Baseline serum blood samples and demographic characteristics used in this cross-sectional study were collected as part of the Cardiovascular and cardiorespiratory Adaptations to Routine Exercise-based Cardiac Rehabilitation (CARE CR) study (29). The CARE CR study protocol was published in detail elsewhere (29). Briefly, clinically stable people with a primary diagnosis of CHD (aged 30–85 years) were referred to the research team by nursing staff, within 2 weeks of a cardiac event or procedure. Participants provided their written informed consent to participate in the study. The CARE CR study was granted ethical approval by the Humber Bridge NHS Research Ethics Committee-Yorkshire and the Humber (12/YH/0278). Ethical approval for assay analysis of serum samples for biomarkers related to sarcopenia was provided by the Northumbria University Health and Life Sciences Ethics Committee (20933). The main findings from the CARE CR study on patient rehabilitation and cardiorespiratory fitness are published elsewhere (4, 30).

Body mass index (BMI; kg m⁻²) was calculated using mass (kg) and stature (m). Waist and hip circumferences (cm) were measured at 1 cm above the iliac crest and at the widest aspect of the hips, respectively. Appendicular lean mass (ALM), defined as total lean mass in both arms and legs (kg), was measured using dual X-ray absorptiometry (DXA; Lunar iDXA GE Healthcare Buckinghamshire, United Kingdom), as a proxy for SMM assessment. ALM is expressed as skeletal muscle index (SMI; kg m⁻²) and as a percentage of total body mass (appendicular skeletal mass; ASM%). Age-adjusted SMI and ASM% were moderately correlated (r = 0.507, p < 0.001). We defined low SMI as <7.0 and <6.0 kg m⁻² (22) and low ASM% as <25.72 and <19.43% (31) for men and women, respectively.

Cardiopulmonary exercise testing was performed using the modified Bruce treadmill protocol (32), as previously described (4, 29). A 12-lead Electrocardiogram (ECG), ECG-gated automated blood pressure, heart rate, and rate of perceived exertion were monitored throughout. Breath-by-breath metabolic gas exchange data were collected using an Oxycon Pro metabolic cart (Jaeger, Hoechburg, Germany). We report V̇O_2peak (ml), defined as the mean V̇O₂ over the last 30 s of the test; V̇O_2peak was adjusted for body mass (ml kg⁻¹ min⁻¹) (4).

Participants abstained from strenuous exercise 24-h prior to attending their baseline study visit. Resting blood samples were drawn by venepuncture and placed in a refrigerated (4°C) centrifuge at 3,000 revolutions per minute, for 15 min. Albumin, aminotransferases and N-terminal pro-brain natriuretic peptide (NT-proBNP) were analysed at the Hull Royal Infirmary in an accredited biochemistry laboratory, as a single measurement on the day of each blood draw. Calibration and quality controls were conducted in accordance with manufacturer’s guidelines. The ABX Pentra 400 biochemistry auto analyser (Horiba, Montpellier, France) was used to analyse high sensitivity C-reactive protein (hs-CRP) in duplicate, in accordance with the manufacturer’s quality control guidance (4). Remaining plasma and serum samples were stored at −80°C until analysis.

We analysed serum samples in duplicate using commercial enzyme-linked immunosorbent assay (ELISA) for CAF (Abcam #ab216945) and transthyretin (Abcam #ab108895) and followed their standard instructions for serum analysis. Concentrations of transthyretin and CAF were assessed in duplicate and the average of the two measures reported. We re-analysed samples with a coefficient of variation (CV) >40% and when biomarker concentrations were not within the limits of the standard curve. The CV for the assay analyses of transthyretin and CAF were 7.9 and 5.1%, respectively. Routine health-related serum biomarkers evaluated as part of the CARE CR study are reported elsewhere, including NT-proBNP, hs-CRP, glucose, white cell count, total cholesterol, low-density and high-density lipoprotein cholesterol, estimated glomerular filtration rate and triglycerides (4, 30).

Normal adult reference values for circulatory markers of interest are:

Statistical analyses were performed by a single researcher using commercially available software (SPSS version 28, IBM, New York, NY, United States). Distribution of the data was assessed using visual inspection of histograms, QQ-plots and using the Kolmogorov Smirnov test. Categorical variables are reported as frequency with percentage. Continuous normally distributed variables are reported as mean ± standard deviation. Continuous non-normally distributed variables are reported as median with interquartile range, or median with range where the sample size is ≤3 people. Demographic characteristics are reported for the whole cohort and separately for people with normal or low SMM (defined as low SMI or low ASM%). Differences in demographic characteristics between the two groups were assessed using the Fisher’s exact test (categorical variables), a Student’s t-test (continuous normally distributed), or Mann–Whitney U test (continuous non-normally distributed). Two-group comparison of blood biomarkers between people with normal or low SMM were evaluated using Mann–Whitney U tests and reported as U statistics, p-values, and effect sizes, calculated using the following equation (36):

Absolute r values of 0.2, 0.5, and 0.8 are considered small, moderate and large effect sizes, respectively (37). The relationship between serum biomarker concentrations, SMI and ASM% were calculated using Spearman’s rank correlations. It is well-established that age and inflammation influence SMM and some serum biomarkers; people with CHD and low SMM are significantly older than those with normal SMM (38), whilst albumin and transthyretin concentrations decrease in the presence of inflammation (39). Accordingly, we also report non-parametric partial correlations adjusted for age and circulatory hs-CRP concentrations, both separately and together. An r value of <0.3, 0.3–0.5, 0.6–0.8, and >0.8 indicated a poor, fair, moderately strong and very strong associations, respectively (40). Scatterplots of associations between SMI and circulatory markers were plotted with linear regression lines. Where a marker was associated with SMI or ASM% or had a significant effect size for low and normal SMM groups, receiver operating characteristic (ROC) curves were used to investigate the sensitivity and specificity of predicting low SMM as the dichotomous ‘state variable’. We report the area under the curve (AUC) with 95% confidence interval (CI) and p-values. The AUC value was interpreted as follows: perfect (1.0), excellent (0.9–0.99), good (0.8–0.89), fair (0.7–0.79), poor (0.51–0.69), and no value (0.5) (41). The biomarker concentration cut-off points for prediction of low SMM were selected based on the highest combination of sensitivity and specificity values. We plotted ROC curves and determined biomarker cut-off points for the whole cohort and then separately for men only. Due to a small sample, ROC curves could not be plotted for women only. Statistical significance was set at p < 0.05.

The prognostic value of transthyretin, ALT, AST, and the AST/ALT ratio for identification of low SMM was assessed using ROC curve analysis. Including all participants, transthyretin (AUC 0.739, 95% CI 0.601, 0.876, p = 0.007) and ALT (AUC 0.731, 95% CI 0.576, 0.887, p = 0.009) had the greatest predictive capacity to identify low SMM. The AUC for AST level was 0.684 (95% CI 0.516, 0.851, p = 0.037) and non-significant for the AST/ALT ratio (AUC 0.636, 95% CI 0.482, 0.790, p = 0.123). The optimal cut-off points to indicate risk of low SMM were: a transthyretin value of ≤37.7654 mg/dl (sensitivity 0.857, specificity 0.520), an ALT value of ≤25.00 U/L (sensitivity 0.857, specificity 0.620), and an AST value of ≤24.50 U/L (sensitivity 0.714, specificity 0.620).

Including men only, ROC curve analyses showed the predictive capacity of transthyretin (AUC 0.773, 95% CI 0.603, 0.943, p = 0.002), ALT (AUC 0.699, 95% CI 0.509, 0.888, p = 0.040) and the AST/ALT ratio (AUC 0.693, 95% CI 0.538, 0.847, p = 0.014). The AUC for AST level was non-significant (AUC 0.560, 95% CI 0.357, 0.764, p = 0.562). In men, the optimal cut-off points to indicate risk of low SMM where: a transthyretin value of ≤30.3284 mg/dl (sensitivity 0.778, specificity 0.723), an ALT value of ≤25.00 U/L (sensitivity 0.889, specificity 0.660), and an AST/ALT ratio of ≥0.9347 (sensitivity 0.778, specificity 0.553).

Albumin is a marker of inflammation-related nutritional risk (20). In agreement with previous studies involving people with liver cirrhosis (46), end-stage renal disease (47) and heart failure (48), we found no association between albumin levels and SMM indices in people with CHD. Interestingly, others report both lower (49–51) and or higher (52) albumin concentrations in older adults with low SMM, compared to those with preserved SMM. The use of albumin levels to infer protein energy malnutrition was previously commonplace in clinical practise (53). Given that lean mass reflects the somatic protein store, the assumption followed that albumin might be useful as a marker of lean mass. However, the use of albumin as a biomarker of malnutrition or body composition has not been without criticism (20, 54). The literature lacks consensus on the existence and/ or direction of the association between albumin and SMM-related variables (46–49, 51, 52), likely due to the role of albumin as an acute-phase response protein.

The inflammation-induced reduction in albumin concentration is underpinned by: decreased albumin synthesis during stress response to prioritise synthesis of essential proteins, increased capillary permeability prompting a shift of albumin from the intravascular to the interstitial space, and a shortened albumin half-life resulting from tissue catabolism (20). In older adults, serum albumin is inversely associated with common inflammatory cytokine, CRP (55). We found no difference in hs-CRP between people with normal or low SMM (Table 2). This could explain the similar albumin levels between groups. Additionally, Chen et al. (56) speculated that sex-specific hormones levels might also impact the association between SMM and albumin levels, after finding these variables to be positively associated in men and negatively associated in women. However, our study included a small sample of women, and we were unable to investigate this hypothesis.

Transthyretin levels were significantly lower in people with low versus normal SMM. Similar to albumin, transthyretin is a marker of inflammation-related nutritional risk (20), a key component of malnutrition related to acute or chronic disease (57). Amino acid availability, from dietary protein intake, was proposed to mediate the relationship between transthyretin and lean mass (34). This is because amino acid ingestion promotes lean tissue accretion (58) and also modulates transthyretin synthesis in the liver (59). A strong, positive association (r = 0.58) between transthyretin levels and SMI was previously reported in people at a geriatric outpatient hospital (51). Around 40% of people in the study by Sergi et al. (51) were underweight (BMI <20 kg m⁻²). Poorer nutritional status likely contributed to a more pronounced inflammatory environment and lean mass loss in this study (51), potentially explaining the strong association between transthyretin and SMI, compared to a non-significant association in the present study (r = 0.246, p = 0.05). Nevertheless, our detection of significantly lower transthyretin levels in people with low compared to normal SMM is a promising finding, as it become increasingly apparent that transthyretin assessment might have clinical utility as part of a comprehensive medical evaluation (60).

Assessment of liver enzymes ALT and AST is routine in clinical practise (61). As a catalyst in the alanine-glucose cycle, ALT converts pyruvate to amino acid alanine in skeletal muscle and converts alanine back to pyruvate (for glucose production) in the liver (62, 63). A similar cycle is catalysed by AST, where the amino acid and product are aspartate and oxaloacetate, respectively (63). Circulatory levels of ALT and AST are elevated in Type 2 diabetes (24) and metabolic syndrome (25), conditions characterised by insulin resistance and hepatic steatosis. We, and others, demonstrate that ALT levels appear to be lower in the presence of low SMM (26, 50). Contrastingly, in a cross-section of >12,000 adults without liver-related disorders, ALT levels were elevated in those with low SMM compared to normal SMM (64). The direction of the relationship between AST and SMM is similarly contested. We found lower AST concentrations in people with low SMM compared to normal SMM. Others report that low SMM coincided with higher AST concentrations in people with (65, 66) and without (64) liver disease.

Multiple factors likely influence the inconsistency in these findings. First, damaged liver cells release ALT and AST into circulation, explaining their higher serum concentrations in people with liver disorders (67). Secondly, participants with low SMM in the study by Yoo et al. (64) were more often obese with higher fasting blood glucose and insulin levels compared to the normal SMM group, consistent with the theory that aminotransferase levels are elevated in the presence of higher metabolic risk. In the present study, people with reduced SMM had higher average body fat and comparable BMI to people with normal SMM. It could be speculated that differences in intra-abdominal and intra-hepatic steatosis, together with diet quality/alcohol consumption might have influenced aminotransferase concentrations.

Additionally, both ALT and AST require vitamin B₆ as a cofactor, meaning that vitamin B₆ deficiency might contribute to low circulatory ALT and AST (68). Furthermore, vitamin B₆ is mostly stored in striated muscle (69); thus, where lean mass is reduced a smaller pool of vitamin B₆ is available to act as a cofactor for AST and ALT. An estimated 31 and 24% of community-dwelling men and women (≥65 years) are at risk of inadequate vitamin B₆ dietary intake (70). Although not assessed in this study, addressing any dietary deficiencies in people with CHD and low SMM should be prioritised.

Studies involving older adults (17, 71, 72), people with lung disease (72, 73) and with heart failure (19, 73) have reported an association between high circulatory CAF levels and low SMM. This association is proposed to originate from degeneration of the neuromuscular junction with ageing. Agrin is cleaved by neurotrypsin during normal neural development (15). Excessive agrin cleavage from over-expression of neurotrypsin causes agrin to become deactivated and the neuromuscular junction to break down (16). The product of this breakdown, CAF, is released into the circulation (74). However, the effect of degeneration and remodelling of the neuromuscular junction on SMM loss is debated, with polarising studies arguing that this process contributes to (75) or is protective against (76) muscle atrophy.

We found no association between CAF levels and SMM indices in people with CHD. Others have reported similar non-significant findings when assessing possible associations between CAF and presence of frailty in people with CHD, although an assessment of SMM was not included in their definition of frailty (77). Sánchez-Castellano et al. (78) found no difference in CAF levels between low and normal SMM groups with hip fracture and suggested that elevated CAF levels in both groups indicated neuromuscular degeneration was present in both. In contrast, median CAF values in the low and normal SMM groups were within the normal limits in the present study (0.86–4.66 ng/ml; 17), suggesting that circulatory CAF has limited utility as biomarker for low SMM in this cohort.

We assessed, in a secondary analysis, multiple proposed biomarkers for low SMM in people with CHD, contributing to our understanding of the factors influencing this complex and under-researched pathology. We included assessment of four biomarkers which are already commonly assessed in clinical practise (albumin, transthyretin, ALT, and AST), aiding the potential transition of our findings into practise.

This study is potentially limited by our use of DXA-derived lean mass to estimate SMM. DXA assessment is the current reference standard, but is limited by the production of variability related to different devices and software versions (79) and the absence of a universally agreed cut-off point for low SMM (80). Furthermore, DXA derived lean mass can be interpreted in several ways (i.e., corrected for stature, body mass or body fat percentage), which often produce conflicting findings when analysed in relation to circulatory biomarkers. This might limit the comparability of our findings with other, similar research. Finally, we included a small sample of women and there was no assessment of muscle strength or function.

Future research should evaluate the association between albumin, transthyretin, aminotransferases, CAF and measures of muscular strength alongside SMM. Whether these markers change with targeted lifestyle interventions also requires investigation. Additionally, there appears to be sex differences in median biomarker concentrations and their correlations with SMM indices, although our small sample of women limits the certainty of this finding. Future research might further investigate sex differences in SMM biomarkers in people with CHD.