BACKGROUND and OBJECTIVE
Several new antidiabetic agents have been introduced over the last two decades, resulting in increased treatment options as well as the complexity of disease management strategies. More clinical studies are now required to ascertain the effectiveness and safety of each antidiabetic therapy. Cardiovascular (CV) safety concerns have surfaced in previous clinical studies on rosiglitazone, which eventually led to a series of regulatory actions in several countries in 2010. However, the US Food and Drug Administration (FDA) backtracked on its regulatory decisions and finally removed heavy access restrictions imposed on rosiglitazone altogether in 2013, reflecting new evidence concerning the drug’s safety profile. The first aim of this study was to describe the utilization patterns of antidiabetic agents, focusing on the effects of safety signals detected from relevant clinical studies and drug regulation changes for rosiglitazone on prescribing patterns of commonly prescribed antidiabetic drugs (ADs). The second aim was to evaluate the influence of preceding AD therapies on the subsequent treatment adjustments in type 2 diabetes mellitus (T2DM) patients.
METHODS
Patient data were extracted from a national health insurance claims database that covered the years 2007 to 2015. We carried out a linear regression and interventional time series analysis to examine longitudinal drug utilization trends and the impact of five milestone events regarding rosiglitazone’s CV safety on AD prescribing patterns. The association between AD class and subsequent treatment adjustment was examined by Cox proportional hazard regression.
RESULTS
Steady growth was observed in AD consumption, with metformin preserving its dominant market share throughout the study period. Dipeptidyl peptidase-4 (DPP-4) inhibitors and pioglitazone gained popularity in the midst of rosiglitazone failure. A significant decline in rosiglitazone use was observed after intervention 1 (Nissen’s meta-analysis and FDA warnings in 2007), associated with a coefficient of -0.52 (P=0.005), and intervention 4 (restriction or suspension of access in 2010), with a coefficient of -0.84 (P<0.001). The dispensing patterns for DPP-4 inhibitors steadily trended upward following each intervention. Of those, intervention 4 had the most substantial influence per magnitude, but none of the upward trends reached statistical significance. Our concomitancy analysis showed that DPP-4 inhibitors have overtaken sulfonylureas (SUs) since 2014 as the most favored add-on to metformin. In the cohort 2 analysis, the use of metformin as a first therapy was less likely to change patients’ treatment compared to other ADs.
CONCLUSIONS
The prevalence of AD use, especially thiazolidinediones (TZDs) and DPP-4 inhibitors, shifted substantially after clinical evidence publication and regulatory actions on rosiglitazone. Despite new evidence from a long-term surveillance study and the FDA’s subsequent decision to eliminate access restrictions on rosiglitazone in 2013, domestic regulations were left intact; hence, its use remained negligible in Korea. In response to the fall of rosiglitazone, the use rates of pioglitazone and DPP-4 inhibitors appeared to increase, suggesting a potential switching pattern from rosiglitazone.