Department of Internal Medicine, Division of Cardiology, Pulmonology and Vascular Medicine, Medical Faculty, Heinrich Heine University Düsseldorf, Moorenstr. 5, 40225, Düsseldorf, Germany.
German Center for Diabetes Research, Partner Düsseldorf, München-Neuherberg, Germany.
ESC Heart Fail. 2023 Feb;10(1):552-567. doi: 10.1002/ehf2.14213. Epub 2022 Nov 7.
Absolute treatment effects-i.e. numbers needed to treat (NNTs)-of novel antidiabetic drugs for cardiovascular outcomes have not been comprehensively evaluated. We aimed to perform a meta-analysis of digitalized individual patient outcomes to display and compare absolute treatment effects.
Individual patient time-to-event information from Kaplan-Meier plots of cardiovascular mortality (CM) and/or hospitalization for heart failure (HHF) endpoints from cardiovascular outcome trials (CVOTs) evaluating dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, and sodium glucose transporter 2 (SGLT2) inhibitors vs. placebo were digitalized using WebPlotDigitizer 4.2 and the R code of Guyot et al.; Weibull regression models were generated, validated, and used to estimate NNT for individual trials; random-effects meta-analysis generated Meta-NNT with 95% confidence intervals. Sixteen CVOTs reported time-to-event information (14 in primary diabetes and 2 in primary heart failure populations). Thirteen studies including 96 860 patients were meta-analysed for CM: At the median follow-up of 30 months, Meta-NNTs were 178 (64 to ∞ to -223) for DPP-4 inhibitors, 261 (158 to 745) for GLP-1 receptor agonists, and 118 (68 to 435) for SGLT2 inhibitors. Ten studies including 96 128 patients were meta-analysed for HHF: At the median follow-up of 29 months, estimated Meta-NNTs were -644 (229 to ∞ to -134) for DPP-4 inhibitors, 441 (184 to ∞ to -1100) for GLP-1 receptor agonists, and 126 (91 to 208) for SGLT2 inhibitors. SGLT2 inhibitors were especially effective for HHF in primary heart failure populations [Meta-NNT 25 (19 to 39)] vs. primary diabetes populations [Meta-NNT 233 (167 to 385)] at 16 months of follow-up.
We found only modest treatment benefits of GLP-1 receptor agonists and SGLT2 inhibitors for CM and HHF in primary type 2 diabetes mellitus populations. In primary heart failure populations, SGLT2 inhibitor benefits were substantial and comparable in efficacy to established heart failure medication.
尚未全面评估新型抗糖尿病药物对心血管结局的绝对治疗效果，即需要治疗的人数（NNT）。我们旨在对数字化个体患者结局进行荟萃分析，以显示和比较绝对治疗效果。
使用 WebPlotDigitizer 4.2 和 Guyot 等人的 R 代码对来自心血管结局试验（CVOT）的评估二肽基肽酶-4（DPP-4）抑制剂、胰高血糖素样肽-1（GLP-1）受体激动剂和钠-葡萄糖共转运蛋白 2（SGLT2）抑制剂与安慰剂的心血管死亡率（CM）和/或心力衰竭住院（HHF）终点的 Kaplan-Meier 图中的个体患者时间至事件信息进行数字化。生成、验证了威布尔回归模型，并用于估计个体试验的 NNT；随机效应荟萃分析生成 Meta-NNT 及其 95%置信区间。16 项 CVOT 报告了时间至事件信息（14 项在原发性糖尿病人群中，2 项在原发性心力衰竭人群中）。对包括 96860 例患者的 13 项研究进行了荟萃分析以评估 CM：在中位随访 30 个月时，Meta-NNT 分别为 DPP-4 抑制剂 178（64 至 ∞ 至 -223）、GLP-1 受体激动剂 261（158 至 745）和 SGLT2 抑制剂 118（68 至 435）。对包括 96128 例患者的 10 项研究进行了荟萃分析以评估 HHF：在中位随访 29 个月时，Meta-NNT 分别为 DPP-4 抑制剂 -644（229 至 ∞ 至 -134）、GLP-1 受体激动剂 441（184 至 ∞ 至 -1100）和 SGLT2 抑制剂 126（91 至 208）。SGLT2 抑制剂在原发性心力衰竭人群中对 HHF 特别有效[Meta-NNT 25（19 至 39）]，而在原发性糖尿病人群中则效果不佳[Meta-NNT 233（167 至 385）]，中位随访 16 个月。
我们仅发现 GLP-1 受体激动剂和 SGLT2 抑制剂对原发性 2 型糖尿病患者的 CM 和 HHF 有适度的治疗益处。在原发性心力衰竭人群中，SGLT2 抑制剂的获益是实质性的，与已确立的心力衰竭药物相当。
