Department of Equity, Ethics and Policy, McGill University, Montreal, Quebec, Canada.
Department of Medical Oncology, Northern Ontario School of Medicine University, Thunder Bay, Ontario, Canada.
J Clin Epidemiol. 2023 Jun;158:134-140. doi: 10.1016/j.jclinepi.2023.03.028. Epub 2023 Apr 5.
Drug developers sometimes launch phase 3 (P3) trials without supporting evidence from phase 2 (P2) trials. We call this practice "P2 bypass." The aims of this study were to estimate the prevalence of P2 bypass and to compare the safety and efficacy results for P3 trials that bypassed with those that did not.
We created a sample of P3 solid tumor trials registered on ClinicalTrials.gov with primary completion dates between 2013 and 2019. We then attempted to match each with a supporting P2 trial using strict and broad criteria. P3 outcomes were meta-analyzed using a random effects model with subgroup contrast between trials that bypassed and those that did not.
129 P3 trial arms met eligibility and nearly half involved P2 bypass. P3 trials involving P2 bypass produced significantly and nonsignificantly worse pooled efficacy estimates using broad and strict matching criteria, respectively. We did not observe significant differences in safety outcomes between P3 trials that bypassed P2 and those that did not.
The risk/benefit balance of P3 trials that bypassed P2 is less favourable than for trials supported by P2.
药物开发商有时在没有 2 期（P2）试验支持证据的情况下启动 3 期（P3）试验。我们将这种做法称为“P2 回避”。本研究的目的是估计 P2 回避的发生率，并比较回避和不回避 P2 的 P3 试验的安全性和疗效结果。
我们创建了一个样本，包括 2013 年至 2019 年期间在 ClinicalTrials.gov 上注册的主要完成日期为 P3 固体肿瘤试验。然后，我们试图使用严格和广泛的标准，将每个试验与支持性 P2 试验相匹配。使用随机效应模型对 P3 结局进行荟萃分析，并在回避和不回避 P2 的试验之间进行亚组对比。
129 个 P3 试验臂符合入选标准，近一半涉及 P2 回避。使用广泛和严格的匹配标准，分别观察到涉及 P2 回避的 P3 试验的疗效综合估计值显著和不显著更差。我们没有观察到回避和不回避 P2 的 P3 试验之间的安全性结果有显著差异。
回避 P2 的 P3 试验的风险/效益平衡不如有 P2 支持的试验有利。
