SOS (Swedish obese subjects) is an on-going intervention trial designed to determine whether the mortality and morbidity rates among obese individuals who lose weight by surgical means (gastric banding, vertical banded gastroplasty and gastric by-pass) differ from the rates associated with conventional treatment. For this purpose, the study is recruiting a sample of obese men and women who constitute a registry of potential subjects from which the participants are drawn. Eligibility criteria for participation in the registry were: age at application 37-57 years and BMI greater than or equal to 34 kg/m2 for men and greater than or equal to 38 kg/m2 for women. Before receiving a health examination, all patients complete extensive questionnaires on current and past health status, utilization of medical care and medications, socio-economic status, psychological profiles, dietary habits, physical activity, weight history, and familial disposition to obesity. Each surgical case is matched to its optimal control in the registry, to ensure that the two groups do not differ systematically with respect to any of 18 matching variables that may affect prognosis. The first 1006 subjects included in the registry have been studied with respect to morbidity and compared with on-going population studies of men and women in Göteborg, Sweden. The relative risks of prevalent disease and symptoms associated with obesity in 50-year-old males and females respectively were 4.3 and 4.7 (dyspnoea), 14.7 and 11.8 (angina), 6.3 (myocardial infarction, males only), 2.1 and 4.5 (hypertension), 5.2 and 6.6 (diabetes), 4.6 and 26.1 (claudication) and 1.7 and 1.8 (gall bladder disease). Correspondingly, obese males and females display elevations of systolic and diastolic blood pressure, fasting glucose, insulin, triglyceride, and uric acid levels. However, total cholesterol was not increased in obese males and was in fact significantly lower in obese compared with reference women. HDL-cholesterol was lower in obese than reference men (data were not available in reference women). The rate of taking sick pensions was over twice as high in SOS obese patients than in population controls. Finally, comparison of measurements with self-reported prevalence estimates revealed a considerable amount of previously undiagnosed hypertension and diabetes in the obese subjects. These data suggest that the excess health risks associated with obesity may not be fully appreciated.