[Accuracy of the procalcitonin test in the diagnosis of occult bacteremia in paediatrics: a systematic review and meta-analysis]. / Exactitud del test de procalcitonina en el diagnóstico de bacteriemia oculta en pediatría: revisión sistemática y metanálisis.

OBJECTIVE: To evaluate the diagnostic accuracy of serum procalcitonin (PCT) to detect severe bacterial infection (SBI) in ambulatory children attended in the emergency room (ER) for fever without source (FWS). MATERIAL AND METHODS: A search was made in MEDLINE, OVID and EMBASE (to January 2010). We searched for papers that evaluated the diagnostic accuracy of serum PCT to detect SBI in children that, being previously well, were seen in the ER for FWS. We rated the methodological quality of each paper using objective validity criteria (QUADAS, CASPE) and included only those with the maximum quality in the analysis. The statistical meta-analysis was performed using the software, Meta-DiSc 1.1.1 for Windows. RESULTS: The search identified 115 papers. Only 6 studies (prospective observational and analytic cohorts) fitted the inclusion criteria, with a sample size of 1139 patients. The prevalence of SBI was between 12.8% and 29% with a weighted mean of 18%. The overall senstivity was 0.771 (95% CI=0.707-0.826), the overall specificity was 0.804 (95% CI=0.777-0.830), the overall positive likelihood ratio was 3.610 (95% CI=2.481-5.253) and the overall negative likelihood ratio was 0.218 (95% CI=0.106-0.446). The diagnostic OR was 18.922 (95% CI=10.076-35.534), the Area under the SROC curve was 0.8801 (95% CI=0.821-0.939), and the optimal diagnostic cut-off value was Q*=0.8106 (95% CI=0.7512-0.8699). CONCLUSIONS: On the basis of our analysis, in children with FWS seen in the ER, the serum PCT test accurately identifies those that have a SBI. We cannot extrapolate these results to other types of patients.