In this large population-based study of all consecutive patients with a confirmed diagnosis of CAD seen in a defined geographic locale over 5 years and followed for a median of 46 months, we found that mildly and/or moderately obese patients with established CAD had reduced mortality compared with normal BMI patients, whether treated with medical management only, PCI, or CABG. Although our findings may appear counterintuitive given the epidemiological link between obesity and cardiovascular disease in the general population,2
they are very much in line with previous reports of an inverse relationship between excess weight and all-cause and cardiovascular mortalities in individuals with established coronary disease. This observation has been referred to as the ‘obesity paradox’ or ‘reverse epidemiology’.25
Our large Canadian-based study adds to the increasing knowledge and awareness in this evolving field of survival paradoxes in chronic disease states.
It has been previously suggested that the obesity paradox post-CABG or post-PCI may reflect the possibility that only the low-risk obese patients are being selected for revascularization;26
however, our data suggest that mild to moderately obese patients who are not revascularized
within 1 year of catheterization and instead are treated with medical management alone and those without a diagnosis of CAD also demonstrate lower mortality rates. The reduction in mortality rates was also consistent among the HF and MI subgroups and persisted after adjustment for known prognostic factors. The lowest adjusted mortality was noted in patients with BMI between 30.0 and 39.9 kg/m2
, whereas a progressive increase in mortality was evident below this range, and a plateau or slight increase was observed above this range in patients with CAD. These findings are similar to recent meta-analyses conducted in patients post-PCI,9
and in patients with HF.10
Our finding of the obesity paradox in patients who have received a cardiac catheterization but without coronary disease is unexpected. It is possible that patients who received a cardiac catheterization but did not show diseased coronaries had other cardiac risk factors that would classify them as having ‘pre-clinical’ disease, and that having a higher BMI offers protection against mortality even in this population. However, another possibility is that this is indicative of a referral and treatment bias in CAD. Being obese is a visible risk factor that may predispose physicians to refer them for a cardiac catheterization earlier than those of normal BMI. We also showed that patients with mild or moderate (but not severe) obesity are more likely to be revascularized with either CABG or PCI in patients with CAD, despite having lower risk coronary anatomy. This is similar to Yancy et al
who also showed that individuals with a BMI of 25.0–35.0 had the highest rates of coronary procedure utilization, whereas patients with a BMI of ≥40 kg/m2
had the lowest odds of receiving cardiac catheterization PCI or CABG.
There have been several proposed explanations for the obesity paradox. Similar to previous studies,29
our data show that compared with normal weight patients, obese patients are younger and have less severe disease at the time of cardiac catheterization. It has been suggested that obese patients who present earlier have more recognizable and aggressively treated co-morbidities. In addition, as patients who are obese tend to have higher systolic blood pressure, this may permit more aggressive upward titration of disease-modifying medications such as ACE-inhibitors and beta-blockers. However, data regarding the association between guideline-recommended treatment and obesity are conflicting30,31
and our data did not suggest that patients with obesity are being more aggressively treated with statins, ACE-inhibitors, beta-blockers, and nitrates, nor that BMI is associated with more invasive treatment strategy. The discrepancies in reported associations between obesity and medical and invasive treatment likely reflect differences in centre-specific practice patterns. Other potential explanations include obesity as protective against protein-energy malnutrition post-revascularization32
and altered neuroendocrine profiles (N-terminal BNP) of obese patients that may play a role in modulating heart disease progression.33
Our study has a number of strengths. We report on a large, well-categorized, consecutively enrolled, population-based cohort of patients. We were able to assess the effect of BMI on mortality in those with and without CAD according to treatment strategy in all patients undergoing cardiac catheterization from multiple centres. The large number of patients provided adequate power to evaluate survival in those with severe obesity, and our results from the stratified analyses are unbiased estimates of the effect of BMI on mortality in both males and females. Our study provides long-term follow-up data (median 46 months, maximum 84 months). In addition, the data are linked to up-to-date records from the Alberta Bureau of Vital Statistics and thus we had no right censoring in our survival analyses.
A limitation of our study is that the observational nature provides associative, not causal, evidence and therefore the possibility of selection bias and residual confounding cannot entirely be ruled out. However, the observational nature of this study also provides real-world data on the largest cohort studied to date. Secondly, although BMI is the most commonly used epidemiological measure of obesity, it is imperfect and does not directly distinguish between adipose and lean tissue or central and peripheral adiposity. Thirdly, we were unable to control for the role of non-purposeful weight loss prior to study entry. Our risk-adjusted analysis, however, did include age, current smoking status, and history of malignancy, which are factors that could lead to non-purposeful weight loss. Fourthly, since this was a record linkage study, we did not have data on cause-specific mortality. Finally, 8.6% of the patients were excluded for missing or implausible BMI values. This excluded group indeed could have differing results than those included in the analysis; however, given the size of the cohort analysed, it is unlikely that the results would have changed substantially.
In conclusion, we found that a paradoxical association exists between higher BMI and survival in patients with established CAD irrespective of medical, interventional, or surgical treatment strategy. This paradoxical observation is also seen in patients who were referred for cardiac catheterization but who have normal coronary arteries. Despite having lower risk coronary anatomy in obese patients with CAD, patients who are overweight, have either mild or moderate (but not severe) obesity, are more likely to receive revascularization procedures compared with those with normal BMI. Further studies are needed to explore the possibility of a treatment/referral bias as an explanation for the obesity paradox in the CAD population.