All large prospective studies (n greater than 20,000) and several smaller studies have found that severe obesity [body mass index (BMI) greater than or equal to 35 kg/m2] is associated with approximately a twofold increase in total mortality and in a severalfold increase in mortality due to diabetes, cerebro-, and cardiovascular disease, and certain forms of cancer. Studies that have not been able to confirm this have been small and/or short term, have failed to control for smoking or early mortality, have controlled for intermediate risk factors in an inappropriate way, or have a reduced internal validity due to misclassification biases. As compared with BMI, abdominal obesity is a stronger predictor of mortality in most studies available. The incidence of sudden death unexplained by autopsy may be up to 40 times higher in severely obese subjects as compared with the general population. A small weight increase since the age of 18 is associated with a decreased risk whereas weight increases greater than 10 kg are associated with an increased mortality. The total mortality ratio for severe obesity decreases from 55 y of age and is not detectable above 80 y of age. Studies lacking adequate control groups indicate that a sustained weight loss may induce a reduced mortality but results from controlled intervention studies are so far not available.