3.Results
3.1Study searches and characteristics of the included studies
The detailed process of the literature search is shown in Fig. 1. A total of 6 eligible studies[15-20], published between 2016 and 2020, were enrolled in the present meta-analysis. The studies were conducted in the following geographical regions:USA(n=2),Germany(n=1), Turkey(n=1), China(n=2). All studies used well-controlled asthma as the control group, and 4 studies also added healthy people as the control group[15, 17, 18, 20]. Our study focused on the role of IOS in predicting asthma exacerbations or loss of control, so we did not include healthy people. The sample size ranged from 54 to 192, including 615 children with an average age of 3.65-11.20 years. They were all case-control studies. All participants performed IOS. Meaningful pulse oscillation parameters varied among the trials. The quality scores of these studies were evaluated by the NOS ranging from 7 to 8. The characteristics of all studies were summarized in Table1.
3.2Association between pulse oscillation parameters and asthma exacerbations
The included six studies revealed that IOS may be able to identify the risk of acute attack or loss of control of asthma. We further analyzed the value of IOS common parameters in predicting the loss of asthma control. In terms of resistance, we found that elevated levels of R5 may indicate the risk of asthma exacerbations (WMD = − 1.21, 95%CI: −1.55 to −0.87, P<0.001)(Fig.2a).However, as for R20(WMD = − 0.81, 95%CI: −1.95 to 0.32, P=0.161), R5-20 (WMD = − 0.68, 95%CI: −1.61 to 0.25, P=0.151), the results did not show a good correlation with the acute attack of asthma (Fig.2b-2c). In terms of reactance, we found X5 (WMD =0.81, 95%CI: 0.56 to 1.01, P<0.001) were significantly related to the loss of asthma control(Fig.2d). Moreover, the Fres (WMD = − 1.34, 95%CI: −2.03 to −0.65,P=0.018) and AX (WMD = − 7.35, 95%CI: −9.94 to −4.76,P<0.001) can also identify the risk of acute attack or loss of control of asthma(Fig.2e-2f).
3.3 Sensitivity analysis and publication bias
Sensitivity analysis confirmed that no individual study influenced the overall results. There was no evidence of publication bias in this meta-analysis indicated by Begg’s funnel plot and Egger’s tests (Fig3).