Diagnostic and prognostic value of serial perfusion index in neonates with suspected sepsis: A prospective observational study

Abstract

Perfusion index (PI), a non-invasive pulse oximetry-derived measure of peripheral perfusion, has shown promise as a marker of illness severity in neonatal sepsis, but most prior studies have relied on single time-point measurements. This study evaluated serial PI as a predictor of culture-positive sepsis, clinical deterioration and mortality in neonates with suspected sepsis. In this prospective observational cohort study conducted in a tertiary care Neonatal Intensive Care Unit (April–July 2025), PI was measured at 0, 6 and 24 h from the time of clinical suspicion of early- or late-onset sepsis in 240 neonates. Diagnostic accuracy was assessed using receiver operating characteristic curve analysis, and independent predictors of deterioration and mortality were identified using multivariate logistic regression. Blood culture was positive in 54 neonates (22.5%). PI did not differ significantly between culture-positive and culture-negative neonates at any time point ( p > 0.19). Sixty neonates (25.0%) developed deterioration and 24 (10.0%) died. PI showed good to excellent accuracy for deterioration (area under the curve [AUC] 0.905–0.910) and mortality (AUC 0.914–0.917), with PI rising over 24 h in stable/surviving neonates and falling in those who deteriorated or died. On multivariate analysis, PI at 0 h remained independently associated with mortality (odds ratio 0.102, 95% confidence intervals 0.015–0.712, p = 0.021), while only PI at 24 h showed a borderline independent association with deterioration ( p = 0.079). Serial PI did not distinguish culture-positive from culture-negative sepsis but showed good discriminative accuracy for deterioration and mortality in neonates, supporting its potential as a non-invasive bedside adjunct for early risk stratification, pending larger multicentre validation.