Diagnostic value of a new procalcitonin cord sample-guided algorithm to manage newborns suspected of early-onset infection

Diagnosis of early neonatal infection remains an emergency. Since clinical symptoms and biological markers are neither sensitive nor specific, many newborns suspected of infection undergo biological analysis and empirical antibiotic treatment while awaiting results. Recent studies underline the benefit of using procalcitonin (PCT) to differentiate inflammatory diseases and viral infections from bacterial infections. Joram shows that it is possible to go beyond the physiological peak of PCT in the first days of life by measuring PCT concentration in cord blood. The aim of this prospective study was to evaluate a new algorithm integrating the value of PCT in blood cord for taking care of newborns who have suspected infection. The diagnostic value of the new algorithm was compared to the diagnostic value of the algorithm currently in use, by analyzing a 9-month prospective database of 1267 newborns suspected of infection. Infection status was established with the ANAES definition and clinical progression. Each infected newborn (n=8) would have been treated without delay with the current algorithm (based on ANAES guidelines) and this new algorithm. The new algorithm had the same diagnostic value as the current algorithm (P=0.5) with 87.5% sensitivity (95%CI [52-98]) versus 100% (95%CI [87-100]) and 87.4% specificity (95%CI [85-90]) versus 83.8% (95%CI [81-86]). Fewer biological analyses 13.1% (95%CI [11-16]) versus 42.2% (95%CI [39-45]) were performed with the PCT cord-guided algorithm than with the current algorithm (P