Evaluation of adverse events of oral antihyperglycaemic monotherapy experienced by a geriatric population in a real-world setting: A retrospective cohort analysis

Background and objective: To evaluate and compare the risk of adverse events (AEs) associated with the use of metformin, sulfonylureas and thiazolidinediones among geriatric patients in a usual care setting. Methods: An electronic medical record database was utilized to identify geriatric patients with type 2 diabetes mellitus aged ≥65 years from 1996 to 2005. Patients naive to oral antihyperglycaemic drug (OAD) therapy were followed for 395 days post initiation of metformin, sulfonylurea or thiazolidinedione treatment. AEs related to study drugs were evaluated during the follow-up period, and the risks of developing an AE were evaluated and adjusted for differences in baseline characteristics by OAD treatment. Results: A total of 5438 patients (mean age 73.2 [SD 5.08] years, 56.1% female) were identified. During the follow-up period, 12.5% of patients experienced an AE (8.3% of metformin, 13.9% of sulfonylurea and 19.8% of thiazolidinedione recipients). Sulfonylurea (odds ratio [OR] 1.74; 95% CI 1.41, 2.13) and thiazolidinedione (OR 2.86; 95% CI 2.23, 3.65) recipients were more likely to experience an AE than metformin recipients, after adjustment for baseline demographic and co-morbidity differences. The average time to onset of a metformin AE (175 days) was less than that for sulfonylurea or thiazolidinedione treatment (192 and 201 days, respectively). The most common AEs were abdominal pain with metformin (42.3%) and weight gain >4.5 kg for sulfonylureas (63.2%) and thiazolidinediones (68.2%). Hypoglycaemia occurred in 2.6% and 2.2% of sulfonylurea and thiazolidinedione recipients, respectively. Discussion and conclusions: Geriatric patients in a real-world setting experienced AEs with metformin, sulfonylurea and thiazolidinedione therapy, although rates differed from those seen in clinical trials, particularly for weight gain and hypoglycaemia. Lactic acidosis occurred at a higher rate with metformin therapy than has been reported in clinical trials, but our results were in the same range for abdominal pain and lower for diarrhoea, nausea/vomiting and dyspepsia. AEs related to sulfonylurea therapy were in the same range as in clinical trials for weight gain but lower for hypoglycaemia, dizziness and headaches. AEs related to thiazolidinedione therapy were more common in our study than in clinical trials, and within the same range for weight gain and elevated liver enzymes but lower for hypoglycaemia and oedema. While AE reporting is likely to be different in a real-world setting than in clinical trials, the observed variances may also be due to the aetiology of diabetes and the physiological response to hypoglycaemia in an older population.