Hershman Dawn L, Till Cathee, Wright Jason D, Ramsey Scott, Barlow William E, Unger Joseph M
Dawn L. Hershman and Jason D. Wright, Columbia University Medical Center, New York, NY; Cathee Till, William E. Barlow, and Joseph M. Unger, SWOG Statistical Center, Fred Hutchinson Cancer Research Center; and Scott Ramsey, Fred Hutchinson Cancer Research Center, Seattle, WA.
JCO Clin Cancer Inform. 2017 Nov;1:1-12. doi: 10.1200/CCI.17.00040.
Diabetes is common, increases with age, and may affect outcomes among people with cancer. Understanding the association between diabetes and cancer outcome is challenging, because patients with diabetes have increased all-cause mortality compared with patients without diabetes.
We systematically examined the phase III trial database of SWOG to identify patients enrolled in trials during the period from 1999 to 2011. We linked the SWOG clinical records to Medicare claims data according to Social Security number, sex, and date of birth. Medicare claims were used to identify diabetes with at least 6 months of continuous Medicare coverage immediately before registration. Multivariable Cox regression was used to compare survival outcomes between patients with and without diabetes for each of 10 tumor cohorts. The primary outcome was overall survival. We also examined progression-free survival and cancer-free survival.
In total, 6,422 patients from 15 trials were ≥ 65.5 years of age, of whom 3,173 patients (49%) met the criteria for linkage to Medicare claims. Thirty percent (n = 952) had claims for diabetes before registration. Patients with diabetes were more likely to be black ( P < .001), but no other differences in demographic characteristics were observed. In multivariable Cox regression, no association was found between baseline diabetes and overall or progression-free survival; in one case, patients with diabetes had marginally worse cancer-free survival (advanced non-small-cell lung cancer; P = .05). A global test found that baseline diabetes was associated with worse overall survival ( P = .03) across the entire panel of analyses.
Diabetes is common among elderly patients enrolled in clinical trials. Unlike prior observational studies, among patients treated with uniform treatment regimens, and controlling for known prognostic factors, we did not observe an association between diabetes and progression-free or cancer-free survival.
糖尿病很常见，且随年龄增长而增加，可能会影响癌症患者的预后。了解糖尿病与癌症预后之间的关联具有挑战性，因为与非糖尿病患者相比，糖尿病患者的全因死亡率更高。
我们系统地检查了SWOG的III期试验数据库，以识别1999年至2011年期间参加试验的患者。我们根据社会保险号、性别和出生日期将SWOG临床记录与医疗保险理赔数据相关联。医疗保险理赔数据用于识别在登记前至少连续6个月享有医疗保险覆盖的糖尿病患者。使用多变量Cox回归比较10个肿瘤队列中糖尿病患者和非糖尿病患者的生存结果。主要结局是总生存期。我们还检查了无进展生存期和无癌生存期。
总共有来自15项试验的6422名患者年龄≥65.5岁，其中3173名患者（49%）符合与医疗保险理赔数据相关联的标准。30%（n = 952）的患者在登记前有糖尿病理赔记录。糖尿病患者更有可能是黑人（P <.001），但在人口统计学特征方面未观察到其他差异。在多变量Cox回归中，未发现基线糖尿病与总生存期或无进展生存期之间存在关联；在一种情况下，糖尿病患者的无癌生存期略差（晚期非小细胞肺癌；P = 0.05）。一项整体检验发现，在整个分析组中，基线糖尿病与较差的总生存期相关（P = 0.03）。
糖尿病在参加临床试验的老年患者中很常见。与先前的观察性研究不同，在接受统一治疗方案治疗且控制已知预后因素的患者中，我们未观察到糖尿病与无进展生存期或无癌生存期之间存在关联。
