The comparative effectiveness and safety of biologic therapies for uncontrolled severe asthma: a systematic review and network meta-analysis

BackgroundTrials have not directly compared biologics for the treatment of asthma. ObjectiveTo comparative the relative efficacy of biologics in asthma. MethodsWe searched MEDLINE, EMBASE, CENTRAL, and clinicaltrials.gov from inception to May 31, 2022, for randomized trials addressing biologic therapies for asthma. Reviewers worked independently and in duplicate to screen references, extract data, and assess risk of bias. We performed a frequentist network meta-analysis and assessed the certainty of evidence using the GRADE approach. We present dichotomous outcomes as absolute risk differences per 1000 patients and relative risk (RR) with 95% confidence intervals (95% CI) and continuous outcomes as mean difference (MD) and 95% CI. ResultsWe identified 64 trials, including 26,630 patients. For patients with eosinophilic asthma, tezepelumab (329 fewer exacerbations per 1000 [95% CI 272.6 to 366.6 fewer]) and dupilumab (319.6 fewer exacerbations per 1000 [95% CI 272.6 to 357.2 fewer]) reduce exacerbations compared to placebo (high certainty). Tezepelumab (MD 0.24 L [95% CI 0.16 to 0.32]) and dupilumab (0.25 L (95% CI 0.21 to 0.29) improve lung function (FEV1) compared to placebo (high certainty). Both tezepelumab (110.97 fewer hospital admissions per 1000 (95% CI 94.53 to 120.56 fewer) and dupilumab (97.27 fewer hospitalizations [4.11 to 124.67 fewer]) probably reduce hospital admissions compared to placebo (moderate certainty). For patients with low eosinophils, biologics probably do not improve asthma outcomes. For these patients, tezepelumab (MD 0.1 L [95% CI 0 to 0.19]) and dupilumab (MD 0.1 L [95% CI 0 to 0.20)] may improve lung function (low certainty). ConclusionTezepelumab and dupilumab are effective at reducing exacerbations. For patients with low eosinophils, however, clinicians should probably be more judicious in use of biologics, including tezepelumab since they probably do not confer substantial benefit.