The terminology "reverse epidemiology" was first proposed by Kamyar Kalantar-Zadeh in the journal Kidney International in 2003 and in the Journal of the American College of Cardiology in 2004. It is a contradiction to prevailing concepts of prevention of atherosclerosis and cardiovascular disease; however, active prophylactic treatment of heart disease in otherwise healthy, asymptomatic people is and has been controversial in the medical community for several years. The mechanism responsible for this reversed association is unknown, but it has been suggested that, in chronic kidney disease patients, "The common occurrence of persistent inflammation and protein energy wasting in advanced CKD seems to a large extent to account for this paradoxical association between traditional risk factors and CV outcomes in this patient population." Other research has proposed that the paradox may be explained by adipose tissue storing lipophilic chemicals that would otherwise be toxic to the body. The obesity paradox was first described in 1999 in overweight and obese people undergoing hemodialysis, and has subsequently been found in those with heart failure, myocardial infarction, acute coronary syndrome, chronic obstructive pulmonary disease, and in older nursing home residents. In people with heart failure, those with a body mass index between 30.0–34.9 had lower mortality than those with what would normally be considered an ideal weight. This has been attributed to the fact that people often lose weight as they become progressively more ill. Similar findings have been made in other types of heart disease. People with class I obesity and heart disease do not have greater rates of further heart problems than people of normal weight who also have heart disease. In people with greater degrees of obesity, however, risk of further events is increased. Even after cardiac bypass surgery, no increase in mortality is seen in the overweight and obese. One study found that the improved survival could be explained by the more aggressive treatment obese people receive after a cardiac event. Another found that if one takes into account COPD in those with peripheral artery disease, the benefit of obesity no longer exists.
Criticisms
The obesity paradox has been criticized on the grounds of being an artifact arising from biases in observational studies. Strong confounding by smoking has been noted by several researchers. Since smokers, who are subject to higher mortality rates, also tend to be leaner, inadequate adjustment for smoking would lead to underestimations of the risk ratios associated with the overweight and obese categories of BMI. In an analysis of 1.46 million individuals, restriction to never-smoking participants greatly reduced the mortality estimates in the underweight group, as well as strengthening the estimates in the overweight and obese groups. A similar 2016 study found that a low BMI that has been "induced by healthy lifestyles" is the best method of reducing the risk of premature death. Another concern is reverse causation due to illness-induced weight loss. That is, it may not be low BMI that is causing death but rather imminent death causing low BMI. Indeed, unintentional weight loss is an extremely significant predictor of mortality. Terminally ill individuals often undergo weight loss before death, and classifying those individuals as lean greatly inflates the mortality rate in the normal and underweight categories of BMI, while lowering the risk in the higher BMI categories. Studies that employ strategies to reduce reverse causation such as excluding sick individuals at baseline and introducing time lag to exclude deaths at the beginning of follow-up have yielded estimates of increased risk for body mass indices above 25 kg/m2. Critics of the "paradox" have also argued that studies supporting its existence almost always use BMI as the only measure of obesity. However, because BMI is an imperfect method of measuring obesity, critics argue that studies using other measures of obesity in addition to BMI, such as waist circumference and waist to hip ratio, render the existence of the "paradox" questionable. One probable methodological explanation for the obesity paradox is colliderstratification bias, which commonly emerges when one restricts or stratifies on a factor that is caused by both the exposure and the outcome.
