CardioExchange Archive

CardioExchange’s John Ryan interviews Richard G. Bach about his study group’s observational analysis of data from the BARI 2D trial regarding outcomes associated with rosiglitazone. The article and its accompanying editorial are published in Circulation.

THE STUDY

Using 4.5 years of follow-up data from 2368 patients with type 2 diabetes and CAD in the BARI 2D trial, the BARI 2D investigators compared outcomes among participants treated with rosiglitazone versus participants not receiving a thiazolidinedione. In multivariable-adjusted analyses, the two groups were similar in their on-treatment risks for death, for MI, and for congestive heart failure (CHF); the composite incidence of death, MI, or stroke and the incidence of stroke alone were significantly lower among rosiglitazone recipients than among patients who did not receive a thiazolidinedione. In propensity-matched analyses, the two groups were similar in their risks for major ischemic cardiovascular events and CHF.

THE INTERVIEW

Ryan: Clinical trials show an increased risk for MI associated with rosiglitazone use. Your observational study of rosiglitazone within the BARI 2D cohort did not show a clear signal of risk. How strong do you consider this evidence? Should this study influence how people think about rosiglitazone’s safety?

Bach: The data suggesting an increased risk for MI associated with rosiglitazone use came from meta-analyses of randomized trials; many of the trials had small sample sizes, short-term follow-up, and non-adjudicated outcomes in low-risk patient populations. Although not a randomized trial of rosiglitazone, BARI 2D has several strengths: It was designed to prospectively assess cardiovascular outcomes among patients who may be considered at high risk for cardiovascular harm; it analyzes a large number of patient-years of exposure to rosiglitazone and a large number of independently adjudicated cardiovascular endpoints; and it employed thorough analyses of long-term outcomes, including propensity matching. Bearing in mind that all of the available data have limitations, the results from BARI 2D contribute significant and relevant information that does not suggest harm from rosiglitazone — information that should be included in any considerations of the drug’s safety.

Ryan: Has this study changed your prescribing habits?

Bach: The results from BARI 2D suggest a lack of cardiovascular hazard, and some analyses even suggest potential cardiovascular benefit from rosiglitazone for patients with type 2 diabetes and established CAD. Nevertheless, the prescription of rosiglitazone has been strongly affected by widely publicized concerns about the potential for harm (despite the limitations of and uncertainty surrounding the data from previous studies) and by regulatory agencies’ tough restrictions on the drug. Given all the available the data, I would feel comfortable prescribing rosiglitazone to similar patients needing improved glycemic control, with suitable counseling regarding the controversy. However, the current obstacles to prescribing continue to make that approach difficult, if not entirely impractical.

Ryan: Please explain how this study evolved. Did GlaxoSmithKline fund the analysis? (It’s hard to tell from the listed sources of funding.) Who had the initial idea for the study?

Bach: The NIH funded the large majority of the BARI 2D trial. However, GSK provided supplementary funding and rosiglitazone medication that was used to support patient treatment and follow-up in the trial. GSK did not provide specific funding for these analyses, and the company was not involved in the design, conduct, or write-up of the analyses. The idea for the study originated when a working group of BARI 2D investigators was established to examine the effect of various hypoglycemic medications used in BARI 2D on outcomes. The group recognized that BARI 2D provided an important dataset where the effect of rosiglitazone on prospectively collected and independently adjudicated cardiovascular outcomes could be examined. The goal was to contribute information relevant to the controversy and uncertainty about the drug’s cardiovascular safety. From those early discussions, an interested group of BARI 2D investigators then specifically designed and performed the analyses needed to examine any associations between rosiglitazone and cardiovascular outcomes. The results are reported in the paper.

JOIN THE DISCUSSION

How do the BARI 2D findings influence your perspective on the safety of rosiglitazone and your willingness to prescribe it?