Deeper Dive Into Rosiglitazone Data Reveals Link to CV Risk, Especially Heart Failure

The most detailed look ever at data for rosiglitazone, the diabetes drug marketed by GlaxoSmithKline (GSK) as Avandia, shows the one-time blockbuster significantly raises the overall risk of heart problems or cardiovascular death, calling attention to the need for more transparency in data collection.

Rosiglitazone is among the most controversial drugs in history. Approved in 1999, sales topped $2.5 billion a year when a 2007 study in The New England Journal of Medicine (NEJM) linked the type 2 diabetes treatment to heart attacks and early death, prompting sales restrictions and lawsuits and, ultimately, an FDA guidance that changed the way diabetes and cardiovascular drugs are evaluated. This paved the way for sodium glucose co-transporter 2 (SGLT2) inhibitors, which appear to be all-purpose weapons against diabetes, heart failure, renal decline, and, possibly, obesity.

Today’s findings, appearing in The BMJ, examine results from more than 130 trials involving 48,000 patients, with each comparing rosiglitazone with a control for at least 24 weeks. For 33 trials, the researchers gained access to individual patient-level data, instead of relying on summary data from clinical trial registries, which may or may not include all adverse events. The team from Yale launched the study after GSK made independent patient-level data available, years after the drug’s patent expired in 2012.

For the data pool using individual patient data, rosiglitazone was associated with a 33% increased risk of a composite of cardiovascular events—heart attack, heart failure, cardiovascular death, and noncardiovascular death (odds ratio [OR], 1.33; 95% CI, 1.09-1.61). Secondary outcomes, for each elements of the composite, were as follows:

• Myocardial infarction (OR, 1.17; 95% CI, 0.92-1.51)
• Heart failure (OR, 1.54; 95% CI, 1.14-2.09)
• Cardiovascular-related death (OR, 1.15; 95% CI, 0.55-2.41)
• Noncardiovascular-related death (OR, 1.18; 95% CI, 0.60-2.30)

When results for trials without individual patient data were included, the results were not as strong. But the heart failure finding stands out, given what has occurred in the diabetes therapy market over the past decade.

The study authors said in a statement that the findings show why sharing individual patient data from different sources matters. "Our study suggests that when evaluating drug safety and performing meta-analyses focused on safety, [individual patient data] might be necessary to accurately classify all adverse events," they write. "By including these data in research, patients, clinicians, and researchers would be able to make more informed decision about the safety of interventions."

Rosiglitazone was approved based on its effectiveness at lowering glycated hemoglobin (A1C). After the link to heart attacks was published in NEJM, other studies refuted this finding, but the controversy lingered. As newer drug classes appeared, rosiglitazone never regained its popularity, particular after SGLT2 inhibitors and glucagon-like peptide-1 receptor agonists were shown to have cardioprotective effects.

In 2008, in the wake of the rosiglitazone incident, the FDA issued a guidance that said showing a diabetes drug could lower A1C was not enough. Manufacturers would be required to conduct large, expensive cardiovascular outcomes trials to demonstrate that the drugs did not cause adverse events—namely heart attacks and strokes.

These trials were designed simply to show that newer drug classes were safe—that they would not cause harm. But the surprise came in 2015, when a trial showed that the mechanisms of SGLT2 inhibitors had a positive effect on heart failure, even though this was a initially a secondary outcome. In response, manufacturers launched dedicated heart failure outcomes trials for this drug class, with DAPA-HF reporting positive results last year.

Reference

Wallach JD, Wang K, Zhang AD, et al. Updating insights into rosiglitazone and cardiovascular risk through shared data: individual patient and summary level meta-analyses. BMJ. 2020;368;l7078. doi: 10.1136/bmj.l7078.