| Objective:It remains uncertain whether smoking status can effect efficacy of immune checkpoint inhibitors in metastatic non-small cell lung cancer.We performed a meta-analysis to address this issue.Methods:Pub Med,Embase,Cochrane Library,Web of Science,and international meetings were searched until April 1,2021,for phase 2 and 3 randomized controlled trials which compared ICIs with chemotherapy and reported overall survival and/or progression-free survival data according to smoking status.The random-effect model was used for statistical analysis.Results:Twenty-eight articles from 24 RCTs including 13918 patients were eligible.ICIs significantly prolonged OS than CT in smokers(HR=0.75,95%CI:0.69-0.81),but not in never-smokers(HR=0.87,95%CI:0.74-1.04);while there was no significant treatment-smoking interaction(P=0.11).Significant heterogeneity was observed for both smokers(OS:I~2=60%,P=0.0002;PFS:I~2=74%,P<0.0001)and never smokers(PFS:I~2=69%,P<0.0001).Subgroup analyses revealed that ICIs monotherapy significantly improved OS in smokers(HR=0.76,95%CI:0.69-0.85)but not in never-smokers(HR=0.93,95%CI:0.77-1.12,P=0.07),and treatment-smoking interaction was significant in patients with PD-L1≥50%(HR,0.61 vs 1.18;P=0.005).ICIs plus C T achieved better OS either in smokers or never-smokers(HR,0.76 vs 0.61;P=0.39),while dual ICIs combination prolonged OS only in smokers but never-smokers(HR,0.68 vs 1.02;P=0.02).Conclusion:Either ICIs monotherapy or combination therapy was superior to CT in smokers.While ICIs monotherapy and dual ICIs combination were less effective in never-smokers,and ICIs plus CT might be the optimal selection. |