) were independent predictors of mortality. CONCLUSION: A lower weight at the onset of sepsis, base excess, and SpO(2):F(IO(2)) ratio are useful to predict nosocomial sepsis-associated mortality in VLBW infants.
