Study objectives: To estimate the incidence of acute liver injury in patients with diabetes and compare it with the corresponding incidence in the general population, and to examine the association between occurrence of acute liver injury and administration of antidiabetic drugs in patients with diabetes.
Design: Retrospective cohort study.
Setting: United Kingdom General Practice Research Database.
Patients: A total of 34,328 patients with adult-onset (type 2) diabetes and a random sample of 50,000 patients with no recorded diagnosis of diabetes; all were aged 20-79 years from 1994-1998 and free of risk factors for hepatic disease.
Intervention: Using the United Kingdom General Practice Research Database as our resource, we followed the two cohorts to identify patients with a recorded diagnosis compatible with acute liver injury Medical records of all patients identified as potentially having liver injury were requested from the general practitioners to validate the diagnosis.
Measurement and main results: After validation, 27 patients met our definition for having acute liver injury: 14 in the diabetes cohort and 13 in the general population cohort. Overall annual incidence was 14.2 and 8.8/100,000 patient-years in the diabetes and general population cohorts, respectively. The adjusted relative risk of acute liver injury in the diabetes cohort compared with that of the general population cohort was 1.0 (95% confidence interval [CI] 0.2-3.7). Patients with diabetes taking antidiabetic drugs had a relative risk of 2.8 (95% CI 0.6-12.5) compared with those not taking these drugs.
Conclusion: Our findings suggest that diabetes itself does not constitute a major risk for acute liver injury and that a small increased risk could be associated with use of antidiabetic drugs. However, due to the low number of cases detected, the estimates of risk with individual drugs are not conclusive.