Diabetes Drug Study in Doubt

A report on Glaxo's Avandia prompted a safety alert, calls for a congressional hearing, and a stock-price tumble, but fears may be overblown

Dr. Steven Nissen, a renowned cardiologist with the Cleveland Clinic, sent shock waves through the diabetes and investment communities with the results of his study, released May 21 by The New England Journal of Medicine. His finding? GlaxoSmithKline's (GSK) diabetes drug Avandia can significantly increase the risk of heart attacks.

But some diabetes specialists charge that those risks are not as serious, or as clear-cut, as the report presented. "That study shouldn't have been published," complains Dr. Joel Zonszein, director of the clinical diabetes center at Montefiore Medical Center in New York. "It could end up doing more harm than good. What [Nissen] really showed is that patients with diabetes have a very high incidence of heart attack, and there is no surprise there."

Deceptive Data?

Nissen's study, a meta-analysis of 42 clinical trials from around the world involving 28,000 diabetic patients, concluded that those taking Avandia had a 64% greater risk of dying from heart disease over a seven-year period and a 43% greater risk of heart attack than patients on other diabetes drugs. The report prompted the Food & Drug Administration to issue a safety alert on Avandia and calls for congressional hearings, sending Glaxo's stock down 8% on the day it was released.

But specialists point out that the actual number of heart attacks recorded in the study was quite small: Nissen found 86 heart attacks among 14,371 patients taking Avandia (or 0.598%), compared with 72 in the 11,634 patients not on the drug (0.619%).

There were 39 deaths from cardiovascular causes in 10,378 patients on Avandia (0.375%) vs. 22 deaths in 9,188 patients not on the drug (0.239%). In other words, though the relative risk of developing a heart attack is much higher in the Avandia group, the overall risk in all patients was relatively low.

Jury Still Out

The difference between the two groups was still statistically significant. But doctors tend not to like meta-analysis because it combines studies with very different criteria that were never meant to be lumped together. Nissen acknowledged this weakness in his report, saying it has "important limitations. We pooled the results of a group of trials that were not originally intended to explore cardiovascular outcomes…. Many of these trials were small and short-term, resulting in few adverse cardiovascular events or deaths." Still, he felt the results were significant enough to warrant a reexamination of the widespread use of Avandia, one of the most popular diabetes drugs in the world. Introduced in 1999, Avandia pulled in some $3 billion in revenues last year for Glaxo.

Not all specialists agree with Nissen. Dr. Sue Kirkman, vice-president of clinical affairs with the American Diabetes Assn., notes that the increase in the absolute, or overall, risk of a heart attack with Avandia is so small that patients should be very cautious about switching. "All drugs have risks and benefits, and there are potential side effects with every drug," she warns. At the University of California in San Francisco Medical Center, Dr. Robert Rushakoff wants most of his 220 patients already on the drug to stay on it for now, until further data comes out that might change his mind. "You always have to be careful when you group together studies that weren't designed to look at cardiovascular disease," he says.

Dr. Rushakoff says the final answer on the heart risks of Avandia won't be in until Glaxo completes a large clinical trial of Avandia it is currently conducting, called RECORD, meant to determine cardiovascular risks. That study will be completed in 2009.

Before it's here, it's on the Bloomberg Terminal.