It seems reasonable that overweight and obesity should be defined based on body composition rather than indirect indices like BMI or waist circumference. The use of conventional parameters like fat mass or visceral fat is however of similar and limited value for disease risk prediction at the population level and does not contribute much beyond the use of simple BMI or waist circumference. This conundrum may be partly explained by using complex phenotypes (e.g., Metabolic Syndrome or whole body insulin resistance) rather than more disease-specific outcomes like liver- and muscle insulin resistance. In addition, there are multifactorial causes of similar body composition phenotypes that may add to explain the variance in metabolic consequences of these phenotypes. An intriguing hypothesis is that fat mass represents the metabolic load that interacts with fat-free mass that stands for metabolic capacity to determine disease risk. This concept has important implications for assessment of healthy growth and development and when it is challenged with weight gain in adults. Integration of body composition information at the whole body, organ-tissue and cellular level is not required to improve the diagnosis of obesity but facilitates a better understanding of the etiology of obesity-associated metabolic complications.