(P-540) The Association Between Type 2 Diabetes and Cancer Incidence in the United Kingdom Between 1988 and 2019

Background: Previous studies have associated type 2 diabetes with certain cancers, but these studies are outdated and may not reflect the current trends. Furthermore, there is little information on the cancer incidence patterns after a diagnosis of type 2 diabetes.

Objectives: To examine the association between type 2 diabetes and cancer and to examine the pattern of cancer incidence after the diagnosis of type 2 diabetes.

Methods: Using the Clinical Practice Research Datalink, we used a base cohort of 7,608,656 patients, at least 18 years of age, to identify all those newly-diagnosed with type 2 diabetes between January 1, 1988 and September 30, 2019. Using risk set sampling, these patients were matched, in a 1:1 ratio, to non-diabetic individuals on age, sex, calendar year of base cohort entry, and time in the database. Poisson regression models, conditional on matched pairs, were used to estimate incidence rate ratios (IRR) with 95% confidence intervals (CIs) of any cancer, as well as 20 site-specific cancers, comparing patients with type 2 diabetes with non-diabetic individuals. We also examined whether the IRR varied over the follow-up period (≤2, >2-4, >4-6, >6-8, >8-10, >10 years).

Results: A total of 241,131 patients newly-diagnosed with type 2 diabetes were matched to an equal number of non-diabetic individuals. Patients with type 2 diabetes had a higher incidence of any cancer, compared with non-diabetic individuals (16.3 vs 14.3 per 1000 person-years, respectively; IRR: 1.14, 95% CI: 1.12-1.17). The top five cancers positively associated with type 2 diabetes were pancreatic (IRR: 2.90, 95% CI: 2.58-3.25), hepatobiliary (IRR: 2.85, 95% CI: 2.47-3.27), renal (IRR: 1.58, 95% CI: 1.40-1.80), cervical (IRR: 1.54, 95% CI: 1.18-2.02), and gastric cancer (IRR: 1.36, 95% CI: 1.17-1.57). In contrast, inverse associations were observed for prostate (IRR: 0.92, 95% CI 0.87-0.97) and non-melanoma skin cancer (IRR: 0.94, 95% CI: 0.91-0.97). Overall, the IRR was highest in the first two years after diagnosis for any cancer (IRR: 1.27, 95% CI: 1.23-1.31). The IRRs were lower in the subsequent time periods (IRRs ranging from 1.06 to 1.11). The site-specific cancers generating the highest initial IRRs were pancreatic (IRR: 7.01, 95% CI: 5.56-8.83), hepatobiliary (IRR: 2.02, 95% CI: 1.52-2.70), and head and neck cancers (IRR: 1.77, 95% CI: 1.39-2.28).

Conclusions: The findings of this large population-based study indicate that type 2 diabetes is associated with an increased incidence of cancer. The magnitude of this association varied by cancer type and duration of follow-up. These patterns should be considered when designing cancer pharmacoepidemiologic studies.