Neurological sequelae of sepsis in very low birth weight infants: Umbrella review and evidence-based outcome tree Eurosurveillance

Sepsis is a frequent cause of death in very-low-birthweight infants and often results in neurological impairment. Its attributable risk of sequelae has not been systematically assessed. To establish an outcome tree for mapping the burden of neonatal sepsis, we performed systematic searches in MEDLINE®, EMBASE® and Cochrane to identify systematic reviews addressing sequelae of neonatal sepsis. We included cohort studies identified in systematic reviews and performed meta-analyses of attributable risks, measured as risk differences between infected and uninfected infants. Evidence quality was assessed using GRADE. Two systematic reviews met inclusion criteria. The first included nine cohort studies with 5620 participants and five different outcomes (neurodevelopmental impairment, cerebral palsy, vision impairment, hearing impairment, death) for analysis. Pooled risk differences varied between 4% (95%CI:2 to 10%) and 13% (95%CI:5 to 20%). From the second review we analyzed four studies with 472 infants. Positive predictive value of neurodevelopmental impairment for later cognitive impairment ranged between 67% (95%CI:22 to 96%) and 83% (95%CI:36 to 100%). Neonatal sepsis increases risk of permanent neurological impairment. Magnitude of this effect varies by outcome, with evidence quality being low to very low. These data were used to construct an outcome tree for neonatal sepsis. Attributable risk estimates for sequelae following neonatal sepsis are suitable not only for burden estimation but may serve as outcome parameters in interventional studies.