Since 1980, obesity prevalence among US adults has soared from 14% to 42%. The commonly accepted explanation is pervasive overeating: ever-increasing energy intake as the population gains weight, year after year. However, evidence does not support this hypothesis. National data on energy intake and energy availability show increases between 1961 and 2000, during modern industrialization of food; but a plateau or declines thereafter-even as obesity continued rising-and while physical activity modestly increased. Thus, Americans appear to be eating relatively less since 2000, for ever-increasing body sizes, as time has progressed. Although both energy intake and energy availability are measured with error, such errors would have to be new since 2000 and systematically increasing over time for these 2 separate, independent measures. Given the tremendous societal consequences of obesity, and failure to date of energy balance-focused interventions to stem the tide, it is critical for the scientific community to consider and test alternative hypotheses. Growing evidence suggests complex, interrelated biological interactions between food processing (including acellular nutrients, depleted prebiotics, additives), gut microbial composition and function, host metabolic expenditure, and intergenerational transmission of risk (including epigenetics, noncoding RNAs, microbial species). In this paradigm, whereas increasing energy intake may have contributed to rising obesity in earlier years, today pervasive adiposity and its physiologic adaptations have created a biological milieu which interacts with industrialized foods to promote escalating obesity, even with stable energy intake-a self-sustaining, difficult-to-reverse cycle. These scientific hypotheses must be rigorously evaluated, because even partial confirmation would dramatically shift and expand current prevention and treatment strategies. Urgent new investment in research is required. Simultaneously, uncertain evidence on the obesity epidemic's primary drivers does not mean there is no evidence on actions that can help, and existing science must be more rapidly translated and refined into clinical, public health, and policy interventions.
