If slimming down to shift out of the "overweight" BMI category is on your list of New Year's resolutions, a new analysis published today may grab your attention: there may be some benefits to carrying a few extra pounds, it suggests.

The review, in The Journal of the American Medical Association, found that overweight (but not obese) people had a 6 percent decreased risk of death.

While the review was extensive — CDC researchers looked at about 100 studies involving almost 3 million people — the surprising news comes with a list of caveats, doctors and obesity experts warn.

BLOG: Does Childhood Obesity Affect Brain Development?

WATCH VIDEO: Inside a Flavor Lab

Even the study's lead author says the findings shouldn't necessarily change people's weight-loss goals.

"This is not meant to suggest that the conventional wisdom is wrong," the Wall Street Journal quotes lead author Katherine Flegal, a senior scientist at the CDC, as saying. Still, it's statistically significant, she points out. "The findings are very consistent across all different ages and continents," she said.

The review concludes that a BMI of 25 to 30 (the World Health Organization's definition of overweight) is associated with a lower risk of death. The analysis included all causes of death in its data. But BMI is a broad measurement that doesn't parse out fat vs. muscle, or the location of extra weight. Both of those factors are key in assessing the healthy weight of an individual, experts say.

Also, if you're a typical American, you may already fall in the overweight category; in other words, don't make the mistake of thinking you should weigh more than your neighbor.

Add in the fact that people often lose weight as they die (and shift into a different weight category), and that 6 percent difference becomes less significant, some say.

BLOG: High-Fat Habits Hard to Break

The authors speculate that overweight people may show signs of disease earlier or be more likely to get screenings for weight-related diseases.

Not convinced? You're not alone. Epidemiologist Walter Willett of the Harvard School of Public Health told NPR "no one should waste their time reading it."

Photo: Corbis