A National Institutes of Health panel said yesterday that obesity is a "killer disease" and that people who are "even five pounds" overweight have cause for concern.

The panel, convened to develop a consensus on the subject, said that persons who are only 20 percent over their desirable weight should be considered obese and in need of medical help. The panel found that 34 million Americans fit that description.

Speaking for the 14-member panel yesterday, Dr. Jules Hirsch, a professor at Rockefeller University, said, "Obesity is a killer. It is a killer as smoking is."

"I think that there has been a great deal of confusion about whether obesity is a biologic disorder or something more involved with appearance than with health hazards," he said. "At the extremes, there has been very little doubt about health hazards, but the evidence we have seen shows us that, at surprisingly low levels, there is a very pervasive health hazard in many systems of the body."

In addition to previously reported risks of high blood pressure, abnormally high levels of cholesterol in the blood and adult diabetes, the panel said obesity has been linked to an increased risk of cancers of the colon, rectum and prostate in men and cancers of the gall bladder, bile passages, breast, cervix, ovaries and uterus in women.

Under the panel's definition, for example, a man 6 feet tall can be considered obese if he weighs 192 pounds or more, while a woman 5-foot-4 fits that description if she weighs 159 or more.

The panel also said that studies have shown that the location of body fat may be at least as important as the quantity of fat in determining the potential hazard to health.

Studies in Sweden and this country show that persons are more likely to suffer complications from obesity when their fat is concentrated in the waist, abdomen and upper body than are people who have fat hips and thighs.

In other words, a "pot belly" or "spare tire" appears to pose more of a health threat than does a wide bottom. If a person's waist measures at least as big as the person's hips, there is a greater health risk, studies showed.

Researchers do not know what accounts for the greater risk, Hirsch told a news conference, but they do know that abdominal fat tissue is more active chemically than fat deposits on the hips.

Panel member Dr. Edward Huth, editor of the Annals of Internal Medicine, suggested that the problem of obesity today is similar to the smoking problem 20 years ago. He cited the parallels that "not everyone who is obese will suffer its adverse effects, the effects take many years to appear and treatment is extremely difficult."

"I think it is fair enough to predict that eventually obesity will be seen in the same historical context with what we've been through with smoking," Huth said.

The specialists noted that "not every fat person will die of one of the related diseases, any more than every cigarette smoker will die of lung cancer." But Hirsch estimated that perhaps as many as 20 percent of those who are considered obese would die because of it. The number is small, he suggested, because such risk factors as hypertension and diabetes can be controlled by medication and other weight hazards, such as osteoarthritis, can be treated surgically.

As a rough guide to desirable weights, the panel suggested using charts issued by the Metropolitan Life Insurance Co. in 1983, and in some cases 1959, and basing calculations on the midpoint of the range of weights for each height category. The consensus statement acknowledged, however, that the charts are not perfect because they do not consider age or percentage of body fat.

The panel said it preferred to use a "body mass index" as a guide to desirable weights.

Under this formula, a man's weight in kilograms is divided by the square of his height in meters. If the answer is more than 30, the panel said, he would be considered to have "medically significant obesity."

The formula for women is more complicated: the weight in kilograms is divided by height in meters to the 1.5th power. The cutoff point would be 30, or possibly slightly less.

The panel recommended that new tables be devised to take into account recent findings about fat distribution.

Although the panel recognized obesity in children as a problem, it expressed its concern, as Hirsch put it, that "excessive zeal or misplaced treatment of children can create more damage than any weight loss could correct."