Cancer, Diabetes, Survival and Glycemic Control: A Large Multisite Analysis

Determine survival, and glycemic control in cancer patients with co-existing diabetes mellitus (DM) from separate geographic locations within a single health system.

Methods:

Breast, colon, lung, pancreas, and prostate cancer data were retrieved from institutional cancer registries from 3 different geographic locations within the same health care delivery system. Data were linked to electronic files to obtain DM status, glucose, and hemoglobin A1c (HbA1c). DM cases were matched to non-DM controls by age, sex, cancer type, staging, geographic location, and year of cancer diagnosis. Overall survival (OS) was compared with Cox regression models, adjusted for Charlson co-morbidity index. Mixed models were used to assess changes in glycemic control over time.

Results:

A total of 7868 patients were analyzed (3934 with DM matched with 3934 without DM). The distribution was: 34% prostate, 23% lung, 18% breast 17% pancreas, and 8% colon. Mean overall age was 67 years, and 90% were white. Significant differences (p < 0.001) in age, Charlson index, and race/ethnicity were present across the 3 sites. In adjusted analysis there were no differences in OS (HR=1.07; 95% CI 0.96, 1.20) according to DM status. Mean A1c was 6.9% and mean glucose 139 mg/dL (DM) and 103 mg/dL (no DM). Among DM patients, hemoglobin A1c decreased significantly (P=0.004), and in both DM and non-DM groups, glucose values decreased during the year following cancer diagnosis (p < 0.001).

Discussion/Conclusion:

This large sampling of five common cancer types with and without DM shows that DM did not affect OS. Cancer did not adversely affect glycemic control. Continued study is needed to confirm findings in more diverse population groups.