Obesity and prognosis: Time to forget about metabolically healthy obesity

This editorial refers to ‘Separate and combined associations of obesity and metabolic health with coronary heart disease: a pan-European case-cohort analysis’†, by C. Lassale et al., on page 397. Obesity is a major epidemic of this century. In western countries, obesity and overweight people make up about half of the population.1 It has never been so easy to get food with so little physical effort, and mankind has never had so many tools to do so many physical tasks. The immediate consequence is that physical activity is decreasing (which raises the probability of increasing overweight/obesity at a population level),2 and a delayed consequence is the increase in prevalence of type 2 diabetes mellitus that parallels the increase in overweight/obesity a decade later.3 Another consequence is the increase in the prevalence of obese adolescents who, in turn, will probably become obese adults. All in all, this increases mortality,4 as has been recently shown by Mendelian analysis where obesity and diabetes are causal to coronary heart disease.5 In fact, primary and secondary prevention treatments have enhanced cardiovascular (CV) prognosis thanks to reductions in cholesterol levels and blood pressure, but increases in obesity and type 2 diabetes mellitus have lessened the overall benefit.6 Among this dark landscape, some sunbeams have been shining. Some authors have described a metabolically healthy obesity phenotype (MHO), where obese individuals show normal blood pressure and have metabolic parameters such as glycaemia, total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides within the normal ranges. There are several diagnostic definitions for MHO, which have an impact from an epidemiological and prognostic point of view. Different series in the literature describe a prevalence ranging between 3 and 32% in men and 11 and 43% in women.7 When initially described, individuals with MHO appeared to have prognoses similar to those with normal weight. However, over the last few years several reports have been published with longer follow-up that challenge these results. In fact, some authors have questioned the concept of MHO itself, including the European Society of Cardiology Guidelines on Cardiovascular Prevention,8 where this type of obesity is only regarded as a transient phase between a normal metabolically healthy state and an obese metabolically unhealthy one according to the long-term results of the Whitehall study.9 In this issue of Eur Heart J, Lassale et al. deal with the so-called MHO phenotype,10 testing the CV risk of MHO patients and metabolically unhealthy obesity patients compared with healthy individuals with normal body mass index (BMI) in a large European prospective study. From the 520 000 individuals included in the European Prospective Investigation into Cancer and Nutrition study (‘EPIC-CVD’), the authors designed a case-cohort study with a representative random subcohort of 18 249 participants (62% women), from which 17 640 participants remained after the exclusion of 609 with a prior history of myocardial infarction or stroke at baseline. Median follow-up was 12.2 years. The study was carefully carried out by trained health professionals. The large number of individuals included allowed a complex statistical analysis, with an extensive database allowing exhaustive adjustment with potential confounders. The results are interesting, and are in accordance with the European Society of Cardiology Guidelines on Cardiovascular Disease Prevention. Obesity increases CV risk, regardless of metabolic abnormalities. It is interesting to note that overweight and obesity are more prevalent in southern countries, but these countries have a larger proportion of MHO members. A significant interaction was observed between waist circumference (visceral adipose tissue) and BMI (peripheral and visceral adipose tissue) although, as expected, the association of waist circumference with coronary heart disease (CHD) was stronger at every level of BMI than any of the associations of BMI with CHD across the tertiles of waist circumference. These results are in accordance with most recent papers on the subject, challenging the very concept of MHO. Other cohort studies that have shown similar conclusions via different approaches [Whitehall 2, Inter99, Multi-Ethnic Study of Atherosclerosis (MESA) study] have also challenged the concept of MHO because, as observed in the present manuscript, there was an excess of CV risk in MHO compared with normal weight individuals.11–13 The definition of metabolic health is also a limitation, but as Hinnouho et al. showed using five definitions, most of them increased mortality risk both for MHO and metabolically unhealthy obese patients.14 Interestingly, the relevance of the media and commercial advertisements has recently been shown. One study observed that it is possible to estimate a nation's future obesity prevalence (in a few years time) based upon the frequency of mentions of sweet snacks (positively associated) and vegetables or fruits (negatively associated).15 Another possible interpretation (not stated by the authors, because the study was not designed with this purpose) could be that the modification of advertisements in the media might reduce obesity prevalence. As acknowledged by the authors, there are several potential limitations of the study. First, there are several definitions of MHO and of metabolic syndrome, and the authors have analysed ‘only’ the most frequently used in the literature. Second, there is no longitudinal follow-up regarding the evolution of weight and metabolic status (i.e. there may be people that change category from normal weight to overweight and vice versa, or may develop metabolic syndrome through follow-up). And finally, there are limitations due to the design of the EPIC-CVD study, which included centres that only recruited women (Utrecht and Florence) or a substantial proportion of vegetarians (Oxford), which could hamper the external validity of their findings. In addition, the population in the manuscript has a 3:2 women:men ratio, although the authors have also included a complete adjusted analysis by gender, and the results were similar for both men and women. All in all, the conclusions of Lassale et al. are relevant: obesity increases the risk of coronary heart disease regardless of metabolic status (Figure 1). Metabolic abnormalities add risk both to normal weight and overweight/obese individuals. In other words, rather than ‘MHO’ it is ‘metabolically healthy not so healthy obesity’. In addition, obesity increases the risk for other risk factors such as hypertension, type 2 diabetes mellitus, and to cardiac conditions such as coronary heart disease or atrial fibrillation. This means that studies with longer follow-up may find further support the CV risk associated with obesity, and will likely corroborate the concept raised by the Whitehall study where MHO is only one step in the process of obesity.11 The most cost-effective way to tackle this epidemic is clearly at a population level, with an urgent need for the design of prevention strategies to decrease its prevalence rather than the development of strategies to treat its consequences. What this study adds: obesity has prognostic implications regardless of its metabolic status.