Diagnostic accuracy of periostin in predicting asthma: a systematic review and Meta-analysis

Objective: We sought to evaluate the accuracy of periostin for the diagnosis of asthma through a systematic review and meta-analysis. Methods: we searched the electronic databases (PubMed, EMBASE, Web of science) to find relevant studies through December 31, 2018. Studies selecting, data extraction and quality assessment were conducted by two researchers independently according to inclusion and exclusion criteria. We used bivariate mixed effects model to conduct the meta-analysis. Results: A total of 9 studies including 1448 subjects with asthma were analyzed. The pooled sensitivity, specificity, diagnostic odds ratio (DOR) were 0.68 (95% CI, 0.42-0.86), 0.87 (95% CI, 0.76-0.93) and 13.78 (95% CI, 4.64-40.93) respectively. The area under the summary receiver operating curves (SROC) was 0.87 (95% CI, 0.84-0.90). The Spearman correlation coefficient indicated that no threshold effect existed in the included studies. The study design had statistical significance for sensitivity results (p < 0.001) through regression analysis. Heterogeneity might be caused by non-threshold effect. Conclusion: Periostin has moderate accuracy in the diagnosis of asthma.