Hypoglycaemia with pioglitazone: analysis of data from the Prescription-Event Monitoring study

Aims and objectives To investigate the relationship between patients' characteristics, use of concomitant anti-diabetic therapies and the incidence of hypoglycaemia, an acute complication of the treatment of diabetes mellitus reported by general practitioners (GPs) during the first 9 months of the treatment with pioglitazone. Methods We used data collected for the Prescription-Event Monitoring (PEM) study conducted by the Drug Safety Research Unit for patients prescribed pioglitazone between November 2000 and June 2001 by their GP in England. A Cox proportional-hazards regression model was used to assess this relationship. Results The original pioglitazone PEM cohort included 12 772 patients (mean age 60.9 years); 53% (6777) were male. A total of 77 patients experienced at least one hypoglycaemic episode (9.64 per 1000 patient-years). Women were estimated to have twice the hazard of having a hypoglycaemic event compared with men [hazard ratio (HR) 2.05; confidence interval (CI) 1.24, 3.41]. Patients taking combination therapy with sulfonylurea or insulin were estimated to have approximately three and four times the hazard of having an event compared with those who were not taking these adjunctive therapies [HR = 3.11 (CI 1.64, 5.88); HR = 4.15 (CI 1.74, 9.91), respectively]. Patients treated with adjunctive metformin were 25% less likely to experience hypoglycaemia than those who did not take concomitant metformin (HR = 0.75; CI 0.44, 1.27). Conclusions This study has shown that the treatment with pioglitazone was associated with a low incidence of hypoglycaemia. The factors possibly increasing the risk of hypoglycaemia were concomitant therapy with sulfonylurea or insulin and female gender.