In briefing documents released ahead of a US Food and Drug Administration advisory panel meeting on Monday to discuss the thiazolidinedione class of antidiabetic drugs, and GlaxoSmithKline's diabetes drug Avandia (rosiglitazone) in particular, agency staff raise questions over the conflicting data relating to cardiovascular events.

Avandia and Eli Lilly's Actos (pioglitazone) were approved in 1999, having been shown to have a much less risk of liver toxicity than troglitazone, which was withdrawn from the market in 2000. Since that time, post-marketing experience has shown both compounds have the potential for anemia, weight gain, oedema, and risk of heart failure. However, while the risk for oedema and heart failure has been well-appreciated and described for TZDs, the effect of these drugs on cardiovascular ischemic risk was less well known, although there were some early concerns that Avandia may confer some changes in lipid profiles that could lead to macrovascular disease. Additional postmarketing studies led to the first label warning for Avandia in April 2006, associating it with a potential risk of cardiac ischaemia, but subsequent analyses threw up conflicting results.

"These various datasets present an array of somewhat inconsistent findings that complicate the interpretation of the available data regarding the effect of rosiglitazone on cardiac ischemic events. Nonetheless, given the findings...FDA views this signal with considerable concern," it stated in the briefing documents.

The agency had always planned a review of Avandia's safety, but brought the meeting forwards after Steven Nissen's meta-analysis showing the risk of cardiovascular ischaemic events with Avandia heightened the public's and physicians' concerns. The clinical trial pool of data up for evaluation on Monday comes from 42 studies of rosiglitazone administered as monotherapy and in combination with sulfonylureas, metformin and insulin.

A doubling of risk due to rosiglitazone added to insulin was seen consistently across all endpoints, the briefing document notes, adding that this combination should be carefully re-assessed. Excluding insulin trials results in a non-significant link to cardiac ischaemia.

The FDA will ask its advisory panel to consider whether the drug's prescribing information should be updated and whether new warnings should be added to Avandia's labelling.

The briefing document is available at: