Is it better to be a little thin than a little fat?

Yes, say a group of government scientists, taking issue with some recent claims that millions of Americans can afford to put on some extra pounds, that being even 20 percent overweight is no danger and that thinness even can kill.

"The message is that obesity is bad news, and you may pay a penalty for every overweight pound," the National Heart, Lung and Blood Institute's Robert Garrison said in an interview.

Writing in today's Journal of the American Medical Association, he and three fellow scientists from the National Institutes of Health (NIH) say some previous studies that have made thinness look bad have confused thinness with what are really the deadly results of cigarette smoking.

Cigarette smokers tend to become leaner than average, they say. So any study that fails to separate smokers from non-smokers may make it seem that leanness can be dangerous, they contend, when actually it is smoking and the diseases it causes that endanger most thin individuals.

The new assertions are based on a 26-year follow-up of 1,976 Framingham, Mass., men who were aged 30 to 62 when they first were examined. The heart institute has been conducting its now-famed "Framingham Study" since the late 1940s to learn how many possible risks, such as high-fat diets, smoking and obesity, affect heart disease and mortality.

In all, the NIH scientists--Garrison, Dr. Manning Feinleib, Dr. William Castelli and Patricia McNamara--say:

* Of men under "desirable" weights as defined by the Metropolitan Life Insurance Co. in 1959, more than 80 percent were cigarette smokers. The smokers consistently have suffered far higher death rates than non-smokers.

* Although obesity's effects generally were small in those who were less than 20 percent overweight, men no more than 10 percent above their desirable weight have lived longest, on average.

* Every pound above the most desirable weight may take its toll. For every such pound, men originally aged 30 to 49 suffered a 1 percent increase in risk of death over their next 26 years. Men originally aged 50 to 62 suffered a 3.2 percent increase for every such pound.

Metropolitan Life issued new weight tables this year, strikingly increasing its 1959 optimum weights, which are the weights that seem to include the longest-lived people. The new tables add three to 13 pounds to the apparently safest weights that had been given in the earlier table.

But the Metropolitan tables are among those that fail to separate smokers from non-smokers, Garrison pointed out, adding that "We--and the American Heart Association--think we should still go by the 1959 tables."