Impact of treatment with pioglitazone on stroke outcomes: A real‐world database analysis.

Aims: Randomized controlled trials have reported an association between pioglitazone and reduced incidence of stroke in type 2 diabetic (T2DM) and insulin‐resistant populations. We investigated this association within a real‐world database. Materials and methods: T2DM patients who initiated pioglitazone between 2000 and 2012 were extracted from the Clinical Practice Research Datalink (CPRD), a UK routine data source. Two non‐exposed control cohorts were matched according to age, gender, HbA1c, diabetes duration, stroke history, co‐morbidities and prior T2DM regimen. Control cohort‐1 comprised patients initiating a new T2DM therapy as their respective case initiated pioglitazone. Control cohort‐2 maintained the same T2DM regimen as their respective case prior to the case initiating pioglitazone. Primary outcome was incident stroke; other outcomes included mortality, length of hospital stay and stroke recurrence. Results: A total of 4234 patients initiating pioglitazone were matched to controls in cohort‐1 and 3604 in cohort‐2. For the primary outcome there were significantly reduced hazard ratios (HRs) for cases: controls. For cohort 1, the HR was 0.666 (95% CI, 0.466‐0.952) during the therapy period and was 0.750 (0.612‐0.919) over the entire observation period; for cohort 2, respective HRs were 0.516 (0.336‐0.794) and 0.773 (0.611‐0.978). There was no significant difference in 30‐day mortality rate or rate of recurrent stroke. For stroke events that required hospitalization, there was a significant difference in length of stay for patients discharged to usual residence (median, 3.0 days vs 7.0 days; P = .008) for control cohort‐2 while undergoing treatment. Conclusions: In support of evidence from 2 large randomized trials, these observational data show that pioglitazone has a potent effect in reducing stroke events in patients with type 2 diabetes. [ABSTRACT FROM AUTHOR]