Ten prospective cohort studies (n=627, range 33 to 100) were included in the review. Blinding of study investigators for interpretation of both index test and references standard was unclear in most studies. Full details of quality assessment results were reported online.
Nine studies assessed a procalcitonin cut-off value of 0.5ng/mL and one study assessed a cut-off of 0.6ng/mL. At these cut-off values, sensitivity ranged from 0.59 (95% CI 0.41 to 0.75) to 1.00 (95% CI 0.82 to 1.00) and specificity ranged from 0.23 (95% CI 0.05 to 0.54) to 0.90 (95% CI 0.77 to 0.97).
The pooled estimate of DOR was 14.25 (95% CI 4.7 to 43.23). Significant between-study heterogeneity was noted (Χ2 test, p < 0.001 and I2 test 80%). When the two studies that reported a DOR with confidence intervals that included values of 1 (which indicated no significant diagnostic utility) were excluded from the analysis, the Χ2 test showed no significant heterogeneity and the I2 value was 36%; the pooled estimate of DOR excluding these two studies was 26.73 (95% CI 10.29 to 69.39).
Sensitivity analysis indicated that for studies that enrolled only children with febrile urinary tract infection, the DOR was 9.56 (95% CI 3.01 to 30.39).