Impact of pioglitazone regulatory withdrawal on antidiabetic drug use and health in diabetic patients.

Abstract

In 2011, pioglitazone was withdrawn from the French market owing to a potential risk of bladder cancer. This study aimed at assessing the impact of this pioglitazone withdrawal (PW) considering (i) trends in antidiabetic uses and (ii) changes in hospitalization/death rates in diabetic patients following PW. We first considered the general population of the Echantillon Généraliste des Bénéficiaires (EGB), a 1/97th representative sample of the French healthcare insurance system beneficiaries, for the 2010-2014 period. In this, for each non-insulinic antidiabetic drug class, changes within the numbers of monthly supplied drug units for 1000 subjects were studied through times series and Unobserved Component Models. Second, we identified from the EGB a cohort of patients who were delivered a non-insulinic antidiabetic between 01 April 2011 and 01 August 2011 (date of PW). In this, post-withdrawal incidences of all-cause hospitalization and death were compared amongst pioglitazone users and non-users using proportional subdistribution hazards models. PW was accompanied by an increase in metformin (+11.7; 95% CI 1.1-22.3) and glinide (+11.0; 95% CI 1.2-20.8) numbers of monthly supplied units for 1000 subjects. No significant change was found for GLP-1 agonists, DPP-4 inhibitors, sulphonylureas or alpha-glucosidase inhibitors. In the cohort of non-insulinic antidiabetic users at the time of PW (1093 pioglitazone users, 17,900 non-users), being a pioglitazone user at PW was not associated with a subsequently higher rate of hospitalization. If PW was accompanied with significant changes in the use of some antidiabetics, no adverse impact of PW on hospitalization or death rates of diabetic type 2 patients was found.