Serum procalcitonin not very accurate in diagnosing children with pneumonia

Référence

Tsou PY, Rafael J, Ma YK, et al. Diagnostic accuracy of procalcitonin for bacterial pneumonia in children—a systematic review and meta-analysis. Infect Dis (Lond) 2020;52(10):683-697.

Sommaire

These authors searched PubMed and EMBASE to identify studies that evaluated the use of serum procalcitonin in diagnosing pneumonia in children. They don't describe efforts to identify unpublished studies. Two authors independently evaluated studies for inclusion and assessed methodologic quality. They allowed, singly or in combination, any of the following as reference standards, making calculations of test properties kind of fun: chest radiographs, blood culture, sputum culture, bronchoalveolar lavage, polymerase chain reaction of nasopharyngeal secretions, and antibody or antigen assays. The authors identified 25 studies with 2864 patients. Overall, the authors report the methodologic quality of the included studies as "acceptable." The studies used a variety of procalcitonin cutoffs (from 0.12 to 35.8 nanograms/mL) and the prevalence of pneumonia ranged from 7% to 82%. Depending on age, the cutoff, and type of pneumonia, the sensitivity ranged from 53% to 78% and the specificity ranged from 48% to 82%. Not all that impressive. Overall, the pooled sensitivity and specificity for procalcitonin was 64% (95% CI 53% - 74%) and 72% (64% - 79%), respectively. This translates to a positive likelihood ratio of 2.3 (0.8 - 3.0) and a negative likelihood ratio of 0.50 (0.38 - 0.66), The overall diagnostic accuracy was 74% (70% - 78%). Finally, the authors identified a high degree of heterogeneity among the data that was not explained by the aforementioned subgroups. Given the methodologic challenges, these estimates suggest that procalcitonin is not all that helpful. If you're interested in seeing how procalcitonin performed among various subgroups, Table 4 has tons of details for you!