In this large longitudinal prospective cohort study of more than 4000 patients with suspected stable angina pectoris, we demonstrate that obese male patients had a 1.8 fold and 1.6 fold increased risk of incident AMI and CV death compared to normal weight men. By contrast, compared to normal weight women, obese women had similar risk of AMI and CV death, while overweight women had nearly half the risk of incident AMI. The risk of all-cause mortality associated with BMI was similar among men and women, and did not differ significantly across BMI categories.
BMI and risk of AMI, CV death and all-cause mortality in men and women
Strong associations between overweight/obesity and risk of CVD and death have been demonstrated in the general population
[19, 20]. By contrast, several studies of patients with CAD have demonstrated that overweight and/or obese patients may have a better morbidity and mortality prognosis than their leaner counterparts; although as one recent review points out, this observation is not supported by all
Only a few studies of patients with CAD have examined the association between BMI and risk of CV events and mortality in men and women separately. Our finding of an increased risk of cardiovascular events among obese men are in accordance with the results from a previous US study of patients with stable CVD, whereof 85% had CHD, as well as with a multi-ethnic sample study of patients with established CAD
[5, 6]. However, while these studies did not observe any significant associations between BMI and risk of major adverse coronary events in women, we report a nearly halved adjusted risk of AMI among overweight women as compared to their normal weight counterparts.
In the present study, there was no interaction between BMI and gender with regards to all-cause mortality. Moreover, the risk of death did not differ between BMI groups. These findings are in accordance with a previous study conducted among European patients with CAD
. To the best of our knowledge only two studies, of patients with CAD, have examined the association between BMI and all-cause mortality in men and women separately
[7, 8]. First, a study of Danish patients with AMI showed that normal weight, overweight and obese men had similar risk of death, whereas overweight women had a slightly decreased risk (HR (95% CI); 0.78 (0.68, 0.90)) of death, as compared to their normal weight counterparts. Furthermore, in a follow up study of the CADILLAC trial, they observed significantly lower in-hospital mortality (0.9% vs. 2.7%), 30 days (1.1% vs. 3.8%) and 1- year (1.8% vs. 7.5%) mortality in obese patients with AMI undergoing PCI when compared to normal weight patients. Statistical significance was, however, only reached in males.
BMI is often used to quantify overweight and obesity owing to a high fat percentage correlation, but does not account for fat distribution. Men have a tendency to store excessive fat in visceral fat deposits, whereas women usually store fat in peripheral subcutaneous distributions
. Excessive visceral fat is associated with an increased risk of developing metabolic syndrome, putting men at a greater risk of developing CVD, while subcutaneous fat in the femoral-gluteal region may be associated with a more favourable CV risk profile
. Furthermore, overweight and obese postmenopausal women may benefit from the increase in circulating levels of estrogen produced by the adipose tissue
Moreover, at baseline, men were more often affected by CV risk factors and had more severe CAD. Inclusion of these potential confounding variables in stratified multivariate analyses did not alter our results. Differing health status at baseline is thus unlikely to be the cause of the observed gender interaction.
Strengths and limitations of the study
The main strength of the present study is its, well defined population with complete follow up of clinical endpoints. Limitations include the single baseline measurements of BMI and other time dependent cofactors such as medication. We did not have sufficient data on possible confounders such as physical activity, socioeconomic status or cardio- respiratory fitness and intentional vs. unintentional weight loss, and thus we cannot exclude the possibility that residual confounding from unmeasured causal factors unevenly distributed between BMI groups may have influenced our results. Unfortunately, we did not have data on recent weight loss prior to inclusion, but as underweight patients (BMI <18.5 kg/m2) were excluded, and adjustment for possible confounders such as cancer, pulmonary disease, extent of significant CAD and LVEF did not significantly alter our results, reverse causation is unlikely. BMI was positively associated with common obesity related characteristics such as higher blood pressure, diabetes, an unfavourable lipid profile, higher eGFR and CRP. Adjustment for these variables did not have a significant effect on our results, but we would in any case not include these variables in a final multivariate adjusted survival model because of the possibility of over-adjustment bias. We did, however, adjust for use of ACE inhibitors and loop diuretics as a proxy of heart failure, and there is the possibility that these variables may have mediated some of the effect of BMI.
It has previously been suggested that BMI is an inadequate marker of overweight and obesity in patients with CAD
, with waist circumference or waist to hip ratio suggested as better predictors of cardiovascular events, especially in women
[6, 26]. Studies supporting an obesity paradox have almost exclusively used BMI as an index of obesity
. We thus suspect that the diverging findings among such studies may be the result of BMI’s inadequacy as a quantifier of true body fatness and fat distribution.
Given that there were relatively few females in the study population and the event rate was low, we thus had a low statistical power to by which to detect the possible effects of BMI on risk of events among women. Further, we cannot rule out that the relatively lower incidence rate of AMI among women is a result of detection bias; women, compared to men, are more likely to experience atypical symptoms of AMI and may consequently delay seeking medical care for symptoms or be misdiagnosed by healthcare providers
. Finally, the inclusion of predominantly white subjects limits the ability to generalise our findings to non-white populations.