Routine screening for colonization by Gram-negative bacteria in neonates at intensive care units for the prediction of sepsis: systematic review and meta-analysis

Background: At neonatal intensive care units, sepsis due to Gram-negative bacteria is an important cause of morbidity and mortality. The benefits of routine microbiological screening of neonatal body surface to predict and prevent sepsis are controversial.

Aim: To evaluate the prognostic value of neonatal body surface screening for colonization with Gram-negative bacteria for the prediction of late-onset sepsis.

Methods: A systematic review was performed, including studies of any design that reported data to calculate prognostic accuracy of surface screening for the prediction of late-onset sepsis. Risk of bias was assessed and a meta-analysis performed. Evidence quality was appraised using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology.

Findings: Eight studies (all cohort design) were identified as eligible. Studies were performed in six countries in Europe, Asia, and North America and comprised a total of 4829 participants. All studies were at high risk of bias. Pooled sensitivity of body surface screening to predict late-onset sepsis was 41% (95% confidence interval: 17-70), whereas pooled specificity was 56% (34-76) (hierarchical summary receiver operating characteristics (HSROC) model). Subgroup analyses showed higher pooled estimates for specificity but not sensitivity when screening focused on Escherichia coli or Klebsiella pneumoniae. GRADE evidence quality was very low.

Conclusion: Limited evidence of very low quality exists regarding the prognostic value of neonatal screening for late-onset sepsis. Carefully planned and conducted prospective studies, including randomized trials, are needed to clarify the potential value of this measure for the prediction and prevention of late-onset sepsis.