Cyclic weight gain may harm the heart.

It may be worse to have lost pounds and regained them than never to have dieted at all -- from a life-expectancy standpoint, at least. That's the message from a new analysis of the effects of weight fluctuation on nearly 3,200 men and women in the Framingham (Mass.) Heart Study.

The Framingham study has monitored participants' health at two-year intervals since 1948. All volunteers were healthy at the outset, their ages ranging from 30 to 62. An international research team has now conducted three different analyses of data spanning 32 years, looking for statistical associations between volunteer's weight variability and each of the following: deaths from all causes; deaths from coronary artery disease; nonlethal coronary artery disease; and cancer risk.

The investigators, led by Lauren Lissner in Gotebory, Sweden, and Kelly D. Brownell of Yale University, found no link with cancer but a significantly elevated risk of premature deaths in general and an even stronger association with heart diseases. "Persons where body weight fluctuates often or greatly have a higher risk of coronary heart disease and death than do persons with relatively stable body weights," they write in the June 27 NEW ENGLAND JOURNAL OF MEDICINE.

Overall, weight variations increased an individual's chances of premature death in general and the threat of lethal or nonlethal coronary disease by 30 to 100 percent, and the increases "tended to be higher for men that for women," the researchers report. Moreover, the heart risk proved independent of obesity and smoking -- and was equal to the risk of being overweight, they assert. The strongest links between weight change and health appeared in men and women under age 45.

In an accompanying editorial, Claude Bouchard of Laval University in St. Foy, Quebec, observes that "this study has considerable strengths," including its large size and its statistical acconting for potentially confounding variables. However, he adds, the finding that variable weight "may have negative consequences for health that are independent of obesity and the rate of [weight] change ...is likely to be controversial" -- especially when animal data have yet to suggest a possible mechanism.

An estimated 25 to 50 percent of U.S. and Canadian adults are attempting to lose weight -- and most will not keep their shed pounds from returning. Lissner's team concludes that the new data "suggest that overweight persons should be taught skills to maintain weight loss and that the prevention of relapse should become a more central focus of weightloss programs."