Background: Overweight and obesity in adults are common and associated with cardiovascular risk and other adverse health effects.
Purpose: To review benefits and harms of screening for and treatment of overweight and obesity in adults to assist the U.S. Preventive Services Task Force (USPSTF) in updating its 2003 recommendation.
Data Sources: We searched MEDLINE, the Cochrane Central Registry of Controlled Trials, and PsycINFO from January 1, 2005 through September 9, 2010. Relevant trials published prior to 2005 were identified through good-quality systematic reviews.
Study Selection: Two investigators independently reviewed 6,499 abstracts and 649 articles against a set of a priori inclusion criteria. Two investigators rated the quality of each study based on USPSTF methods. We included trials that involved behavioral-based treatment (38 trials, n=13,495) or the use of orlistat (18 trials, n=11,256) or metformin (3 trials, n=2,652) for weight loss or weight maintenance in adults in settings that are generalizable to U.S. primary care. Additional studies were included for the evaluation of weight loss treatment harms (4 additional behavioral trials, 6 additional orlistat trials, and 1 additional metformin trial).
Data Extraction: Selected elements were abstracted into standardized tables from each study by one investigator and checked by another investigator.
Data Synthesis: Data were qualitatively and quantitatively (using meta-analysis) synthesized separately for each type of intervention. Behavioral treatment resulted in an average weight loss of 3.0 kg more in intervention participants compared with control, with greater weight loss in trials with more treatment sessions (generally 4–7 kg lost in the intervention group in trials with 11–26 treatment sessions in the first year). Orlistat was additive to behavioral counseling, resulting in even greater weight loss (generally 6–9 kg total). Metformin trials were heterogeneous, but one large, good-quality trial showed a weight loss of 2.3 kg more in the intervention group. Weight loss treatments did not improve health outcomes, but they were sparsely reported and most trials were not powered for outcomes such as death and cardiovascular events. Weight loss treatment resulted in a reduction in diabetes incidence in two large, good-quality behavioral-based trials of diabetes prevention. Behavioral-based treatment showed small positive effects on blood pressure. Orlistat improved blood pressure and lowered low-density lipoprotein cholesterol (by 7–16 mg/dL) and plasma glucose (by 12 mg/dL in patients with diabetes) compared with placebo. Metformin did not improve lipid levels or blood pressure, but reduced the incidence of diabetes. Withdrawals due to adverse effects were more common among medication users than placebo users and were primarily related to gastrointestinal complaints.
Limitations: There were minimal data on the distal health outcomes of death and cardiovascular disease. Many intermediate outcomes were sparsely reported, especially in the behavioral treatment literature. There were minimal data on behavioral-based treatment in people with class III obesity (body mass index >40 kg/m2). Behavioral-based treatments were heterogeneous and specific elements were not always well reported. Many medication trials had high attrition and most were conducted outside of the United States. There was one good-quality trial of orlistat and one of metformin but no data on maintenance of weight loss after medications were discontinued. Medication trials were not powered to identify group differences in rare but serious adverse effects.
Conclusions: Behavioral-based treatments are safe and effective for weight loss, although they have not been studied in persons with class III obesity. Medication may increase weight loss beyond behavioral approaches alone, although side effects are common.