Patient-Level Data Affirm That Rosiglitazone Increases Cardiovascular Risk

Patient-level data support previous evidence that the antidiabetic drug rosiglitazone is associated with increased cardiovascular risks, especially heart failure, researchers say.

"Our study highlights that when conducting meta-analyses evaluating drug safety, access to raw clinical trial data is necessary to accurately identify and classify adverse events," Dr. Joshua Wallach of Yale School of Public Health in New Haven told Reuters Health by email. Traditional data sources such as publications and clinical trial registrations are often incomplete, he noted.

Recent efforts by GlaxoSmithKline to make patient-level data from rosiglitazone trials available to external investigators prompted the team to re-analyze all available data.

"Overall, the study confirms previous concerns regarding rosiglitazone's cardiovascular risk, especially for heart failure," he said. "However, using the raw data from 33 trials and numerous different meta-analytical methods, we found the risk of myocardial infarction to be somewhat lower than what was previously reported. This suggests that different results can be derived from different data sources, and that raw data may be necessary to produce more accurate risk estimates."

Overall, the study included 33 trials with more than 21,000 participants for whom patient-level data were available. The meta-analyses for myocardial infarction included 103 trials with about 23,000 subjects for whom patient-level data were not available, and a similar population for cardiovascular-related deaths

As reported in The BMJ, among 29 trials for which patient-level data were available and that had been included in previous meta-analyses that used summary data, the new analysis found more myocardial infarction events in 26 trials and fewer cardiovascular-related deaths for five trials.

When analyses were limited to trials for which patient-level data were available, patients treated with rosiglitazone had a 33% increased risk of a composite event compared with controls (odds ratio 1.33). Specifically, the ORs were 1.17 for myocardial infarction; 1.54 for cardiovascular-related death; 1.15 for heart failure; and 1.18 for non-cardiovascular-related death.

By contrast, in analyses for which patient-level data were not available, ORs were attenuated for myocardial infarction (1.09) and cardiovascular-related death (1.12).

Based on their findings, the authors conclude that use of patient-level data "might be necessary when performing meta-analyses focused on safety."

Dr. Wallach noted that rosiglitazone, which was approved about 20 years ago, is no longer available in Europe and rarely prescribed in the US. "However, the rosiglitazone story is an important lesson in the value of clinical trial transparency," he said. "Our study demonstrates how data-sharing platforms can be used to improve our understanding of drug safety."

Dr. Sabrina Islam, Assistant Professor of Medicine, Section of Cardiology, at the Lewis Katz School of Medicine at Temple University in Philadelphia, commented by email, "This study adds to the body of medical data that is available regarding rosiglitazone.... The results ... further emphasize that attention should also be paid to patients with coronary disease, as there is a likelihood of increased adverse effects with use of this medication. Physicians have approached the medication with caution, and it is not generally prescribed for diabetes management."

—Marilynn Larkin

Source