文献速递流行病学与生物统计临床试验与 GCP

良性前列腺增生试验结局变更追踪:ClinicalTrials.gov 版本历史综述

Tracking outcome changes in benign prostatic hyperplasia trials: a version history review of ClinicalTrials.gov records.
PubMed 文献 · Trials · 2026-07-16

该研究基于 ClinicalTrials.gov 对 2017 至 2024 年间注册的良性前列腺增生干预性试验进行横断面分析,纳入 36 项符合条件且已发表结果的试验。全部试验均存在至少一次结局变更且均未说明理由,变更仅发生在主要完成之后,次要结局变更更频繁。

要点

  • 全部 36 项试验均存在结局变更,无任何变更在注册库或论文中说明理由。
  • 变更仅发生于主要完成之后,约半数在发表前、半数在发表后。
  • 次要结局变更频率显著高于主要结局(p < 0.01)。

核心结论

  • 良性前列腺增生试验的结局变更普遍存在,且均未在注册库或发表论文中说明理由。
  • 变更仅发生在主要完成之后,可能影响试验结果的可解释性与透明度。

研究设计与样本

  • 纳入标准包括入组成年男性、已发布结果、原始与最新版本均含预设结局、且与同行评议论文关联。
  • 筛查 1383 条记录,最终纳入 36 项试验。
  • 变更按类型、时机和严重程度分类,采用 Cochran's Q 检验、Kaplan-Meier 分析和相关性检验。

主要发现

  • 全部 36 项试验(100%)至少存在一次结局变更,且均未说明理由。
  • 变更仅发生在主要完成之后,约半数在发表前、半数在发表后。
  • 轻微措辞修改和从宽泛到具体的定义调整最常见。
  • 次要结局的变更频率显著高于主要结局(p < 0.01)。
  • 结局负荷指数与变更频率呈弱相关(r = 0.17)。

关键数值与推荐

  • 纳入 36 项试验,均满足入组标准且已发表同行评议论文。
  • 100% 试验存在至少一次未说明理由的结局变更。
  • 次要结局变更频率显著更高(p < 0.01)。
  • 结局负荷指数与变更频率相关系数 r = 0.17。

临床怎么用

  • 解读良性前列腺增生试验结果时,应核查注册库版本历史以识别未说明的结局变更。
  • 研究者应在注册库和论文中主动披露结局变更及其理由。
  • 期刊和监管机构可加强对注册库版本历史的审查与披露标准。

局限性

  • 仅纳入已发表结果且与论文关联的试验,可能存在选择偏倚。
  • 样本量较小(36 项试验),结果外推需谨慎。
  • 未评估变更对结果解释的实际影响。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: Benign prostatic hyperplasia (BPH) is a common condition with trials that frequently rely on subjective outcomes. While federal regulations require prospective outcome registration in ClinicalTrials.gov, modifications to these outcomes are permitted and often unjustified, potentially compromising transparency and interpretability. This study examined the frequency, timing, type, and transparency of prespecified outcome modifications in BPH trials. METHODS: We conducted a registry-based cross-sectional analysis of interventional BPH trials registered on ClinicalTrials.gov between January 18, 2017, and December 31, 2024. Eligible trials enrolled adult males, reported posted results, contained prespecified outcomes in both original and most recent registry versions, and were linked to peer-reviewed publications. Modifications were categorized by type (e.g., additions, removals, rewordings), timing, and severity. Key variables included the Outcome Load Index (OLI) and a Severity-Weighted Change Score. Statistical analyses included Cochran's Q tests, Kaplan-Meier time-to-event analysis, and correlation tests. RESULTS: Of 1383 records screened, 36 trials met inclusion criteria. All trials (100%) exhibited at least one outcome modification, none of which were justified. Modifications occurred exclusively after primary completion; half occurred before publication and half after. Minor rewording and broad-to-specific definitional shifts were most common. Changes were significantly more frequent in secondary outcomes (p < 0.01). Severity scores varied widely but were not consistently associated with funder or intervention type. OLI showed weak correlation with modification frequency (r = 0.17). CONCLUSIONS: In this registry-based analysis of BPH clinical trials, outcome modifications were universal, occurred exclusively after primary completion, and were never justified in either the registry or corresponding publications. These findings highlight a persistent gap in transparency that could affect the interpretability of trial results. Greater oversight of registry version histories and stronger disclosure standards are needed to ensure the credibility of outcome reporting in BPH research and beyond. TRIAL REGISTRATION: Registered on the Open Science Framework. The study was additionally submitted to PROSPERO; the OSF registration is the protocol of record. Registered on July 17, 2025.

来源:Trials,查看原文。版权归原作者所有。

引用

  1. Livsey T, Itani M, Elghzali A, et al. Tracking outcome changes in benign prostatic hyperplasia trials: a version history review of ClinicalTrials.gov records. Trials. 2026. PMID: 42458564. doi:10.1186/s13063-026-09892-0

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。