Introduction Obesity is a rising public health crisis and a major risk factor for heart failure (HF). However, underlying mechanisms are incompletely understood. In this study, we investigate longitudinal associations of obesity with incident HF and cardiovascular imaging phenotypes in the UK Biobank (UKB). Importantly, we use cardiovascular magnetic resonance (CMR) to investigate potential mechanisms driving the obesity-HF relationships.

Methods The UKB cohort comprises over half a million individuals recruited from across the UK between 2006–2010. The UKB imaging study, which includes CMR, commenced in 2015 with plan to scan a random 20% subset of the original cohort. We defined obesity using body mass index (BMI) and waist-to-hip-ratio (WHR) measured at baseline recruitment. Incident HF events were identified through linked Hospital Episode Statistics data (censor date December 2021). CMR scans were analysed using an automated pipeline. We used Cox proportional hazard regression models adjusted for potential confounders to estimate the associations of BMI and WHR with incident HF in the whole sample. We used linear regression to characterise obesity (BMI, WHR) associations with CMR phenotypes. Finally, we used multiple mediation analysis to define the role of obesity-related cardiac remodelling in driving its associations with incident HF, independent of cardiometabolic diseases (diabetes, hypertension, high cholesterol).

Results In 491,606 UK Biobank participants (mean age 56.6 years, 54.3% women) over 12.2±0.9 years of prospective follow-up, higher BMI [HR 1.34 (1.32, 1.37)] and WHR [HR 1.30 (1.30, 1.36)] were associated with a greater hazard of HF. In the subset of participants with CMR (n=31,107), greater obesity was associated with adverse left ventricular (LV) structure (higher LV mass, greater concentricity), poorer LV global functional index, and lower myocardial native T1. In multiple mediation analysis, hypertension had an important role in mediating the associations of obesity with incident HF. Adverse LV remodelling (higher LV mass, greater concentricity) were also major mediators of the obesity-HF associations, independent of cardiometabolic disease. Notably, higher native T1 (indicated greater myocardial fibrosis) mediated a significant fraction of the relationship between obesity and incident HF. Given that in the whole cohort greater obesity was related to lower native T1, this observation may indicate different stages of progression in obesity-related cardiac remodelling.

Conclusion Our findings demonstrate the association of obesity with a greater risk of HF and adverse alterations of LV structure, function, and myocardial character. Importantly, we highlight the role of specific adverse cardiovascular remodelling patterns in driving obesity-HF associations. Thus, our findings highlight the growing public health importance of obesity as a driver of HF and propose novel channels for dedicated mechanistic research.

Conflict of Interest None