Statistics from Altmetric.com
If we were sitting with two patients who had coronary heart disease (CHD), possibly post myocardial infarction, 20 years ago, and one had a body mass index (BMI) of 23 kg/m2, or ‘normal’ BMI, and the other had a BMI of 32 kg/m2, or class I obesity, we would have likely been much more concerned about the short-term and long-term prognosis in the heavier patient. However, during the past 15 years, considerable evidence from our own studies and meta-analyses tell us that we were not just incorrect, but actually flat out wrong regarding our prognostication in these two patients!1–3 Using a meta-analysis of 89 studies in over 1.3 million patients with established CHD, by far the largest of such studies, Wang et al4 confirmed the information from prior studies, and provided several important new observations on this interesting and controversial topic.
How can obesity be protective?
When we (CJL, RVM) and others first made these observations 15 years ago, about the obesity paradox in heart failure (HF), there were concerns that there may be fatal flaws in our data. Certainly, excess body weight is associated with adverse effects on cardiovascular (CV) risk factors, including hypertension (HTN), lipids, glucose and inflammation, and is associated with an increased risk of almost all CV diseases (CVD), including HTN, CHD, HF and atrial fibrillation.1–3 Therefore, no expert to our knowledge is suggesting that obesity is a ‘good thing’, and support for the obesity paradox does not make this claim. However, 15 years later, it is hard to deny the fact that at least mild increases in BMI in patients with CVD appear to be associated with a better prognosis than noted in patients with low BMI.
Is BMI the problem?
As Wang et al4 alluded to, some have suggested that since BMI includes total weight (muscle, skeletal and fat), it may be partly the reason why there is a surprising relationship between BMI and prognosis in patients with CVD.1–3 Although there may be a discrepancy between BMI and other assessments of obesity,2 ,3 nevertheless, we have demonstrated the obesity paradox even with per cent body fat (BF) in patients with CHD2 ,3 and HF,1 ,2 and this has even been demonstrated with waist circumference (WC) or central obesity in both HF1 and CHD.2 ,3 Despite the potential problems with BMI for assessing true body fatness in individual patients, BMI is quite useful to reflect body composition in large populations.
Reverse causation and cardiorespiratory fitness
As Wang et al4 suggest, most of the studies are not able to exclude non-purposeful weight loss prior to study entry. In fact, besides lower body weight, low levels of cholesterol and blood pressure are also associated with poor prognosis in HF, although this has not been demonstrated to our knowledge in cohorts with CHD. Nevertheless, in the present study,4 the obesity paradox is more evident during early follow-up, and appears to disappear after 5 years of follow-up in this large meta-analysis. Additionally, patients who have CHD with moderate and severe obesity (class II/III) have a higher mortality during long-term follow-up, supporting the data published in primary prevention. However, we demonstrated that in large cohorts of patients with CHD followed for close to 15 years, patients with decent levels of cardiorespiratory fitness (CRF), defined as not being in the bottom tertile of CRF for age and gender, have a good prognosis regardless of BMI, BF or WC, whereas in those with low CRF, an obesity paradox was present during long-term follow-up, meaning that those with low CRF and the lowest categories of body composition (BMI, BF and WC) had a worse CVD-cause and all-cause mortality than heavier patients.2 ,3
Body fatness versus CRF
As discussed above, our data suggest that CRF markedly alters the relationship between adiposity and subsequent prognosis. In both cohorts with CHD2 ,3 and HF,1 we have only found an obesity paradox among those with low levels of CRF, whereas those with preserved CRF have a relatively good prognosis regardless of body composition. In a recent meta-analysis by Barry et al5 of 10 studies that quantified the combined association of CRF and weight on mortality, they demonstrated that compared with normal-weight-fit individuals, unfit individuals have a twofold higher mortality regardless of BMI, whereas overweight-fit and obese-fit individuals have similar mortality risk as normal-weight-fit individuals. Therefore, the constellation of these data suggests that CRF is much more important than obesity regarding long-term prognosis. As Wang et al4 addressed, the lack of data on physical activity (PA), much less CRF, is certainly a limitation of their meta-analysis.
What about obesity severity?
Some evidence have suggested that the obesity paradox mostly refers to an overweight or mild obesity paradox,2 ,3 and we have even suggested that this is more of a ‘lean paradox’, and this information may not apply to more severe obesity.2 Interestingly, during short-term follow-up, Wang et al4 even showed that CHD patients with class II/III obesity had better survival than those with normal BMI. However, during long-term follow-up, at least over 5 years, the very heavy patients with BMI ≥35 kg/m2 have a higher mortality. This certainly suggests that moderate and severe obesity may take a ‘heavy toll’ on the CV system in the long term. Therefore, avoiding more severe obesity and moving the class-II/III obese patients into the overweight or class-I level, along with obtaining a higher level of CRF, should be a long-term goal for obese patients with CHD. This also supports the recent Flegal et al6 meta-analysis of 97 studies in 2.9 million individuals, which showed that the best survival occurred in the overweight group, and a trend for better survival was present in the class-I obese patient compared with those in the normal-BMI group. Although obese as a group had worse survival, this was all due to higher mortality in the more severe grades of obesity.
Is weight loss still recommended?
If overweight and obese patients with CHD have a better survival than leaner patients, does this lessen the need for weight loss? Although the data regarding long-term safety and efficacy of purposeful weight loss in general and in patients with CVD, including CHD, are limited, weight loss combined with PA and exercise training (ET) has many potential advantages.2 ,3 In a recent meta-analysis by Pack et al7 of 12 studies and 14 cohorts (n=35 335), overall weight loss was associated with a significant 30% increase in major CVD events. However, whereas observational weight loss in 10 cohorts was associated with a 62% increase in risk of CVD events, presumed intentional weight loss in four cohorts was associated with a 33% reduction in risk. Another meta-analysis recently published of 15 randomised controlled trials of 17 186 participants suggested that purposeful weight loss in obesity was associated with approximately a 15% reduction in all-cause mortality.8 Considering the decreasing protection of obesity noted in the long-term follow-up by Wang et al,4 especially in those with class II/III obesity, long-term weight loss still appears advantageous, especially when combined with increasing PA, ET and efforts to improve CRF. Additionally, obese patients have more BF, and have more muscle mass than leaner patients.1–3 With weight loss, BF is reduced, but muscle mass also declines. Since muscle and lean mass are also important,2 and muscular strength is associated with better CVD risk factors and lower mortality, resistance training during weight loss and afterwards is also important to increase muscle mass and muscular strength.
In conclusion, this meta-analysis by Wang et al,4 which was by far the largest and with the most detailed follow-up, extends the observations from earlier and smaller cohorts and adds valuable information about short-term versus long-term prognosis and the impact of obesity severity. Supporting the obesity paradox, however, does not mean that obesity is being promoted or that one is suggesting that obesity is a good thing. Quite possibly, many overweight and obese patients with CHD may not have developed CHD in the first place had weight gain been prevented. However, at least in the short-term, overweight and obese patients with CVD, including CHD, appear to have a very favourable prognosis. Quite likely, this prognosis will be improved even further with long-term weight loss, especially if this is combined with increased PA, ET and improved CRF. Certainly, promoting PA throughout our healthcare system is urgently needed, and this may go a long way to prevent obesity, especially more severe degrees of obesity, as well as improving CRF and prognosis in the primary and secondary prevention of CVD, including CHD.
Contributors All authors contributed to this editorial.
Competing interests CJL is the author of the book, The Obesity Paradox, and in the past has received honorarium from the Coca-Cola Company on fitness/obesity, but not related with their products.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.