Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada.
Center for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Canada.
Clin Pharmacol Ther. 2021 May;109(5):1219-1223. doi: 10.1002/cpt.2213. Epub 2021 Mar 16.
The US Food and Drug Administration is considering replacing cardiovascular outcome trials of antidiabetic drugs with trials that better represent patients with type 2 diabetes. However, designing such representative trials requires understanding the underlying target populations (i.e., populations intended to receive the drug in the real-world setting). Thus, we used the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial as a motivating example to illustrate how different target populations impact trial representativeness. Using the United Kingdom Clinical Practice Research Datalink, we identified three target populations: (i) all patients with type 2 diabetes; (ii) patients prescribed liraglutide; and (iii) patients who would have been eligible to receive liraglutide based on treatment stage (i.e., patients with poorly controlled diabetes eligible to receive a second-to-fifth line antidiabetic drug). We then examined the representativeness of the LEADER trial by applying its eligibility criteria to each target population. The target populations of patients with type 2 diabetes (n = 279,763), those prescribed liraglutide (n = 14,421), and those eligible to receive liraglutide based on the treatment stage (n = 85,610) differed substantially in terms of hemoglobin A1c, body mass index, prevalence of heart failure, and chronic kidney disease. Applying the LEADER trial eligibility criteria to these target populations resulted in the inclusion of 19.1%, 20.7%, and 34.8% patients, respectively. This study highlights how real-world data can be used to define different target populations. Explicitly defining these target populations can help in the design of future trials of antidiabetic drugs.
美国食品和药物管理局正在考虑用更能代表 2 型糖尿病患者的试验来替代抗糖尿病药物的心血管结局试验。然而，设计这样具有代表性的试验需要了解潜在的目标人群（即，打算在实际环境中使用该药物的人群）。因此，我们以利拉鲁肽效应和行动在糖尿病：心血管结局评价（LEADER）试验为例，说明了不同的目标人群如何影响试验的代表性。我们使用英国临床实践研究数据链接，确定了三个目标人群：（i）所有 2 型糖尿病患者；（ii）接受利拉鲁肽治疗的患者；（iii）根据治疗阶段有资格接受利拉鲁肽的患者（即，血糖控制不佳、有资格接受二线至五线抗糖尿病药物的患者）。然后，我们通过将 LEADER 试验的入选标准应用于每个目标人群，考察了 LEADER 试验的代表性。2 型糖尿病患者（n=279763）、接受利拉鲁肽治疗的患者（n=14421）和根据治疗阶段有资格接受利拉鲁肽的患者（n=85610）在糖化血红蛋白、体重指数、心力衰竭和慢性肾病的患病率方面存在显著差异。将 LEADER 试验的入选标准应用于这些目标人群，分别纳入了 19.1%、20.7%和 34.8%的患者。这项研究强调了如何使用真实世界的数据来定义不同的目标人群。明确界定这些目标人群有助于设计未来的抗糖尿病药物试验。
