Diabetes Associated with Short Survival in Pancreatic Cancer.

Yixiang Mao,Min Tao,Xiaoyan Jia,Donghui Li

DOI: https://doi.org/10.1200/jco.2014.60.2896

2015-01-01

Abstract:association between long-term diabetes and poor overall survival among patients with diabetes and pancreatic cancer in three longitudinal studies. The authors reported a 40% to 50% increased risk of death for patients with long-term diabetes (duration of 4 years) compared with those without diabetes (P .05) in two prospective cohort studies 2,3 and a clinic-based study. 1 In contrast, no statistically significant association between short-term diabetes (duration of 4 years) and survival was observed. Thus, the authors concluded that there is a link between chronic glucose intolerance and pancreatic cancersurvival,whereasshort-termdiabetesdoesnotportendaworse prognosis among patients with pancreatic adenocarcinoma. We have a concern about the authors’ conclusion regarding the association between short-term diabetes and survival. Because the information on diabetes was collected via a biennial questionnaire in the cohort studies, it is possible that the number of patients with short-term diabetes was greatly underestimated, which may have biased the risk estimate toward null. The prevalence of patients with short-term diabetes was 4.67% (47 of 1,006) and 18.65% (72 of 386), whereas the prevalence of patients with any type of diabetes was 17.99% (181 of 1,006) and 31.35% (121 of 386) in the cohort and the clinic-based studies, respectively. The prevalence of diabetes in the clinic-based study was quite similar to the numbers previously reported by other studies. Short-term diabetes was associated with a nonsignificant but increased risk of death in this clinic series (hazard ratio [HR], 1.22; 95% CI, 0.89 to 1.67). In our recently completed studyof1,328patientswithpathologicallyconfirmedpancreaticductal adenocarcinoma of all stages at MD Anderson Cancer Center, we foundthatshort-termdiabetes(durationof3years)wasassociated with a significantly increased risk of death (HR, 1.23; 95% CI, 1.09 to 1.40).Wehavealsoconductedameta-analysisof29studiesondiabetesandoverallsurvivalofpatientswithpancreaticcancerandfounda reduced survival of patients with diabetes in 25 clinic-based studies (pooled HR, 1.13; 95% CI, 1.02 to 1.25), but not in four cohort studies 4-7 (pooled HR, 1.12; 95% CI, 0.95 to 1.32). The prevalence of patients with diabetes was 28.61% (4,133 of 14,446) and 20.92% (1,124 of 5,372) in the clinic and cohort studies, respectively. These observations suggest that short-term diabetes is associated with a higher risk of death in patients with pancreatic cancer. The inconsistentobservationswithrespecttotheassociationbetweendiabetesand