(USA TODAY) -- A company-sponsored study found that the weight-loss drug Meridia raised heart attack and stroke risk in patients with pre-existing heart disease, according to a report out today.

The publication comes just two weeks before the Food and Drug Administration is to convene an advisory panel of outside experts to discuss the fate of Meridia, which the agency approved in 1997.

"It is difficult to discern a credible rationale for keeping this medication on the market," concludes an accompanying editorial in The New England Journal of Medicine.

Based on preliminary data, the FDA's counterpart for the European Union suspended Meridia's marketing license in January, but the FDA only added stronger information on the label about use by patients with pre-existing cardiovascular disease.

"The label has always had a warning about people with cardiovascular disease," Abbott Laboratories spokesman Scott Davies said Wednesday. The drug, which isn't actively marketed in the USA, is expected to have sales of only $30 million this year, Davies said.

Abbott will propose "risk-management activities" at the upcoming FDA advisory committee meeting, Davies said, but he declined to describe them.

In the study, nearly 10,000 patients 55 and older with pre-existing cardiovascular disease, type 2 diabetes or both were randomly assigned to take either Meridia or a placebo. All of the patients also participated in a weight-management program. On average, the patients took their pills for about 3 1/2 years. Researchers found that the risk of a heart attack or stroke was 16% higher in those on Meridia, although they weren't more likely to die of cardiovascular disease.

The increased risk of non-fatal heart attack or stroke was found only in patients who had pre-existing heart disease.

"On the basis of these results," the researchers conclude, "sibutramine should continue to be excluded from use in patients with pre-existing cardiovascular disease."

The problem, according to the editorial authors, is that in the real world outside a clinical trial, "it can be difficult to reliably identify patients" who have heart disease but no symptoms.

Copyright 2010 USA TODAY, a division of Gannett Co. Inc.