Yang Minxia, Qiu Di, Huang Minguang, Yu Shengjian, Xuan Feng
Department of Radiology, Shaoxing People's Hospital, Key Laboratory of Functional Molecular Imaging of Tumor and Interventional Diagnosis and Treatment of Shaoxing City, Shaoxing, China.
Department of Hematology, Zhuji Affiliated Hospital of Wenzhou Medical University, Shaoxing, China.
Cancer Control. 2025 Jan-Dec;32:10732748251381428. doi: 10.1177/10732748251381428. Epub 2025 Sep 19.
IntroductionImmune checkpoint inhibitors (ICIs) have redefined cancer therapeutics. However, they may provoke immune-related adverse events (irAEs), with diabetes potentially altering their patterns. We aimed to investigate whether diabetic cancer patients exhibit a distinctive or intensified irAE pattern.MethodsWe performed a real-world, retrospective pharmacovigilance study of ICIs using the FDA Adverse Event Reporting System from 2011 to 2025. Reports listing anti-PD-1 (Nivolumab, Pembrolizumab, Cemiplimab), anti-PD-L1 (Atezolizumab, Avelumab, Durvalumab), and anti-CTLA-4 (Ipilimumab, Tremelimumab) agents as suspected drugs were extracted. Disproportionality signals were identified with 4 algorithms: Bayesian Confidence Propagation Neural Network, Reporting Odds Ratio, Proportional Reporting Ratio, and Multi-item Gamma Poisson Shrinker. Time-to-onset was calculated from therapy start to event date, modelled with Weibull distributions, and compared across subgroups with non-parametric tests.ResultsOf 22,775,812 FAERS reports, 1886 involved ICIs used in cancer patients with comorbid diabetes. 423 (22.4 %) were fatal and 1463 (77.6 %) non-fatal. Men predominated (71.5 %), and 63.0 % of patients were aged 65-85 years. Combination therapy (anti-CTLA-4 plus PD-1 or PD-L1) accounted for the highest death proportion (29.6 %). Disproportionality analysis revealed the strongest preferred-term signals for pneumonitis/interstitial lung disease, hypothyroidism, and colitis among all diabetic cancer patients receiving ICI therapy. At the system-organ-class level, endocrine, hepatobiliary, and blood/lymphatic disorders showed the most consistent risk across agents. Weibull modelling demonstrated an early-failure pattern (shape β < 1) with a median time-to-onset of 126.6 days overall, shortening to 90.9 days with combination therapy. Fatal subgroup occurred sooner than non-fatal subgroup (median 106.7 vs 132.5 days; = 0.004).ConclusionDiabetic cancer patients experienced the full spectrum of ICI-associated toxicities, with combination treatments linked to greater lethality. Multidisciplinary surveillance during the first 3-4 months of therapy, glycemic control, and long-term follow-up may be essential to optimize benefit and minimize harm in this expanding population.
引言
免疫检查点抑制剂（ICI）重新定义了癌症治疗方法。然而，它们可能引发免疫相关不良事件（irAE），糖尿病可能会改变其模式。我们旨在调查糖尿病癌症患者是否表现出独特或加重的irAE模式。
方法
我们使用美国食品药品监督管理局（FDA）不良事件报告系统，对2011年至2025年期间使用ICI的情况进行了一项真实世界的回顾性药物警戒研究。提取了将抗PD-1（纳武单抗、帕博利珠单抗、西米普利单抗）、抗PD-L1（阿替利珠单抗、阿维鲁单抗、度伐利尤单抗）和抗CTLA-4（伊匹木单抗、曲美木单抗）药物列为可疑药物的报告。使用4种算法识别不成比例信号：贝叶斯置信传播神经网络、报告比值比、比例报告比值和多项目伽马泊松收缩器。从治疗开始到事件发生日期计算发病时间，用威布尔分布进行建模，并通过非参数检验在亚组间进行比较。
结果
在22775812份FAERS报告中，1886份涉及在患有糖尿病合并症的癌症患者中使用的ICI。423例（22.4%）为致命病例，1463例（77.6%）为非致命病例。男性占主导（71.5%），63.0%的患者年龄在65 - 85岁之间。联合治疗（抗CTLA-4加PD-1或PD-L1）的死亡比例最高（29.6%）。不成比例分析显示，在所有接受ICI治疗的糖尿病癌症患者中，肺炎/间质性肺病、甲状腺功能减退和结肠炎的首选术语信号最强。在系统器官类别层面，内分泌、肝胆和血液/淋巴系统疾病在所有药物中显示出最一致的风险。威布尔模型显示出早期失败模式（形状参数β<1），总体发病时间中位数为126.6天，联合治疗时缩短至90.9天。致命亚组的发病时间早于非致命亚组（中位数106.7天对132.5天；P = 0.004）。
结论
糖尿病癌症患者经历了与ICI相关的全部毒性，联合治疗与更高的致死率相关。在治疗的前3 - 4个月进行多学科监测、血糖控制和长期随访，对于在这个不断扩大的人群中优化获益和最小化危害可能至关重要。
