Higher grades of obesity were associated with greater all-cause mortality than was normal weight, but low-grade obesity and overweight were linked to lower mortality in a meta-analysis of 97 clinical studies reported in the Jan. 2 issue of JAMA.
These findings, from a systematic review that included nearly 3 million study subjects on five continents, are "broadly consistent" with those of several previous studies and two previous meta-analyses: Excess mortality occurs only at the highest levels of obesity, while mortality is lower among overweight and slightly obese individuals than among people of normal weight, said Katherine M. Flegal, Ph.D., of the National Center for Health Statistics, Hyattsville, Md., and her associates.
The investigators reviewed the literature and included in their meta-analysis all prospective, observational cohort studies that reported all-cause mortality in adults whose body mass index was measured using standard categories. A BMI of less than 18.5 was underweight, that of 18.5 to less than 25 was normal weight, a BMI of 25 to less than 30 was overweight, that of 30 to less than 35 was grade 1 obesity, a BMI of 35 to less than 40 was grade 2 obesity, and a BMI of 40 or greater was grade 3 obesity.
The researchers included 41 studies from the United States or Canada, 37 from Europe, 7 from Australia, 4 from China or Taiwan, 2 from Japan, 2 from Israel, 2 from Brazil, 1 from India, and 1 from Mexico. Overall there were 2.8 million subjects and 270,000 deaths during follow-up.
Overweight and grade 1 obesity were associated with significantly lower mortality than was normal weight, with hazard ratios of 0.94 and 0.95, respectively. In contrast, grade 1 and grade 2 obesity were associated with significantly higher mortality than was normal weight, with an HR of 1.29, Dr. Flegal and her colleagues said (JAMA 2013;309:71-82).
These findings remained robust in several sensitivity analyses.
The reasons why overweight and slight obesity might be protective are still unknown. Some have proposed that slightly heavy patients might present for medical care earlier than normal weight and severely obese patients when symptoms arise, and they also may be more likely to receive optimal medical treatment because caregivers recognize that their weight confers risk. It also is possible that slightly increased body fat may exert cardioprotective metabolic effects and provide better metabolic reserves when illness occurs, the investigators said.
Their meta-analysis also demonstrated that studies relying on patients’ self-report of weight and height are subject to important bias, because these measures are not accurately reported across all categories of age, sex, and race.
No financial conflicts of interest were reported.