Abstract
BACKGROUND: Diabetes is considered a risk factor for mortality following a diagnosis of cancer. We hypothesised that the risk will vary due to the heterogeneous nature of the population and accurate classification of vascular complications will improve prediction of clinical outcomes. METHODS: The COloRECTal cancer data Repository (CORECT-R) was used to identify individuals with primary colorectal cancer, who underwent surgical resection in England (2005-2016). Diabetes was recorded using ICD10 codes (E10-E14) during inpatient hospital admission in the six years preceding cancer diagnosis, complication status was determined using the adapted Diabetes Complications Severity Index (aDCSI). Survival and post-operative outcomes were compared between groups. RESULTS: Of 232,367 individuals, 28,642 (12.3%) were recorded as having diabetes, 49.2% of whom had complications according to the aDCSI. Patients with diabetes complications had increased incidence of adverse post-operative outcomes (90-day post-operative mortality (6.6% versus 3.2%) and death during the surgical episode (7.9% versus 3.6%)), compared to those without diabetes. Those without complications had rates comparable to the population without diabetes. The odds of death within a year of diagnosis were higher for those with complicated diabetes compared to those without diabetes [OR 1.58 (95%CI 1.51-1.66) p < 0.01], but no difference was observed between those with uncomplicated diabetes and those without diabetes [OR 1.05 (95%CI 0.99-1.11) p = 0.10]. CONCLUSIONS: Prediction of outcome following surgery in colorectal cancer patients with diabetes relies on the accurate assessment of complications. This study suggests that the poor post-operative outcomes in diabetes patients may be associated with diabetes complication rather than diabetes itself.