The latest findings on Avandia, a top-selling diabetes drug, raise concerns both about its safety and about the way the manufacturer and the Food and Drug Administration have responded to signs of danger. It would be rash to make definitive judgments until the F.D.A. completes a detailed analysis. But the handling of this case bears disturbing resemblances to the Vioxx debacle, in which early warning signs were ignored by its manufacturer until the evidence of serious harm became inescapable and the drug was pulled from the market.

Avandia was approved for sale in 1999 based on studies showing that it could lower blood glucose levels in patients suffering from Type 2 diabetes, also known as adult-onset diabetes. The assumption was that the drug could alleviate some of the most damaging effects of the disease, such as heart attacks and other cardiovascular ailments. But a paper just published in The New England Journal of Medicine suggests that Avandia may instead increase the risk of a heart attack by 43 percent and perhaps the risk of cardiovascular deaths as well.

The study - an analysis of the combined results of 42 previous studies that compared people who took the drug with people who did not - is not definitive, and the absolute risk to any given patient is small. But the study points to a risk that could potentially harm thousands of patients a year. Its lead author was Steven Nissen, a cardiologist at the Cleveland Clinic, who blew the whistle on the cardiovascular risks of Vioxx, which adds to the sense of déjà vu.

What's most troubling is the possibility that early warning signs were not adequately heeded.

Seven years ago a leading diabetes doctor warned the F.D.A. of "a worrisome trend in cardiovascular deaths," two years later a safety monitoring group within the agency expressed concern over cases of heart failure in patients taking the drug. In late 2005 and again in August 2006 the drug's manufacturer - GlaxoSmithKline - informed the F.D.A. that its own combined analysis of various studies showed a 31 percent increase in adverse cardiovascular events. But the company also cited studies that it said showed no evidence of harm, and the F.D.A. concluded the findings were too mixed to take immediate regulatory action.

Now, prodded by Dr. Nissen's article, the agency has issued a safety alert, is completing its own safety analysis and will convene a panel of experts to review the data. The best hope for a definitive answer may lie in a clinical trial that won't finish monitoring all enrolled patients until late next year.

If the verdict goes against Avandia, the F.D.A. and Glaxo will have a lot to explain. Congress will need to probe hard to determine what risks the agency and the manufacturer were aware of and what they did - or didn't do - to mitigate them.