Comparison of the sensitivity and specificity of bronchiolitis severity scores in infants: a systematic review and meta-analysis

Abstract

Background: Respiratory syncytial virus (RSV) represents the most common cause of acute respiratory hospitalization in children <24 months of age. Several clinical scores have been proposed to estimate disease severity.

Objective: This study aims to compare the sensitivity and specificity of clinical scores assessing the severity of bronchiolitis in children <2 years old.

Methods: PubMed, Scopus and Web of Science databases were approached to identify all studies published before 31^st December 2024 using bronchiolitis severity scores and providing data regarding their sensitivity and specificity.

Results: At the end of the selection process, 18 studies evaluating 13 clinical severity scores were analysed. A total of 6552 children (weighted mean age 5.1 months - range 0-24 months) with a diagnosis of bronchiolitis were included. The summary ROC analysis demonstrated that the GRSS (Global Respiratory Severity Scale) performed better than other scores in discriminating children at risk for severe disease. In particular, the GRSS showed cumulative sensitivity and specificity values of 0.87 (95 %CI: 0.80-0.92) and 0.92 (95 %CI: 0.88-0.95) respectively, with elevated accuracy (0.90).

Limits: We limited our analysis only to studies reporting specificity and sensitivity values, possibly excluding other validated scores, but we aimed to perform a balanced analysis involving the most possible homogenous population.

Conclusions: There is limited but moderate-to-adequate evidence that the GRSS score has better sensitivity and specificity for clinically assessing the severity of bronchiolitis in infants and children <2 years old. However, further studies are needed to validate these results, ideally using larger datasets.

Keywords: Bronchiolitis; Children; GRSS; Infant; Meta-analysis; RSV; Severity score.