CHICAGO, Illinois (AP) -- U.S. life expectancy will fall dramatically in coming years because of obesity, a startling shift in a long-running trend toward longer lives, researchers contend in a report published Thursday. By their calculations -- disputed by skeptics as shaky and overly dire -- within 50 years obesity likely will shorten the average life span of 77.6 years by at least two to five years. That's more than the impact of cancer or heart disease, said lead author S. Jay Olshansky, a longevity researcher at the University of Illinois at Chicago. This would reverse the mostly steady increase in American life expectancy that has occurred in the past two centuries and would have tremendous social and economic consequences that could even inadvertently help "save" Social Security, Olshansky and colleagues contend. "We think today's younger generation will have shorter and less healthy lives than their parents for the first time in modern history unless we intervene," Olshansky said. Already, the alarming rise in childhood obesity is fueling a new trend that has shaved four to nine months off the average U.S. life span, the researchers say. With obesity affecting at least 15 percent of U.S. school-age children, "it's not pie in the sky," Olshansky said. "The children who are extremely obese are already here." The report appears in the New England Journal of Medicine. In an accompanying editorial, University of Pennsylvania demography expert Samuel H. Preston calls the projections "excessively gloomy." Opposing forecasts, projecting a continued increase in U.S. longevity, assume that obesity will continue to worsen, but also account for medical advances, Preston said. Still, failure to curb obesity "could impede the improvements in longevity that are otherwise in store," he said. Americans' current life expectancy already trails more than 20 other developed countries. Dr. David Ludwig of Children's Hospital Boston, a study co-author, cited sobering obesity statistics: Two-thirds of U.S. adults are overweight or obese; one-third of adults qualify as obese. Up to 30 percent of U.S. children are overweight, and childhood obesity has more than doubled in the past 25 years. Childhood diabetes has increased 10-fold in the past 20 years. "It's one thing for an adult of 45 or 55 to develop type 2 diabetes and then experience the life-threatening complications of that -- kidney failure, heart attack, stroke -- in their late 50s or 60s. But for a 4-year-old or 6-year-old who's obese to develop Type 2 diabetes at 14 or 16" raises the possibility of devastating complications before reaching age 30, Ludwig said. "It's really a staggering prospect." While national attention is starting to focus on contributors to obesity, including the prevalence of fast-food, soft drinks in schools and cuts in physical education classes, "what we presently lack is a clear, comprehensive national vision for addressing the obesity epidemic," Ludwig said. The calculations are a stark contrast with Social Security Administration forecasts for slow improvement in life expectancy, and with projections publicized in 2002 that said the maximum human life span will reach 100 in about six decades. In an interview, Olshansky said he hoped the new research would play a role in the current discussion about overhauling Social Security. Critic calls report 'very one-sided' James Vaupel, director of the Max Planck Institute for Demographic Research in Rostock, Germany, and a research scientist at Duke University, co-authored the 2002 forecast, based on data from developed nations including the United States. Vaupel called the new report "very one-sided" and said he doubts that obesity will negate the effects of other medical progress in improving mortality. Emory University health policy expert Dr. Kenneth Thorpe said that while obesity is clearly damaging public health and driving up health care spending, rising rates aren't enough to resolve Social Security's woes. "That's too simplistic," he said. Other life expectancy forecasts rely on past mortality trends; the Olshansky group used obesity prevalence data and previously published estimates of years of life lost from obesity. They calculated in reverse, assessing the fall in death rates that would occur if all obese Americans had a normal weight. Their estimate shows that, if not for obesity, life expectancy at birth should be four to nine months higher than the record 77.6 years announced by the government last month. That slight gain translates into a loss that will worsen if current trends continue, the researchers said. Richard Suzman, a researcher at the National Institute on Aging, which helped fund the study, said the projections are "possible, but I would say unlikely." He said the best approach is to estimate life expectancy using historical trends. The Center for Consumer Freedom, an advocacy group for the restaurant and food industry, which argues the obesity problem has been exaggerated, said the paper should be discredited because co-author David Allison has done consulting for makers of weight-loss products. Allison, a biostatistician at the University of Alabama at Birmingham, noted that the journal mentions his financial ties. While the study methods are partly based on assumptions, they are also sound, Allison said. Obesity researcher Dr. JoAnn Manson said she agrees with the paper's message, if not the methods. "The calculations that were made may not be perfect," but the emphasis on obesity's dangers "should serve as a wake-up call for policy makers and the public health community," said Manson, chief of preventive medicine at Harvard's Brigham and Women's Hospital. U.S. Sen. Tom Harkin said Wednesday that the report supports his efforts to have government regulation of junk food marketing to children. If the dim life expectancy forecast doesn't demonstrate a need for action, "I don't know what will," the Iowa Democrat said.