Researchers looking at cholesterol-reduction studies have stumbled upon a curious phenomenon: What's good for your heart may not be so great for those around you.

Both humans and monkeys on cholesterol-reducing programs may be more prone to violence than those who remain on a more typical high-fat diet, according to two reports, which many researchers say are yet to be confirmed. Researchers also said they do not know why cholesterol-reducing diets might lead to violence-prone behavior.

The controversial findings raise questions about the wisdom of advising everyone to switch to a low-fat diet, as the American Heart Association recommends. High-fat diets, especially those high in animal fats, are believed to contribute to the body's production of excess amounts of artery-clogging cholesterol.

If it is confirmed that lowering cholesterol leads to violent behavior, the findings might suggest that people not at high risk of developing heart disease -- such as those without a family history of heart disease, whose cholesterol levels are normal and who do not smoke -- should think twice before starting a cholesterol-reducing program.

"At present the recommendation is that everybody try to reduce their cholesterol," said Matthew F. Muldoon, a University of Pittsburgh researcher who was chief author of a British Medical Journal report on the phenomenon. "It may be necessary to focus only on those individuals who are at elevated risk of heart disease."

"Of everyone living, approximately one-third will die of a heart attack," he added. "There's a lot of people . . . who aren't going to die of a heart attack. At the least, it's inefficient to treat everyone."

But many dispute the findings, and even Muldoon stresses that they need to be confirmed before any changes are made in recommendations concerning low-fat diets. And even if his findings are repeated, the long-term risk of developing heart disease may far outweigh any short-term risk to others from violent tendencies prompted by the diet.

Muldoon's study is one of two that have appeared during the past year that link cholesterol reduction with violent behavior.

The first, presented at the American Psychosomatic Society's annual meeting last fall, looked at data from studies of monkeys conducted in the early 1980s. It found that monkeys shifted from a high-fat diet (43 percent of calories from fat) to one similar to that endorsed by the American Heart Association (30 percent of calories from fat) were about 50 percent more likely to start fights than monkeys who remained on the high-fat diet.

"The findings were pretty straightforward," said the researcher who presented the data, Jay Kaplan of the Bowman Gray School of Medicine at Wake Forest University. "The observation in the monkeys establishes the possibility of the relationship {between cholesterol reduction and violence}. That's what makes you think there's something significant going on."

Kaplan said it was not a case of monkeys getting meaner as they got leaner. They were all the same weight at the beginning and end of the study; they all consumed the same number of calories. Nor was it a case of the animals on rich diets getting sluggish as their arteries were getting clogged. The increased aggressiveness was present in monkeys during the first three months of the study as well as during the rest of the two-year study.

Muldoon's article, published Aug. 11, used a technique called meta-analysis to look at death rates in six different heart disease-prevention studies. It found that people who adopted cholesterol-reducing regimens -- such as going on low-fat diets or taking medications such as gemfibrozil and cholestyramine -- were as likely to die during the study as people who did not take steps to lower their cholesterol.

Why? Because the death rate from suicide, homicide or accidents for people on a cholesterol-reduction program turned out to be 1.76 times higher than that of people who did not adopt a cholesterol-reducing regimen. Their rate of death from cancer was the same, while their rate of death from heart disease was significantly lower.

"The bottom line of our study is: There's no evidence from primary prevention trials that longevity is increased, and there's a suggestion that cholesterol may be associated with some adverse effects," Muldoon said.

Researchers had noted an increase in deaths from violent causes when some of the primary studies were published, Muldoon said. But because the number of such deaths reported in each of these studies was small, they were interpreted as insignificant and dismissed as being a fluke.

Meta-analysis, which combines data from several studies to get one result, allowed Muldoon to analyze information collected on almost 25,000 male patients -- a pool big enough to produce statistically significant differences in death rates.

Whether the subjects used a drug regimen or a diet program to lower their cholesterol did not seem to matter. People using either approach were equally likely to die a non-illness-related death.

That similarity is important because people in such studies did not know whether they were taking a placebo or a cholesterol-reducing drug, suggesting that the stress of being on a cholesterol-reduction program was not what was altering their behavior, some said.

Neither Muldoon nor Kaplan could say why such regimens might lead to altered behavior. One theory is that as blood cholesterol goes down, the chemical composition of cell membranes, which contain cholesterol and other fatty molecules, changes. Such changes might affect the activity of brain cells, it is theorized.

Muldoon said he thinks the act or process of modifying blood cholesterol, not lower cholesterol per se, may account for the altered behavior. Previous studies have suggested that people with high blood cholesterol levels are just as likely to die from accidents, suicide or violence as people with low blood cholesterol levels.

But Muldoon and others stress that the British Medical Journal findings are limited because they looked only at short-term death rates. People in the meta-analysis were followed for an average of about five years. The true effects of a cholesterol-reduction plan on heart disease death rates probably would not be seen until much later, many researchers said.

"It may be as long as 10 to 20 years before you see the impact," said Alan Chait, a fat-metabolism researcher at the University of Washington in Seattle who is chairman of the American Heart Association's nutrition committee. "If you're going to look at mortality, you've got to design trials that are big enough and go on for long enough. Otherwise you're simply not going to find any effect."

Both Muldoon's and Kaplan's reports have been criticized because they rely on data from studies that were not designed to look for an increase in violence among patients on cholesterol-lowering regimens. Kaplan's data came from 30 monkeys used as controls in a study of whether environment increases the development of atherosclerosis, or hardening of the arteries.

Jacques Rossouw, a South African cardiovascular epidemiologist who is a visiting researcher at the National Heart, Lung and Blood Institute, said that while Muldoon's study of heart patients was "interesting, it doesn't carry the same weight as a study which was set up to prove the hypothesis."

He said there were only 103 deaths from accidents, suicide or violence in Muldoon's study. And within this group of non-disease deaths, the study did not make a separate count of deaths from accidents, from suicides or from homicides. "Meta-analysis notwithstanding, chance could still be playing a role," Rossouw said.