BACKGROUND - Type II diabetes increases liver cancer risk but the risk may be mitigated by anti-diabetic medications. However, choice of medications is correlated with diabetes duration and severity, leading to confounding by indication.
METHODS - To address this association, we conducted a nested case-control study among persons with type II diabetes in the Clinical Practice Research Datalink. Cases had primary liver cancer and controls were matched on age, sex, practice, calendar time, and number of years in the database. Exposure was classified by type and combination of anti-diabetic prescribed and compared to non-use. Odds ratios (ORs) and 95% confidence intervals (95% CI) were calculated using conditional logistic regression.
RESULTS - In 305 cases of liver cancer and 1151 controls, there was no association between liver cancer and anti-diabetic medication use compared to non-use (OR=0.74 (95% CI=0.45-1.20) for metformin-only, 1.10 (95% CI=0.66-1.84) for other oral hypoglycaemic (OH)-only, 0.89 (95% CI=0.58-1.37) for metformin+other OH, 1.11 (95% CI=0.60-2.05) for metformin+insulin, 0.81 (95% CI=0.23-2.85) for other OH+insulin, and 0.72 (95% CI=0.18-2.84) for insulin-only). Stratification by duration of diabetes did not alter the results.
CONCLUSIONS - Use of any anti-diabetic medications in patients with type II diabetes was not associated with liver cancer, though there was a suggestion of a small protective effect for metformin.