A substantial amount of studies in the last decade have determined the detrimental effects of overweight and obesity on the risk of myocardial infarction, stroke, cancer, and overall mortality. Nonetheless, there is also evidence in favour of the so-called “obesity-paradox” which states that overweight - or even obesity - exerts a protective role or has no impact on the incidence of cardiovascular disease (CVD) and overall mortality, because of increased cardiovascular mortality in the lowest BMI strata (J-curve).
In March 2018, a study was published by researchers from the University of Glasgow, challenging the obesity-paradox and heating up the debate of this much-discussed topic .
Aim of the study
Conflicting data exist regarding the associations of BMI with CVD risk – especially for those with a low BMI. The aim of the study was to examine the association between body measures and the incidence of CVD outcomes in healthy subjects.
Study design and participants
Approximately 300.000 participants from the UK Biobank were included. They were all of white European descent without pre-existing CVD at baseline. The UK Biobank is a large prospective study with about 500.000 participants (recruited between 2006 and 2010), aged 40-69 years, who consented for their records to be linked with research data (national hospital and death registries). Subjects attended one of 22 assessment centres across the UK. They completed a questionnaire and underwent physical measurements as well as blood sampling for biological samples.
Adiposity was assessed with in 5 different measures: BMI, waist circumference, waist-to-hip ratio, waist-to-height ratio and percentage body fat.
Primary outcomes were fatal and non-fatal CVD events. Secondary outcomes were CVD mortality, non-fatal CVD events, and a composite outcome of ischaemic heart and cerebrovascular events.
Very low BMI (≤18.5 kg/m2) was associated with higher incidence of CVD and the lowest risk of CVD was exhibited at BMI of 22-23 kg/m2. Thereafter, the incidence of CVD increased up to a BMI of 35 kg/m2 for men and 45 kg/m2 for women. This J-shaped association of the BMI attenuated drastically in subgroup analyses, when participants with comorbidities where excluded, and when data were adjusted for non-smokers (men).
Notably, for the remaining adiposity measures the associations were all linear, implicating that higher adiposity per se was associated with greater risk of CVD events.
In detail, one standard deviation (SD) increase in waist circumference (12.6 cm for women and 11.4 cm for men) was associated with a hazard ratio (HR) of 1.16 (95% CI 1.13–1.19) for women and 1.10 (95% CI 1.08–1.13) for men for CVD events. Associations for 1 SD of the other adiposity measures were similar.
The J-shaped associations of BMI with the incidence of CVD almost disappears in subgroup analysis in subjects without comorbidities or non-smoking men. The associations of the remaining adiposity measures were linear and unchanged after these adjustments. This may suggest that the adverse effects of a low BMI on the risk of CVD is rather a result of confounding by comorbidities. The other studied adiposity measures (waist circumference, waist-to-hip ratio, waist-to-height ratio and percentage body fat) appeared less susceptible to this bias, as these are not so much influenced by muscle mass loss and better reflect fat mass.
The study confirms that being overweight or obese has a detrimental impact on health by increasing the risk of CVD in middle-aged men and women.