Increased cardiovascular risk and cardiac dysfunction in cancer survivors from UK Biobank

Abstract Background Cancer survivors have an elevated risk of cardiovascular disease (CVD). Cardiovascular magnetic resonance (CMR) offers detailed assessment of cardiac structure and function. Few researchers have examined the long-term health of cancer survivors on a population level or incorporated large-scale CMR imaging. Purpose This study leverages linked health records and CMR data from the UK Biobank to evaluate 1) excess risk of specific CVDs and 2) adverse CMR remodelling in cancer survivors, compared to matched controls. Methods Participants with a record of cancer (20 cancer subtypes) prior to baseline recruitment were identified using linked hospitalisation and cancer registry data. Individuals with codes for neoplasms of uncertain behaviour and non-melanoma skin cancer were excluded. Each cancer-exposed participant was propensity score-matched to two participants without cancer. Matching was performed on an extensive range of sociodemographic, lifestyle, baseline morbidity, and clinical biomarkers. Missing covariates were imputed using predictive mean matching. Cox regression was used to assess hazard ratios for the following CVD outcomes: chronic ischaemic heart disease, non-ischaemic cardiomyopathy, heart failure, myocardial infarction, atrial fibrillation, pericardial disease, and venous thromboembolism. Incident outcomes were prospectively ascertained from linked hospital and death registry records over a median of 13.6 years. In participants with CMR data available, linear regression was used to examine the association of cancer exposure with CMR-derived metrics of cardiovascular structure, function, and myocardial tissue character. These included ventricular and atrial volumes and function, strain parameters, myocardial T1, and arterial stiffness. Results A total of 36,175 cancer survivors and 72,354 controls were identified. The cohort comprised 70% men and 30% women with median age of 61 years [IQR: 54-65]. The three most common cancers were breast cancer (25%), colorectal cancer (6.6%), and melanoma (5.6%). Cancer survivors had an increased incident risk for almost all selected CVDs (Figure 1A), with the highest risk observed in relation to pericardial disease [HR=2.10, 95% CI: 1.75-2.51] and venous thromboembolism [HR=1.74, 95% CI: 1.62-1.86]. CMR revealed unhealthy alterations in cardiac structure and function in cancer survivors, including significantly increased LV volumes, decreased LV ejection fraction, poorer right and left ventricular function, and greater myocardial fibrosis (higher T1) (Figure 1B). Conclusion Cancer history is associated with increased incidence CVD risk and pathological cardiac remodelling patterns. These findings highlight the promise of CMR for early disease detection and the importance of improving cardiovascular health of cancer survivors through routine monitoring and long-term prevention strategies.