The incidence of LNS in each score cluster is presented in Table 4. In design of our scale, we had defined a score of 3 as a cutoff that discriminate between neonates with 'low risk' of LNS and 'medium' to 'high risk' of LNS. Application of this score to neonates in the current study, however, misclassified one neonate with LNS with a score of 2; all other neonates with LNS (n=34) had a clinical score of 4 or higher. The only neonate who was misclassified had a CONS septicemia. This was a 30-week, 14-day neonate without a CVC. The recorded indication for performing of sepsis evaluation in this neonate was an abnormal white blood cell count from a 'routine' CBC check. Despite absence of clinical signs, this neonate was treated with vancomycin for 10 days. No quantitative or repeat blood culture was performed on this neonate. We have tried to modify the LNS score by incorporating utilization of all CVCs instead of only UVC, but the performance of the modified LNS score was poorer with AUC of 0.73.