PURPOSE: To facilitate future economic evaluations of weight loss interventions, we examined the association of body weight with preference-based quality of life in a nationally-representative sample of U.S. adults. METHODS: We analyzed data on 27,332 adults in the 2000 and 2001 Medical Expenditure Panel Surveys, two nationally representative, cross-sectional surveys of the noninstitutionalized civilian population of the United States. Height and weight were self-reported, and quality of life was assessed using the EuroQol EQ-5D, a generic, preference-based instrument that has 5 dimensions (mobility, self care, usual activities, pain/discomfort, and anxiety/depression). We used linear regression to assess the associations between National Institutes of Health body mass index (BMI) categories and utility weight son the EQ-5D. We used logistic regression to predict the probability of having at least some problems on each of the dimensions of the EQ-5D, by BMI categories. Analyses were multivariable adjusted for sociodemographic characteristics (age, gender, race, income, education, marital status, region of country, insurance type), smoking status, and 6 chronic health conditions (hypertension, diabetes mellitus, coronary heart disease, stroke, asthma, and emphysema) using STATA SE 8.2 and survey commands. RESULTS: Thirty-nine percent of US adults had normal body weight (BMI 18.5-24.9), 36% were overweight (BMI 25.0-29.9), 15% had class I obesity (BMI 30.0-34.9), 5% had class II obesity (BMI 35.0-39.9), 3% had class III obesity (BMI * 40.0), and 2% were underweight (BMI < 18.5). When compared with normal weight adults, EQ-5D utility weights were significantly lower for underweight adults, overweight adults, and those with class I, II, and III obesity (P < 0.05). Above a BMI of 25, we observed a linear, inverse relationship between BMI and EQ-5D utility scores, where a five point change in BMI was associated with an 0.03 change in utility (95% CI: 0.024 to 0.031; adjusted R-squared = 0.15). Obese adults were more likely than normal weight adults to report any limitations in mobility, self care, and usual activities, and more likely to have chronic pain. CONCLUSIONS: Body mass index is associated with lower utility in a dose-dependent fashion. Further research is needed to test the cost-effectiveness of weight loss interventions in U.S. adults using health state preferences derived from the general population.