基于SMAA模型的肠促胰素类降糖药的获益/风险评价

Benefit-risk Assessment of Incretin-based Antidiabetic Treatment Based on SMAA Model

目的:选择随机多准则可接受性分析(SMAA)模型对肠促胰素类降糖药中的二肽基肽酶-4抑制药(DPP-4I)和胰高血糖素样肽-1受体激动药(GLP-1RA)进行获益/风险评价(BRA)。方法:确立评价指标,同时系统检索Medline、Embase、the Cochrane Library、Clinical Trials数据库,搜集在2型糖尿病(T2DM)患者中开展的DPP-4I与GLP-1RA相互比较、DPP-4I或GLP-1RA与其他降糖药比较且至少报告一项本研究所关注结局指标的随机对照试验(RCT),检索时限均从建库至2019年3月29日,采用SMAA模型和网状Meta分析(NMA)对DPP-4I和GLP-1RA相互之间或各自与其他降糖药之间进行BRA,并进行敏感性分析。结果:确定了与肠促胰素类药物治疗相关的26项BRA指标,最后BRA结果提示GLP1-RA在此评价体系内优于DPP-4I的概率为84.5%。此外,DPP-4I相较于胰岛素、二甲双胍、钠-葡萄糖共转运蛋白-2抑制药(SGLT-2I)、磺脲类、噻唑烷二酮类(TZD)和α-糖苷酶抑制药(AGI),其BRA的可接受度分别为93.1%,38.5%,6.5%,80.4%,41.4%和72.1%;GLP-1RA相较于胰岛素、二甲双胍、SGLT-2I、磺脲类、TZD与AGI,其BRA的可接受度分别为90.6%,70.4%,24.0%,94.2%,69.8%和89.4%。结论:GLP-1RA在此评价体系内有较大概率优于DPP-4I,DPP-4I优于胰岛素与磺脲类,劣于二甲双胍和SGLT-2I的可接受度较高,GLP-1RA优于胰岛素、磺脲类、TZD与AGI,劣于SGLT-2I的可接受度较高,敏感性分析前后结论一致,较为稳健。但由于研究间异质性、药物的分类方式、结局指标的选择以及偏倚风险的存在,可能均会对NMA和SMAA结果的准确性造成一定的影响,因此下结论时需谨慎;此外,进行决策时,还应考虑药物定价。

Objective:To conduct benefit-risk assessment of incretin-based therapies including dipeptidyl peptidase-4 inhibitor(DPP-4 I)and glucagon-like peptide-1 receptor agonists(GLP-1 RA)by using stochastic multi-criteria acceptability analysis(SMAA)for the treatment of type 2 diabetes mellitus(T2 DM).Methods:Evaluation criteria were determined and then we conducted a systematic review of English database such as Medline,Embase,the Cochrane Library and ClinicalTrials.gov from inception to March 29,2019,to identify randomized controlled trials(RCTs)comparing DPP-4 I with GLP-1 RA,or DPP-4 I,GLP-1 RA with other antidiabetic drugs in patients with T2 DM.At the same time,these studies must report at least one of the indicators that we had identified.Next,SMAA was conducted and sensitivity analysis was performed on the results.Results:26 evaluation criteria were finally determined.The results of SMAA showed that the acceptability of GLP-1 RA over DPP-4 I in this evaluation system was 84.5%.In addition,the rank acceptability indexes of BRA of DPP-4 I was 93.1%,38.5%,6.5%,80.4%,41.4%and 72.1%compared to insulin,metformin,sodium-glucose co-transporter 2 inhibitor(SGLT-2 I),sulfonylurea,thiazolidinedione(TZD)and alpha-glucosidase inhibitor(AGI);compared with insulin,metformin,SGLT-2 I,sulfonylurea,TZD and AGI,the rank acceptability of BRA of GLP-1 RA was 90.6%,70.4%,24.0%,94.2%,69.8%and 89.4%.Conclusion:GLP-1 RA was superior to DPP-4 I with a higher probability in this evaluation environment.Besides,the results of SMAA analysis indicated that the rank acceptability indexed of DPP-4 I was higher than insulin and sulfonylurea,but lower than metformin and SGLT-2 I.Meanwhile,GLP-1 RA was better than insulin,sulfonylurea,TZD and AGI.SGLT-2 I had a better performance than GLP-1 RA.Sensitivity analysis indicated the results were robust.But due to the existence of heterogeneity,drug classification and the influence of bias,the accuracy of the results may be affected.In addition,drug pricing should be considered when making decisions.