Clearing Up the Obesity Paradox in Cardiac Surgery

To the Editor:

We read with interest the recent article by Burgos et al. on the presence of the “obesity paradox” in cardiac surgical patients.¹ The authors found that overweight patients experienced lower unadjusted mortality rates and fewer adverse outcomes after cardiac surgery. However, after adjusting for preoperative variables, body mass index (BMI) did not have an independent effect on mortality, which calls the existence of an “obesity paradox” into question. We commend the authors for their excellent work. We wish to delve deeper into the debate surrounding the “obesity paradox” and discuss how one might reconcile this medical hypothesis with the extensive body of literature on atherosclerosis and cardiovascular disease.

First described in 2003,² the obesity paradox, or reverse epidemiology principal, states that there are better health outcomes for obese individuals than for normal weight–matched individuals with certain medical conditions, such as diabetes, end-stage renal disease, hypertension, heart failure, coronary artery disease, and peripheral artery disease. There is even some evidence that a higher BMI confers a positive survival value on sudden cardiac arrest, both out-of-hospital and in-hospital cardiac arrest, with shockable rhythms (ventricular fibrillation and pulseless ventricular tachycardia), suggesting that the obesity paradox applies to the post-arrest population.³ Since its inception, the obesity paradox has generated hundreds of articles and studies on disparate clinical subpopulations.

The obesity paradox is an example of a statistical phenomenon known as bias resulting from observational studies. The mere association of 2 variables does not in and of itself demonstrate causality as the old adage, “correlation is not causation” states. This is particularly true for retrospective studies. Lack of statistical adjustment for cigarette smoking, for example, may lead to erroneous underestimation of the risk conferred by obesity in some early studies⁴ because smokers tend to have lower BMIs. The same logic holds true for nursing home residents or terminally ill or cancer patients. It is necessary to account for individual disease severity and frailty. Use of the term “obesity,” or use of BMI alone, is insufficient and may be misleading. Obesity should be subdivided into the following categories: class 1 (BMI of 30-34.9 kg/m²), class 2 (BMI of 35-39.9 kg/m²), and class 3 (BMI >40 kg/m²). Studies that cannot parse obese participants into these categories should not assume subgroup homogeneity. Some studies that took a more rigorous view of this problem still appear to validate the obesity paradox.³ This may be a result of the varying levels of fitness of study participants, a variable difficult to quantify post hoc. Outcome studies that take a rigorous look at the effect of BMI,⁵ on the other hand, often do not support the obesity paradox. Perhaps we're looking at obesity the wrong way. Obesity is not a static variable, but consists of several different subpopulations. Obesity class 1 participants frequently may be otherwise healthy, with varying degrees of cardiovascular fitness, whereas obesity class 3 subjects have higher incidence of cardiovascular disease, metabolic syndrome, and obstructive sleep apnea. Frequently used phenotypic indicators of the latter 2 of these conditions are abdominal circumference and waist-to-hip ratio.^{6 , 7} Central obesity correlates better than BMI as a measure of cardiovascular risk.^{6 , 7}

The debate on the obesity paradox is continuing and moving into the subspecialty realm. As anesthesiologists, we're struck that the wrong message is being given to the general public. The obesity paradox is a misnomer.⁸ Obesity is not a benign condition, and obese individuals are not homogenous; they are affected by their personal physical fitness and fragility—items all- too- often not included in the routine assessment tools. Nonetheless, obesity, particularly severe obesity, increases cardiovascular, diabetic, cancer, and stroke risks; shortens life expectancy; and has been shown to be associated with worse coronavirus disease 2019 outcomes.9, 10, 11, 12, 13 We believe that the obesity paradox should be revealed for what it is—a statistical bias of a limited subset that is sometimes correct.

Conflict of Interest

None.