Validation of a clinical score for the diagnosis of late onset neonatal septicemia in babies weighing 1000-2500 g.

There is paucity of data about the predictive values and likelihood ratios of clinical signs of late onset nosocomial sepsis in neonates. A clinical score comprising of seven items had been derived from analysis of individual signs and had been published by this group in the Journal of Tropical Pediatrics in 2003. The current study was done to validate the score in a fresh validation cohort, to evaluate the score at 0 and 24 h after onset of clinical signs of sepsis and to evaluate the sepsis screen in combination with the clinical score. The seven clinical signs in the clinical score included grunting, abdominal distension, increased prefeed aspirates, tachycardia, hyperthermia, chest retractions and lethargy. A total of 220 episodes of sepsis among 208 babies were evaluated. The clinical score was calculated at 0 h and 24 h. A sepsis screen (micro erythrocyte sedimentation rate, C reactive protein, absolute neutrophil count and immature/total neutrophil ratio) and blood culture were performed in all subjects at enrollment. Sepsis screen was considered 'positive' if any two parameters were positive. The outcome of interest was 'definite sepsis', defined as blood culture positive. The 0-h clinical score had sensitivity, specificity, PPV, NPV, LR(+) and LR(-) of 90, 22.5, 30.3, 85.7, 1.16 and 0.44%, respectively. The 24-h score had higher specificity (60.6%) but lower sensitivity than the 0-h score. Sepsis screen per se had a sensitivity and NPV of 48.3 and 78.3% but when combined with the 0-h clinical score, the sensitivity and NPV rose to 95 and 90.6%, respectively. The 'clinical score' in combination with sepsis screen result can be used by clinicians to rule out sepsis.