SGLT2抑制剂试验安全性报告:注册库与期刊发表数据不一致性分析
本研究纳入176项在ClinicalTrials.gov注册且配有全文发表的SGLT2抑制剂临床试验,比较注册库与期刊论文中严重不良事件(SAE)和全因死亡率数据的一致性。结果显示40%试验的SAE患者数、44%试验的SAE总数、34%试验的全因死亡率存在差异,注册库报告的SAE患者数和SAE总数均高于发表论文,提示安全性报告透明度不足。
要点
- 40%试验的SAE患者数在注册库与论文间不一致,注册库多报5185例。
- 44%试验的SAE总数不一致,注册库多报16012例。
- 34%试验的全因死亡率不一致,论文中21%试验少报死亡、10%多报死亡。
核心结论
- SGLT2抑制剂试验的死亡与SAE数据在注册库和期刊论文间存在明显差异,需提升报告一致性与透明度。
研究设计与样本
- 观察性横断面研究,纳入176项在ClinicalTrials.gov注册且配有全文发表的SGLT2抑制剂临床试验。
- 从注册库和对应论文提取SAE患者数、SAE总数、全因死亡率和死亡数,分析一致性。
主要发现
- 70项(40%)试验的SAE患者数不一致,注册库报告33303例 vs 论文28118例(差5185;χ2=22808,P<.001)。
- 77项(44%)试验的SAE总数不一致,注册库44163例 vs 论文28151例(差16012;χ2=196897,P<.001)。
- 27项(34%)试验的全因死亡率不一致;论文中17项(21%)少报死亡共4106例,8项(10%)多报死亡共43例。
- 产业资助试验在注册库与论文中全因死亡率报告相同的可能性显著更高(χ2=22.72,P<.001)。
- 达格列净试验的发表报告差异率最高。
临床怎么用
- 解读SGLT2抑制剂安全性证据时应同时核查注册库与发表论文,警惕SAE和死亡数据的不一致。
- 产业资助试验的全因死亡率报告一致性较高,但其他SAE指标仍需交叉验证。
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 AND OBJECTIVES: Sodium-glucose cotransporter 2 (SGLT2) inhibitors have gained prominence over the last decade due to their multifaceted benefits. However, systematic reviews and meta-analyses reveal conflicting results regarding their adverse event profiles compared to placebo. This study aimed to compare safety data from clinical trials on SGLT2 inhibitors reported in ClinicalTrials.gov with data published in scholarly journals. METHODS: In this observational, cross-sectional study, we included clinical trials on SGLT2 inhibitors registered at ClinicalTrials.gov with a matching publication in a scholarly journal with full text available. From ClinicalTrials.gov and the corresponding publication, we extracted data on the number of serious adverse events (SAEs), the number of patients with SAEs, all-cause mortality (ACM), and the number of deaths. We analyzed the consistency of data between the registry and publications. RESULTS: In 176 included trials, comparisons of safety data between ClinicalTrials.gov and corresponding publications indicated discrepancies for the total number of patients with SAE in 70 (40%) trials, for the reported overall number of SAEs in 77 (44%) trials and ACM in 27 (34%) trials. Overall, trialists reported 33,303 patients with SAEs in ClinicalTrials.gov, compared to 28,118 in matching publications (difference 5185; χ2 = 22,808, P < .001). For the number of SAEs, there were 44,163 SAEs reported in ClinicalTrials.gov, compared to 28,151 SAEs reported in publications (difference 16,012; χ2 = 196,897, P < .001). Compared to the registrations, 17 (21%) trials reported fewer deaths (a total of 4106 fewer deaths), and 8 (10%) trials reported more deaths (43 more deaths) in the publication. Industry-sponsored trials were significantly more likely to report the same number of ACM in registrations and publications (χ2 = 22.72, P < .001). Trials on dapagliflozin had the highest rate of discrepant reporting in publications compared to other medicines. CONCLUSION: The pronounced discrepancies in mortality and SAEs data for SGLT2 inhibitors between clinical trial registries and journal publications highlight the need for improved alignment and transparency in safety reporting. Strengthening reporting standards and editorial processes may enhance the consistency and interpretability of trial safety data.
引用
- Šaler F, Paladin I, Blažeković I, et al. Reporting of sodium-glucose cotransporter 2 (SGLT2) inhibitor trial safety: an observational analysis of ClinicalTrials.gov registrations and journal publications. J Clin Epidemiol. 2026. PMID: 41856422. doi:10.1016/j.jclinepi.2026.112227
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。