Lack of evidence for obesity paradox in patients with cardiovascular disease: A UK BioBank cohort study

Abstract Aims The obesity paradox has been controversial and is mostly observed when body mass index (BMI) is used. We aimed to examine the presence of the obesity paradox using body fatness (body fat percentage, BF%) and central obesity (waist‐to‐hip ratio adjusted for BMI, WHRadjBMI). We also used Mendelian randomisation to explore causality for the associations of BMI, BF% and WHRadjBMI with cardiovascular disease (CVD) mortality. Materials and Methods A total of 85 926 participants with CVD from the UK Biobank were included. Prospective associations of BMI, BF% and WHRadjBMI with CVD mortality in these patients were examined. Polygenic risk scores (PRSs) for BMI, BF% and WHRadjBMI were used as instrumental variables in Mendelian randomisation analyses. Results A total of 5432 patients died of CVD causes during a median follow‐up period of 13.6 years. BMI in the overweight and class‐I obesity ranges was associated with reduced mortality, with class‐II or more severe obesity associated with increased mortality; however, there was a linear trend toward increased mortality with increasing BF% and WHRadjBMI. There was no clear indication that increased obesity‐PRSs were associated with reduced risk of CVD mortality among patients with known CVD. Sensitivity analyses by sex, age group and disease type, and by using a single variant from the FTO gene rs1558902 as an instrumental variable showed similar results. Conclusion Increased obesity does not show a protective effect in patients with CVD. Previously reported obesity paradox in observational studies may be a result of confounding or other biases.