Confounders can partly explain obesity paradox in STEMI patients


Fabienne Witassek (AMIS Plus Data Centre, Institute of Social and Preventive Medicine, University of Zurich, Zurich, Switzerland) report in Swiss Medical Weekly that the supposed obesity paradox—in which being obese seems to confer a mortality benefit—observed in some studies of patients with ST-segment elevation myocardial infarction (STEMI) may be at least partly explained by confounders. In the study, obese class I patients were younger and had less renal disease.

Witassek et al write that there are conflicting results for the obesity paradox in STEMI patients undergoing percutaneous coronary intervention (PCI); therefore, they reviewed data from the AMIS Plus register, which collects data for patients with acute myocardial infarction in Switzerland, to evaluate the relationship between obesity and mortality in STEMI patients undergoing PCI. They categorised patients (7,891 overall)—according to World Health Organization definitions—as being underwent (body mass index <18.5kg/m2; 1%), normal weight (BMI 18.5–24.95kg/m2; 33%), overweight (BMI 25–29.9kg/m2; 45%), obese class I (BMI 30–34.9kg/m2; 15.9%), and obese class II/III (BMI ≥35kg/m2; 5%).

The authors found a U-shaped distribution for in-hospital mortality with class I obesity patients having the lowest mortality and underweight patients having the highest mortality. They note: “In relation to normal weight, the odds for in-hospital mortality were significantly lower for patients obesity class I (odds ratio 0.56) and significantly higher for underweight patients (odds ratio 2.72). The odds for overweight and obese class I/III did not differ significantly from the normal weight category.”

However, after statistical adjustment for confounders, no significant difference for the odds of in-hospital mortality was found for any BMI group compared with normal weight. The independent predictors of in-hospital mortality were age per additional year (OR 1.06), history of diabetes (OR 2.15), moderate to severe renal disease (OR 2.32), resuscitation prior to admission (OR 5.58), and Killip class >2 (OR 8.01). According to Witassek et al, these findings imply that the mortality benefit initially observed in patients with class I obesity “can be, at least partly, explained by confounders”.  They add: “In our study, obese class I patients were younger and had a lower renal disease rate than the other BMI groups. In the multivariate model, these two factors were independent predictors of in-hospital mortality and may therefore provide a partial explanation for the favourable mortality outcome in this group.” Furthermore, the occurrence of Killip class >2—the strongest predictor of in-hospital mortality in the study—was lowest among patients with class I obesity.

Witassek et al conclude: “Further studies are needed to investigate the reasons for lower mortality in obese patients, which may have implications for the treatment of BMI groups at higher risk.”

Study author Dragana Radovanovic (head of AMIS Plus Data Center, Institute of Social and Preventive Medicine, University of Zurich, Zurich, Switzerland) told Cardiovascular News: “Obese patients who survived a STEMI should be encouraged to lose weight as there is enough evidence showing that weight reduction is associated with lower risk for future events. However, overweight patients must be individually assessed by taking into account their age as well as other risk factors as the recommendation to lose weight strongly depends on how overweight the patient actually is. We do not dare tell patients with a BMI of 26 or 32 not to worry about carrying a few extra kilos, but we must recommend an overall change in lifestyle: quit smoking, control blood pressure, improve diet and increase physical activity – which is more important than merely reducing overweight. Long-term studies are needed to evaluate which specific subgroups of patients with established cardiovascular disease can possibly benefit from being overweight, and the underlying pathophysiological mechanisms.”