McGill University, Biomedical Ethics Unit, Montreal, QC, Canada.
The Anticancer Fund, Brussels, Belgium.
PLoS One. 2022 Sep 12;17(9):e0274115. doi: 10.1371/journal.pone.0274115. eCollection 2022.
Once a drug gets FDA approved, researchers often attempt to discover new applications in different indications. The clinical impact of such post-approval activities is uncertain. We aimed to compare the clinical impact of research efforts started after approval with those started before for cancer drugs.
We used Drugs@FDA to perform a retrospective cohort study of secondary approvals for cancer drugs that were initially FDA approved between 2005 and 2017. Clinicaltrials.gov was used to identify the beginning of each research trajectory that resulted in a secondary FDA approval. Each trajectory was classified as pre- or post-approval depending on if it was initiated before or after initial drug licensure. Clinical impact was assessed by comparing secondary approvals and NCCN off-label recommendations deriving from pre- vs. post-approval trajectories, pooled effect sizes, incidence, and level of evidence.
We identified 77 broad secondary approvals, 60 of which had at least 6 years follow-up. Of these, 9 (15%) resulted from post-approval trajectories, a proportion that is significantly lower than would be expected if the timing of research didn't impact approval (McNemar's test p = 0.001). Compared to pre-approval trajectories, approvals resulting from post-approval trajectories were for cancers with lower mean incidence (6.11 vs 14.83, p = 0.006) and were based on pivotal trials with smaller pooled effect sizes (0.69 vs 0.57, p = 0.02) that were less likely to be randomized (38.5% vs 64.1%, p = 0.145). We identified 69 NCCN off-label recommendations. The proportion stemming from post-approval trajectories was similar to that for pre-approval (56.5% vs. 43.5%). However, recommendations from post-approval trajectories were significantly more likely to involve rare diseases (76.7% vs 51.4%, p = 0.019) and nonsignificantly less likely to be based on level 1 evidence (11.6% vs 22.9%, p = 0.309).
Secondary FDA approvals are less likely to result from post-approval trajectories and tend to be less impactful compared to approvals originating from research started before first FDA licensure. However, post-approval trajectories may be as likely to lead to NCCN recommendations for off-label use. Limitations of this work include our use of indirect measures of impact and limited follow-up time for trajectories. Our study protocol was pre-registered (https://osf.io/5g3jw/).
一旦药物获得 FDA 批准，研究人员通常会尝试在不同的适应症中发现新的应用。这些批准后的活动的临床影响是不确定的。我们旨在比较批准后和批准前开始的癌症药物研究的临床影响。
我们使用 Drugs@FDA 对 2005 年至 2017 年间首次获得 FDA 批准的癌症药物的二次批准进行了回顾性队列研究。Clinicaltrials.gov 用于确定导致二次 FDA 批准的每个研究轨迹的开始。根据其是否在初始药物许可之前或之后开始，每个轨迹都被分类为批准前或批准后。通过比较批准前和批准后轨迹的二级批准和 NCCN 标签外推荐、汇总效果大小、发生率和证据水平来评估临床影响。
我们确定了 77 种广泛的二级批准，其中 60 种有至少 6 年的随访。其中，9 项（15%）来自批准后轨迹，这一比例明显低于如果研究时间不影响批准，则预期的比例（McNemar 检验，p=0.001）。与批准前轨迹相比，来自批准后轨迹的批准适用于发病率较低的癌症（6.11 与 14.83，p=0.006），并且基于汇总效果较小的关键性试验（0.69 与 0.57，p=0.02），这些试验更不可能是随机的（38.5%与 64.1%，p=0.145）。我们确定了 69 项 NCCN 标签外推荐。来自批准后轨迹的比例与批准前相似（56.5%比 43.5%）。然而，来自批准后轨迹的建议更有可能涉及罕见疾病（76.7%比 51.4%，p=0.019），不太可能基于 1 级证据（11.6%比 22.9%，p=0.309）。
与来自首次 FDA 许可前开始的研究的批准相比，来自批准后轨迹的二次 FDA 批准不太可能发生，并且往往影响较小。然而，批准后轨迹可能同样有可能导致 NCCN 建议进行标签外使用。本研究的局限性包括我们使用间接措施来衡量影响以及对轨迹的随访时间有限。我们的研究方案已预先注册（https://osf.io/5g3jw/）。
