ABSTRACT
AIM: To quantify the relationship between adherence to oral anti-diabetic drugs and incident hypoglycaemia in Type 2 diabetes. METHODS: Utilizing a claims database, we identified patients with Type 2 diabetes initiating metformin, sulphonylureas or thiazolidinediones and classified adherence over the next 6 months, creating markers of changes in therapy (switches/additions). We created nine mutually exclusive exposure groups, including: metformin ≥ 80% adherence; metformin < 80% adherence; sulphonylurea ≥ 80% adherence; sulphonylurea < 80% adherence; thiazolidinediones ≥ 80% adherence; thiazolidinediones < 80% adherence; switching to a new class; adding on therapy; and switching to two or more different classes of medication. We followed patients for incident hypoglycaemia medical visits and developed a Cox proportional hazards model to compare rates of hypoglycaemia across exposure groups. RESULTS: Adherence to monotherapy was high (86.0 ± 17.8% for metformin, 87.2 ± 17.5% for sulphonylureas and 87.8 ± 16.9 for thiazolidinediones). The incidence of hypoglycaemia ranged from 93.1 to 259.9 per 10 000 person-years in the nine exposure groups. Relative to metformin users with ≥ 80% adherence, those switching from any monotherapy to combination therapy had a 32% increased rate (hazard ratio 1.32; 95% CI 1.07-1.64) of hypoglycaemia. Thiazolidinediones users with ≥ 80% adherence had a decreased hazard rate (hazard ratio 0.67; 95% CI 0.46-0.98) relative to metformin users with ≥ 80% adherence. All other groups on oral anti-diabetic drugs, regardless of adherence, were not associated with hypoglycaemia CONCLUSIONS: We found that the relative rate of hypoglycaemia was highest in patients switching from monotherapy to combination therapy, while rates of hypoglycaemia in monotherapy users were largely unrelated to level of adherence.
