Associations of diabetes and mortality among colorectal cancer patients from the Southern Community Cohort Study

Abstract

We investigated associations between diabetes and mortality among participants with incident colorectal cancer (CRC) from the Southern Community Cohort Study. 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. 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. 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.