Sarcopenia Amplifies Risk of Adverse Cardiovascular Outcomes in Individuals with Left Ventricular Hypertrophy 

​​​Authors

Mihir M. Sanghvi1,2, Hannah L. Nicholls1, Thomas M. Kaplan1, Ross J. Thomson1,2, Vijay Shyam-Sundar1,2, Mohammed Y. Khanji1,2, Steffen E. Petersen1,2, James A. Timmons1, Patricia B. Munroe1*, Nay Aung1,2* 

​

1. William Harvey Research Institute, NIHR Barts Biomedical Research Centre, Queen Mary University of London, Charterhouse Square, London EC1M 6BQ, United Kingdom 

2. Barts Heart Centre, Barts Health NHS Trust, West Smithfield, London EC1A 7BE, United Kingdom 

 

​*denotes equal contribution 

​​​

Introduction: Sarcopenia is the reduction in skeletal muscle mass and function, and is associated with a broad range of adverse health outcomes but its quantification in clinical practice is challenging. Left ventricular hypertrophy (LVH), the presence of increased myocardial mass, is a well-established biomarker for cardiovascular disease. This study aims to establish if the presence of sarcopenia in the presence of LVH amplifies risk of developing adverse cardiovascular outcomes. 

​

Methods: We developed a deep learning pipeline to quantify pectoralis major muscle mass from 77,586 cardiovascular magnetic resonance (CMR) examinations in the UK Biobank. Sarcopenia was defined as the sex-specific lowest decile of pectoralis major muscle mass indexed to height², and LVH as the highest sex-specific decile of LV mass indexed to body surface area. Using multivariable Cox models adjusted for demographic, lifestyle, metabolic, renal, and imaging covariates, we examined associations with incident heart failure, atrial fibrillation, composite CVD, cardiovascular mortality, and all-cause mortality, excluding individuals with prior CVD.  

​

Results: Over a median follow-up of 4.8 years, the coexistence of sarcopenia and LVH conferred the highest risk across all outcomes, with hazard ratios (HR) 1.3x to 3x higher when compared to the presence of LVH or sarcopenia alone. The combined phenotype showed markedly elevated hazards for heart failure (HR 4.26, 95% CI 2.61–6.95), atrial fibrillation (HR 2.06, 95% CI 1.26–3.37), composite CVD (HR 2.55, 95% CI 1.71–3.82) and cardiovascular death (HR 6.54, 95% CI 2.66–16.07) and all-cause mortality (HR 2.91, 95% CI 1.65–5.13).  

 

Discussion: The presence of sarcopenia appears to amplify the deleterious prognostic impact of LVH. A number of conditions which are followed-up in inherited cardiovascular disease clinics are characterised by hypertrophy, the majority of whom will have undergone cross-sectional imaging. This work highlights the importance of sarcopenia identification in risk stratification and provides an automated tool with which to achieve this from routine imaging. 

 

Extended Text 

 

Introduction

Sarcopenia is the reduction in skeletal muscle mass and function, and is associated with a broad range of adverse health outcomes including cardiovascular disease events.1 Left ventricular hypertrophy (LVH), the presence of increased myocardial mass, is a well-established biomarker for cardiovascular disease.2 

 

Increasing degrees of LVH have been shown to predict higher rates of incident of clinical events (heart failure, stroke, myocardial infarction) and cardiovascular death in both population settings and individuals with diseases characterised by LVH such as hypertrophic cardiomyopathy.3 However, it remains unclear whether the coexistence of sarcopenia amplifies the risk conferred by LVH. 

 

A major limitation in the incorporation of sarcopenia assessment are in the common modalities used to assess sarcopenia: bioelectrical impedance analysis (BIA) and dual energy x-ray absorptiometry (DEXA). They are not routinely performed in patients with cardiometabolic disease, only indirectly estimate muscle mass and furthermore are subject to a number of limitations and assumptions such as hydration status and fat distribution which limit their precision and accuracy.4,5  By contrast, cross-sectional imaging of the chest is one of the most commonly performed radiological investigations within healthcare systems.6  

 

Here, we have developed a deep learning algorithm to perform direct, three-dimensional measurement of pectoralis major muscle mass from cardiovascular magnetic resonance (CMR) examinations. We hypothesise that low muscle mass amplifies the risk of adverse cardiovascular outcomes among individuals with LVH. In this study, we examine the independent and combined associations of sarcopenia and LVH with heart failure (HF), atrial fibrillation (AF), a composite of CVD events, cardiovascular death, and all-cause mortality in 76,586 individuals in the UK Biobank. Our aim was to determine whether presence of sarcopenia in combination with LVH identifies individuals who are at particularly high risk, which could have implications for risk stratification and preventive strategies. 

 

Methods 

Deep learning segmentation algorithm

We developed a deep learning pipeline to measure pectoralis major muscle mass from the axial stack of UK Biobank CMR examinations. Ground truth comprised of manual, three-dimensional segmentation of pectoralis major in 200 cases (>2,100 slices). Of these 200 cases, participants were partitioned into training (n=150) and test (n=50) sets. We implemented nnU-Net (v2) in both two dimensional (2D) and three-dimensional (3D) full-resolution U-Net architectures.7 In the evaluation of 50 independent test studies the highest performance was achieved by the 2D U-Net ensemble incorporating uncertainty estimation, with a mean Dice similarity coefficient of 0.832 (standard deviation 0.050; median 0.835; range 0.680–0.923), surpassing the single-fold 2D model and 3D models. Exemplar segmentation is shown in Figure 1.  

 

Definition of exposures, covariates, and outcomes

Pectoralis major muscle mass (g) was indexed to height2. Sarcopenia was defined as the sex-specific lowest decile of indexed values. LV mass as measured from CMR was indexed to body surface area (BSA). Individuals categorised into the LVH group were from the highest sex-specific decile for LV mass. Covariates included age, sex, ethnicity, body mass index, Townsend deprivation index, systolic blood pressure, haemoglobin A1c, estimated glomerular filtration rate, smoking status, alcohol intake, physical activity level and indexed LV end-diastolic volume. Continuous variables were scaled to a mean of 0 and standard deviation of 1. We examined five pre-specified outcomes: incident heart failure, incident atrial fibrillation, composite atherosclerotic CVD events (myocardial infarction, stroke, coronary revascularisation), cardiovascular mortality and all-cause mortality which were ascertained through linkage with hospital episode statistics. Individuals with pre-existing CVD were excluded. 

 

Statistical analysis 

Cox proportional hazards models to estimate hazard ratios for each outcome. Proportional hazards assumptions were checked and found to be satisfied. Three models were constructed for each outcome: Model 1 included sarcopenia (low muscle mass) as the exposure of interest; Model 2 included LVH (high LVM) as the exposure; Model 3 examined the combined effect of sarcopenia and LVH, by including a variable identifying those with both conditions. Each model was adjusted for all covariates listed above. 

 

Results 

Mean age was 65.5 years (SD 7.8) and 51.7% were female. The association of sarcopenia, LVH and a combined sarcopenia and LVH group are presented in Figure 2. At a median follow-up time of 4.8 years, the combination of sarcopenia and LVH conferred the highest risk of adverse outcomes following adjustment for conventional cardiovascular risk factors. 

 

Individuals with both sarcopenia and LVH had a markedly higher risk of incident heart failure (HR 4.26, 95% CI 2.61–6.95), exceeding the risk associated with LVH alone (HR 2.68, 95% CI 2.03–3.54) or sarcopenia alone (HR 2.26, 95% CI 1.69–3.01). The combined phenotype remained the strongest predictor of AF (HR 2.06, 95% CI 1.26–3.37), relative to LVH alone (HR 1.54, 95% CI 1.25–1.90) and sarcopenia alone (HR 1.33, 95% CI 1.08–1.64). Risk was highest for the combined phenotype for the composite CVD endpoint (HR 2.55, 95% CI 1.71–3.82), compared with LVH alone (HR 1.58, 95% CI 1.36–1.84) and sarcopenia alone (HR 1.21, 95% CI 1.03–1.41). 

 

The greatest relative hazard was observed for CV death with coexistent sarcopenia and LVH (HR 6.54, 95% CI 2.66–16.07). Both sarcopenia (HR 2.12, 95% CI 1.21–3.71) and LVH (HR 2.43, 95% CI 1.38–4.25) were also independently associated with higher CV mortality. The combined phenotype again showed the highest risk for all-cause mortality (HR 2.91, 95% CI 1.65–5.13), followed by sarcopenia alone (HR 1.97, 95% CI 1.59–2.45). LVH alone was not statistically significant (HR 1.25, 95% CI 0.94–1.65). 

 

Discussion: 

In this large cohort, we have developed an automated way to derive a measure of sarcopenia from individuals with CMR examinations, a common test performed in inherited cardiovascular conditions. We found that sarcopenia and LVH each independently increase the risk of adverse cardiovascular outcomes. Importantly, the coexistence of the two conditions identifies a subset of individuals at markedly elevated risk with sarcopenia appearing to amplify the deleterious prognostic impact on LVH. A number of conditions which are followed-up in inherited cardiovascular disease clinics are characterised by hypertrophy and this work highlights the importance of sarcopenia identification, in particular given that it is potentially modifiable with targeted lifestyle and therapeutic strategies. 

​

​

References 

1. Benz E, Pinel A, Guillet C, et al. Sarcopenia and Sarcopenic Obesity and Mortality Among Older People. JAMA Network Open. 2024;7:e243604. 

​

2. Bluemke DA, Kronmal RA, Lima JAC, et al. The relationship of left ventricular mass and geometry to incident cardiovascular events: the MESA (Multi-Ethnic Study of Atherosclerosis) study. Journal of the American College of Cardiology. 2008;52:2148–55. 

​

3. Spirito P, Bellone P, Harris KM, Bernabo P, Bruzzi P, Maron BJ. Magnitude of left ventricular hypertrophy and risk of sudden death in hypertrophic cardiomyopathy. N Engl J Med. 2000;342:1778–1785. 

​

4. Lee SY, Ahn S, Kim YJ, et al. Comparison between Dual-Energy X-ray Absorptiometry and Bioelectrical Impedance Analyses for Accuracy in Measuring Whole Body Muscle Mass and Appendicular Skeletal Muscle Mass. Nutrients. 2018;10:738. 

​

5. Buckinx F, Landi F, Cesari M, et al. Pitfalls in the measurement of muscle mass: a need for a reference standard. Journal of Cachexia, Sarcopenia and Muscle. 2018;9:269–278. 

​

6. Statistics. Statistics » Diagnostic Imaging Dataset Accessed September 1, 2025. https://www.england.nhs.uk/statistics/statistical-work-areas/diagnostic-imaging-dataset/. 

​

7. Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods. 2021;18:203–211. 

​

Contributions

 

I generated the hypothesis, performed the manual segmentation of the CMR examinations, assisted Nay Aung in setting up the deep learning pipeline, performed all of the statistical analyses and wrote up the work.