Abstract
BACKGROUND: We investigated associations between diabetes and mortality among participants with incident colorectal cancer (CRC) from the Southern Community Cohort Study. METHODS: Participants (73% non-Hispanic Black; 60% income < $15,000) were recruited between 2002-2009. Diabetes was self-reported at enrollment and follow-up surveys at approximately 5-year intervals. Incident CRC and mortality were identified via state registries and the National Death Index. Proportional hazards models calculated associations between diabetes with overall, CRC-specific mortality among 1059 participants with incident CRC. RESULTS: Diabetes prior to diagnosis is associated with elevated overall (hazard ratio [95% confidence interval]: (1.46[1.22-1.75]), and CRC-specific mortality (1.36[1.06-1.74])) after adjustment for tumor stage. For non-Hispanic Black and non-Hispanic White participants, consistent associations were observed for overall (1.35[1.10-1.66] vs. 1.89[1.31-2.72], respectively, p-interaction = 0.11) and CRC-specific mortality (1.30[0.99-1.71] vs. 1.77[1.06-2.95], respectively, p-interaction = 0.28). For individuals with incomes <$15,000/year, associations with overall (1.44[1.15-1.79]) and CRC-specific mortality (1.28[0.94-1.73]) were similar to the full sample. Associations with overall (1.71[1.37-2.13]) and CRC-specific mortality (1.65[1.22-2.22]) were highest for diabetes ≥ 10 years at diagnosis. CONCLUSIONS: Pre-diagnosis diabetes is associated with higher mortality among participants with incident CRC from a predominantly non-Hispanic Black cohort with lower socioeconomic status. The higher prevalence of diabetes in this population may contribute to racial disparities in CRC mortality.