The study was paid for by Abbott Laboratories, Meridia's maker, and published in The New England Journal of Medicine. And the authors of the study, three of whom are Abbott employees, concluded that the trial results did little more than show that patients with heart problems should not be prescribed Meridia - a restriction already included in Meridia's label.

But in an unusual rebuke, The Journal's top editors wrote an editorial concluding that the study actually showed that Meridia, also known as sibutramine, should be removed from the market.

"It wasn't that we disagreed with the interpretation of the authors," said Dr. Gregory D. Curfman, The Journal's executive editor. "It's just that we thought they didn't quite go far enough." Many patients, Dr. Curfman added, "have cardiovascular disease and don't know it. How are you supposed to identify those patients who might be put at risk by putting them on drugs like sibutramine?"

After seeing preliminary results of the trial in January, the European Medicines Agency ordered Abbott to remove Meridia from the European market. The Food and Drug Administration took a less forceful step and instead requested that Abbott state on the drug's label that Meridia should not be used in patients with a history of cardiovascular disease.

The agency will ask a committee of experts on Sept. 15 whether further steps are needed.

Dr. David J. Graham, a drug-safety official at the F.D.A., said the agency's decision to allow continued marketing of Meridia was mistaken and resulted because officials who approved the drug remain in charge of assessing its safety.

"This once again emphasizes the need for a separate center to oversee the safety of presently marketed drugs," Dr. Graham said.

Karen Riley, an F.D.A. spokeswoman, said the agency had been aware of most of the trial's findings since last year and issued a public warning on Nov. 20.

In the study, called the Scout trial, 10,744 overweight or obese patients with a mean age of 63 and histories of cardiovascular problems were all given Meridia for six weeks. Patients who showed immediate problems on the drug were dropped from the study, and the 9,804 remaining were given either Meridia or a placebo for the remainder of the study. Patients on Meridia lost about nine pounds in the first year; those who received placebos lost about four pounds. The Meridia patients suffered 28 percent more heart attacks and 36 percent more strokes, although there was no difference in heart-related deaths.

Scott Davies, an Abbott spokesman, described Meridia "as an important option to treat a serious condition for which there are few treatments available."

Dr. W. Philip T. James, a professor of nutrition at the London School of Hygiene and Tropical Medicine and the lead author of the study, said that Meridia was beneficial for many obese people without heart problems. Dr. James dismissed the editorial that attacked his study's conclusions. He was paid by Abbott to conduct the study, and his co-authors disclosed a dizzying array of research and marketing relationships with drug companies.

"I think the editors are cavalier in their approach in basing their decision on statistics and not looking at it in clinical terms," Dr. James said in an interview.

Dr. Rudolph L. Leibel, an obesity researcher at Columbia University, said in an interview that the Meridia dispute revolved around an argument between those who viewed drugs from a population basis and those who viewed them from an individual patient basis.

A small number of patients lose considerable weight while taking Meridia and therefore benefit enormously from the drug, Dr. Leibel said. But a far larger number of people get no benefit from the drug, and some of these patients may suffer heart attacks and strokes as a result of taking it.

"The question is do you withdraw the drug to protect the large number of individuals who have no benefit and could have a bad response and thereby eliminate the opportunity for that small number of people who respond well?" Dr. Leibel asked.

It is a dilemma often faced by the F.D.A. in part because the agency has no power over how doctors use medicines. Even when agency officials warn against the use of drugs in certain patients, some doctors prescribe them dangerously anyway and patients die.

In some cases, the agency has forced drugs off the market after some doctors have proven unable to prescribe them safely. In other cases, the F.D.A. has only strengthened warnings - particularly when the injuries associated with the drug's use are common problems like heart attacks and strokes that in individual patients are difficult to blame the medicine for.

Meridia's safety woes and modest efficacy are part of a long history of disappointing efforts by the drug industry to cure the obesity epidemic and cash in on the estimated $59 billion that Americans spend annually fighting fat. Xenical, another diet drug made by Roche, can cause flatulence and oily discharges. A federal advisory panel voted in July against endorsing Qnexa, an experimental fat pill developed by Vivus. And highly promoted scientific efforts to control appetite through leptin, a hormone involved in appetite and metabolism, and through suppression of appetite receptors activated by marijuana have so far failed.