Impact of whole-body and skeletal muscle composition on peak oxygen uptake in heart failure: a systematic review and meta-analysis

Abstract

Aims

Heart failure (HF) has a major impact on exercise tolerance that may (in part) be due to abnormalities in body and skeletal muscle composition. This systematic review and meta-analysis aims to assess how differences in whole-body and skeletal muscle composition between patients with HF and non-HF controls (CON) contribute to reduced peak oxygen uptake (VO₂peak).

Methods and results

The PubMed database was searched from 1975 to May 2024 for eligible studies. Cross-sectional studies with measures of VO₂peak, body composition, or muscle biopsies in HF and CON were considered. Out of 709 articles, 27 studies were included in this analysis. Compared with CON, VO₂peak [weighted mean difference (WMD): −9.96 mL/kg/min, 95% confidence interval (CI): −11.71 to −8.21), total body lean mass (WMD: −1.63 kg, 95% CI: −3.05 to −0.21), leg lean mass (WMD: −1.38 kg, 95% CI: −2.18 to −0.59), thigh skeletal muscle area (WMD: −10.88 cm² , 95% CI: −21.40 to −0.37), Type I fibres (WMD: −7.76%, 95% CI: −14.81 to −0.71), and capillary-to-fibre ratio (WMD: −0.27, 95% CI: −0.50 to −0.03) were significantly lower in HF. Total body fat mass (WMD: 3.34 kg, 95% CI: 0.35–6.34), leg fat mass (WMD: 1.37 kg, 95% CI: 0.37–2.37), and Type IIx fibres (WMD: 7.72%, 95% CI: 1.52–13.91) were significantly higher in HF compared with CON. Absolute VO₂peak was significantly associated with total body and leg lean mass, thigh skeletal muscle area, and capillary-to-fibre ratio.

Conclusion

Individuals with HF display abnormalities in body and skeletal muscle composition including reduced lean mass, oxidative Type I fibres, and capillary-to-fibre ratio that negatively impact VO₂peak.

Graphical Abstract

Open in new tabDownload slide

Introduction

Decreased exercise tolerance, measured as reduced peak oxygen uptake (VO₂peak), is a hallmark feature in heart failure (HF) and is associated with a reduced quality of life (QoL) and an increased risk of hospitalizations and mortality.^1,2 The mechanisms underpinning the lower VO₂peak in HF are multifactorial and traditionally have focused on impaired cardiac function.³ Non-cardiac ‘peripheral’ factors (i.e. reduced lean mass, oxidative fibres, and capillarity) may also contribute to reduced exercise tolerance in HF.² However, due to inter-individual variability coupled with differences in physical activity levels between patients with HF and age-matched healthy controls, uncertainty remains regarding the magnitude of the decline in lean mass and skeletal muscle composition in HF and their contribution to the reduced VO₂peak in HF. To address this knowledge gap, we performed a systematic review and meta-analysis to examine how differences in whole-body and skeletal muscle composition contribute to the lower VO₂peak in HF compared with CON.

Methods

Data sources and search strategy

A PubMed literature search was conducted for English-language articles published between 1975 and May 2024. The search strategy was structured using the following three main terms: (i) heart failure; (ii) exercise tolerance, oxygen consumption, or VO₂peak; and (iii) skeletal muscle and body composition or biopsy. In addition, a manual search was conducted using the reference list of the included studies in this review using Google Scholar. Studies identified from the database search were exported in full sets and transferred into the Covidence review management software (Melbourne, Australia). All studies were initially screened by title and abstract and subsequently by full text if they met the relevant inclusion criteria. Data from all included studies were extracted and reviewed by two independent researchers (V.S. and D.W.). In the context of studies with duplicate data, the study with the most relevant information was included to avoid overlapping populations. No ethical approval for this study was necessary as all data were sourced from previously published studies and did not involve any personally identifiable information.

Study selection and inclusion and exclusion criteria

Only studies were included that: (i) compared subjects with HF [HF with reduced ejection fraction (HFrEF) or preserved ejection fraction (HFpEF)] with a non-HF control group (CON); (ii) reported VO₂peak measured from expired gas analysis; and (iii) measured whole-body composition [measured by dual-energy X-ray absorptiometry (DEXA), magnetic resonance imaging (MRI), or computed tomography (CT)] and/or skeletal composition by muscle biopsy. The following criteria were excluded: (i) no original or duplicate data; (ii) no control group without HF; (iii) non-human cohorts (i.e. animal models); (iv) non-English studies; (v) no peak exercise data; and (vi) no body composition (by DEXA, MRI, or CT) and/or no skeletal muscle biopsy data.

Study quality assessment

The quality of the included studies was evaluated by using the AXIS appraisal tool. The AXIS tool is a validated 20-point tool designed to assess the quality of cross-sectional studies (a maximum score of 20 indicates the highest quality).⁴

Data synthesis and statistical analysis

Differences, measured as effect sizes, in primary and secondary outcomes between HF and CON group were quantified using a meta-analysis based on the random-effects models. From the models, the weighted average effect size, defined as the weighted mean difference (WMD) with 95% confidence intervals (CIs), was calculated for each outcome between HF and CON. The weight assigned to each study was the inverse of the variance within study. The larger the sample size and smaller the variance of the study, the greater the assigned weight. Heterogeneity within individual effect sizes was calculated by I² and τ². Forest plots were generated to illustrate individual effect sizes, standard deviations, and the associated P-value for hypothesis testing at an alpha level of 0.05. In addition, to determine the association between VO₂peak and parameters of body and skeletal muscle composition, meta-regression analyses were performed using the R metacont package (R Core Team 2016, R Foundation for Statistical Computing, Vienna, Austria).

Results

The initial search yielded 709 articles. Following screening, 27 studies that met the inclusion criteria were included in the analysis (Figure 1). Overall, 844 patients with HF (mean age: 64 years, 77% male, 67% HFrEF, and 33% HFpEF) and 515 CON participants (mean age: 62 years and 75% male) were included (Table 1). Cardiopulmonary exercise testing (CPET) was performed on a cycle ergometer (n = 15 studies) or treadmill (n = 12 studies) to quantify VO₂peak. As detailed in Table 1, in addition to measuring VO₂peak (n = 27 studies), 19 studies also measured body composition, whereas 13 studies measured skeletal muscle morphology. Body composition was assessed with either DEXA (n = 16 studies), MRI (n = 2 studies), or CT (n = 3 studies), while muscle fibre composition was evaluated with biopsy from the m. vastus lateralis (n = 13 studies).

Peak oxygen uptake

Peak oxygen uptake was significantly and markedly lower in HF when measured in absolute values (WMD: −725.20 mL/min, 95% CI: −919.33 to −53.07 mL/min, I² = 89%, n = 899) or indexed to whole-body mass (WMD: −9.96 mL/kg/min, 95% CI: −11.71 to −8.21 mL/kg/min, I² = 87%, n = 1239) (Figure 2).

Body composition outcomes

Total body lean mass (WMD: −1.63 kg, 95% CI: −3.05 to −0.21 kg, I² = 35%, n = 938) and leg lean mass (WMD: −1.38 kg, 95% CI: −2.18 to −0.59 kg, I² = 54%, n = 675) were significantly lower (Figure 3A and C), while total body fat mass (WMD: 3.34 kg, 95% CI: 0.35–6.34 kg, I² = 78%, n = 862) and leg fat mass (WMD: 1.37 kg, 95% CI: 0.37–2.37 kg, I² = %, n = 527) were significantly elevated in HF compared with CON (Figure 3B and D). Thigh skeletal muscle cross-sectional area (WMD: −10.88 cm², 95% CI: −21.40 to −0.37 cm², I² = 64%, n = 268) was also significantly reduced in HF relative to CON (see Supplementary material online, Figure S1). Furthermore, the absolute VO₂peak (mL/min) was significantly associated with total body lean mass (HF: β = 68.52, P = 0.004, 95% CI: 22.34–114.70; CON: β = 76.17, P = 0.001, 95% CI: 29.31–123.03), leg lean mass (HF: β = 145.34, P = 0.001, 95% CI: 89.88–200.79; CON: β = 146.49, P = 0.008, 95% CI: 38.65–254.34) (Figure 4A and B), and thigh skeletal muscle cross-sectional area (HF: β = 8.21, P < 0.001, 95% CI: 3.83–12.60; CON: β = 27.96, P < 0.0001, 95% CI: 15.16–40.75). However, no association was found between VO₂peak and fat mass (see Supplementary material online, Table S1).

Skeletal muscle morphology

Compared with CON, HF displayed a significantly lower percentage of Type I muscle fibres (WMD: −7.76%, 95% CI: −14.81 to −0.71%, I² = 86%, n = 273) (Figure 5A). No significant difference was found between groups for Type IIa fibres (WMD: −2.74%, 95% CI: −10.59 to 5.11%, I² = 79%, n = 172), while the percentage of Type IIx fibres was significantly higher in HF compared with that in CON (WMD: 7.72%, 95% CI: 1.52–13.91, I² = 67%, n = 126) (see Supplementary material online, Figure S2). Additionally, the capillary-to-fibre ratio was markedly lower in HF than in CON (WMD: −0.27, 95% CI: −0.50 to −0.03, I² = 72%, n = 194) (Figure 5B) and also significantly associated with absolute VO₂peak in HF (HF: β = 584.21, P = 0.003, 95% CI: 60.92–394.85; CON: β = −100.86, P = 0.501, 95% CI: −394.85 to 193.14). There was no association between the percentage of muscle fibre types and absolute VO₂peak (see Supplementary material online, Table S1).

Study quality and risk of bias

The risk of bias of the included studies quantified by the AXIS tool was 13.5 points, indicating a moderate risk of bias. The main weaknesses of the included studies were the lack of sample size justification. Detailed AXIS scoring for each study is given in the supplementary material (see Supplementary material online, Table S2).

Discussion

The following are the major new findings of this systematic review and meta-analysis: (i) individuals with HF exhibit a significant and marked reduction in VO₂peak (∼10 mL/min/kg) compared with CON, and across all studies, the patients with HF displayed an average VO₂peak of 16 mL/kg/min, which is well below the 18 mL/kg/min VO₂ threshold required for full and independent living;³² (ii) total body and leg lean mass as well as thigh skeletal muscle area are significantly lower in patients with HF vs. CON and this shares a positive relationship to absolute VO₂peak; and (iii) m. Vastus lateralis Type I (oxidative) fibres and capillary-to-fibre ratio were significantly lower in patients with HF vs. CON and capillary-to-fibre ratio was positively associated with absolute VO₂peak in HF.

Body composition and exercise intolerance in heart failure

Our findings in the current novel meta-analysis illustrate that muscle atrophy is a major component of the HF syndrome. Specifically, we demonstrate that individuals with HF have reduced total and lower body lean mass and skeletal muscle area relative to CON. Moreover, our analysis also illustrates, across several influential studies, the repercussion of reduced lean mass for exercise intolerance in HF by showing a significant association between total body and leg lean mass with absolute VO₂peak. These findings align with studies in healthy CON that report a linear relationship between VO₂peak and lean mass, but not with total body mass.^33,34 Notably, our findings implicate that leg mass plays a critical role in limiting exercise tolerance as each kilogram increase in lean mass is associated with a concomitant increase in absolute VO₂peak of about 145 mL/min. Indeed, with a rise in lean mass, the amount of (aerobic) metabolically active tissue presumably increases and thereby the total capacity to utilize the delivered O₂ for oxidative phosphorylation-dependent generation of adenosine triphosphate for muscle contraction is enhanced.³⁵

These findings support the results from a meta-analysis of >1700 patients with HF showing a relatively high prevalence of sarcopenia (34%, 95% CI: 22–47%),³⁶ and it is also consistent with several reports showing that individuals with HF who have sarcopenia and/or decreased lean mass have decreased VO₂peak, reduced physical performance, and lower QoL.^37,38 Taken together, these findings highlight the importance of targeting skeletal muscle (an important component of lean mass) as a strategy to improve HF-related disability and clinical outcomes, particularly among patients with HF with additional risk factors for sarcopenia (i.e. older patients with HF).³⁹ The mechanisms driving the increased muscle atrophy may be related to features inherent to HF such as deconditioning, increased neurohumoral activation, disruption of insulin-like growth factor 1 signalling, pathological systemic and local immune responses, oxidative stress, alterations in protein signalling, and chronic inflammation.⁴⁰ Chronic inflammation is particularly facilitated by fat mass, as adipose tissue promotes the secretion of proinflammatory cytokines such as tumor necrosis factor-alpha and interleukin-6.⁴¹ These cytokines accelerate the breakdown of muscle proteins and inhibit protein synthesis in muscles, thereby contributing to muscle atrophy. In contrast, anti-inflammatory cytokines such as interleukin-10, cytokine inhibitors such as interleukin-1 receptor antagonist, so-called myokines (which are synthesized and released by muscle tissue during muscle contraction), counteract chronic inflammation.^42,43

Our results demonstrated that total body and leg fat were 3.3 and 1.4 kg higher, respectively, in HF vs. CON, highlighting not only a transition towards sarcopenia but also a sarcopenic obesity phenotype, which is associated with a poorer prognosis and increased functional limitations.⁴⁴ Nevertheless, our meta-regression analysis revealed no significant association between fat mass and VO₂peak. This finding may partly be due to collinearity between fat mass and lean mass,⁴⁵ where heavier individuals with HF may also have higher lean mass (and therefore greater absolute VO₂peak) due to the increased mechanical loading from greater fat and total body mass.³⁸ While increased fat mass may not directly reduce VO₂peak in absolute terms, the metabolic inefficiencies associated with higher fat mass mean that even basic limb movements against gravity require significantly more O₂ and a larger proportion of an individual’s VO₂peak. This highlights the importance of interventions targeting increases in muscle and decreases in fat mass, such as combined dietary (e.g. caloric restriction + increased dietary protein intake) and resistance-based exercise interventions.⁴⁶ This may be particularly important to maintain optimal skeletal muscle health and physical function when patients are undergoing medically induced weight loss with bariatric surgery or medications such as glucagon-like peptide-1 agonists, which have the potential to cause marked reductions in lean body mass in addition to their effects on fat mass.⁴⁷

Skeletal muscle morphology and exercise intolerance in heart failure

In addition to the quantity of lean mass, skeletal muscle quality is also seen as a key factor in exercise intolerance in HF.²⁴ Indeed, our meta-analysis indicates that the composition of skeletal muscle fibres in individuals with HF differs from CON. Specifically, the proportion of highly oxidative Type I muscle fibres is reduced by 7.8%, while the proportion of Type IIx (glycolytic) fibres is increased by 7.7% in HF compared with CON. The proportion of oxidative-glycolytic Type IIa fibres also tended to be slightly lower (2.7% lower) in HF. Type I fibres have a high oxidative capacity compared with Type II fibres, especially Type IIx fibres, owing to differences in mitochondrial content, oxidative enzyme activity, and capillarity.⁴⁸ Accordingly, a lower per cent of oxidative fibres would be expected to contribute to decreased VO₂peak, aerobic performance, and fatigue resistance.⁴⁹ Indeed, Bekedam et al.¹¹ noted that the cross-sectional area of highly oxidative fibres is relatively small compared with low oxidative fibres, indicating an inverse relationship between fibre cross-sectional area and VO₂peak.

Additionally, the capillarization of different fibre types is a major determinant of VO₂peak, which may be impaired in HF.^25,50 Indeed, the results of our analysis indicate that individuals with HF have a lower capillary-to-fibre ratio and a significant association with VO₂peak in patients with HF. A relatively higher capillary-to-fibre ratio enhances muscle O₂ diffusive conductance by reducing the diffusion distance from blood to tissue, increasing O₂ exchange surface area for O₂ transport from the microvasculature to the mitochondria, and therefore improving peripheral O₂ extraction and VO₂peak.⁵¹ Ingjer et al.⁵² demonstrated that capillarization was highest in Type I fibres and lowest in Type IIx fibres, potentially explaining in part why patients with HF had reduced a proportion of Type I fibres and decreased capillarization in our analysis. Altered muscle fibre composition and capillarity in HF will decrease muscle O₂ diffusive conductance and O₂ utilization by skeletal muscles.⁵³ Interestingly, we did not observe a significant association between mean study values for muscle fibre composition and VO₂peak. However, this could be owing to the limited number of studies that concurrently examined VO₂peak and parameters of muscle fibre composition necessary to perform a meta-regression analysis (Type I fibre %, n = 7 studies; Type IIa fibre %, n = 4 studies, Type IIx fibre %, n = 4 studies, capillary-to-fibre ratio, n = 3 studies). Additional research to gain a deeper understanding of this complex relationship in individuals with HF is warranted.

Skeletal muscle function and exercise intolerance in heart failure

Beyond total body and skeletal muscle composition, there are also additional factors that affect skeletal muscle function that could impact exercise tolerance in HF, such as alterations in microstructural, functional, and biochemical muscle factors. Unfortunately, due to the limited number of studies and variability in their methods and results reporting, we could not incorporate these variables into our meta-analysis. However, the following are the general findings from studies assessing these outcomes. Several studies reported a diminished succinate dehydrogenase, citrate synthase, and 3-hydroxyacyl-CoA dehydrogenase in HF group compared with CON, suggesting a decrease in aerobic oxidative enzyme activity of their skeletal muscle (see Supplementary material online, Table S3).^5,9,10,13 Importantly, the elegant study by Mettauer et al.⁹ noted that muscular oxidative capacity (V_max) is decreased in HF compared with active CON, with no significant difference between sedentary CON and HF, suggesting that these differences may be mediated by physical activity rather than HF condition itself. Additionally, the authors observed significantly lower mitochondrial creatine kinase levels in HF, indicating compromised mitochondrial energy supply.⁹ This could be linked to the low percentage of Type I fibres, as these fibres exhibit an increased oxidative enzyme activity and a higher mitochondrial density, contributing to enhanced endurance performance and fatigue resistance.⁴⁹

In contrast to the decreased oxidative enzyme activity and capacity observed in HF, data from our included studies also indicate a tendency for glycolytic enzyme content and activity to increase in patients with HF. This increase could be attributed to the higher proportion of Type IIx fibres, which are characterized by their fast-twitch and quick-fatiguing properties that favour reliance on glycolytic metabolism (see Supplementary material online, Table S3).^5,13,49 Given the predisposition of patients with HF to decreased lean body mass and size, elevations in catabolic markers are another mechanism warranting further investigation in the context of HF, muscle wasting, and exercise intolerance. Although only a few studies have measured the expression of catabolic markers in skeletal muscle in patients with HF, the results are controversial. For instance, while Miller et al.¹⁷ reported no significant differences in the expression of catabolic markers between patients with HF compared to CON, while Forman et al.²³ found elevated expression of atrophy-promoting genes (such as forkhead box proteins O1 and O3 and ubiquitin B), some of which were associated with VO₂peak.

Taken together, the existing scientific literature indicates a reduction in mitochondrial oxidative metabolism, coupled with increased glycolytic metabolism and up-regulation of catabolic gene expression within skeletal muscle in patients with HF. This shift may contribute to the observed decrease in VO₂peak and functional limitations in patients with HF. However, due to small sample sizes and variability in study results reporting, methodologies, and control groups, further research is necessary to elucidate the complex relationship between skeletal muscle function and exercise intolerance in patients with HF.

Limitations

The present analysis is limited by the heterogeneity of the included studies. Firstly, there was a considerable heterogeneity in the study populations in terms of varying sample sizes, sex, HF phenotypes, and HF severity. Moreover, potential differences between HF phenotypes (HFrEF and HFpEF) could not be thoroughly assessed due to limited sample representation, as only 33% of the patients were classified as HFpEF. Additionally, the majority of included patients were male, and no sex-specific analysis was performed; therefore, the generalization of these findings to both sexes should be approached carefully. Comorbidities, and particularly diabetes, may also have contributed to the heterogeneity in body composition outcomes. Unfortunately, the presence of major HF comorbidities such as diabetes was only reported in nine of the included studies (33% of included studies). Furthermore, differences in the characteristics of CON participants may also act as a confounding factor. In most studies, the CON and HF groups were explicitly matched for age and sex, and the mean values for body size were generally well matched between groups. However, as demonstrated in Mettauer et al.⁹, the differences in mitochondrial oxidative capacity between HF and CON groups were largely mediated by the active, fitter CON participants, with no significant differences seen when patients with HF were compared with sedentary controls. Although, overall 17 of the 27 included studies (63% of included studies) controlled for physical activity of the CON participants by either matching for physical activity levels of the participants with HF or by excluding participants who reported performing regular physical activity. Notably, at an individual study level, VO₂peak was significantly lower in HF than sedentary CON groups in all of these studies, and total or leg lean mass was also numerically lower in the vast majority, suggesting that differences in physical activity levels of patients with HF and CON participants were unlikely to explain the overall differences in VO₂peak or body composition (although it may have contributed to heterogeneity). Despite applying the strict inclusion criteria, the inter-study methodological differences for the reported outcomes represent a further limitation. For one, the use of different CPET modes and protocols may affect the measured VO₂peak values and their relationship with body and skeletal muscle composition outcomes. Secondly, varying techniques were utilized for the analysis of body composition and muscle biopsies, thus limiting the comparability between the studies.

Conclusions

Individuals with HF display abnormalities in whole-body and skeletal muscle composition including reduced lean mass, oxidative Type I fibres, and capillary-to-fibre ratio that negatively impact VO₂peak.

Lead author biography

Veronika Schmid is a PhD student at the Department of Preventive Sports Medicine and Sports Cardiology, School of Medicine and Health, Technical University of Munich, in Germany. She holds a master’s degree in Sports Science. Her primary research focuses on heart failure and the role of skeletal muscle in the ageing population. For further information, she can be contacted at [email protected].

Data availability

Not applicable.

Supplementary material

Supplementary material is available at European Heart Journal Open online.

Authors’ contribution

M.J.H. and V.S. contributed to concept and design. V.S., S.J.F., D.W., C.T., J.W., and M.J.H. were involved in acquisition, analysis, or interpretation of data. V.S., J.W., S.J.F., and M.J.H. drafted the manuscript. V.S., S.J.F., C.T., W.T., S.S.A., J.W., M.H., and M.J.H. were involved in critical revision of the manuscript. J.W. performed statistical analysis. M.J.H. was involved in supervision.

Funding

M.J.H. was funded by an endowed research chair in ageing and QoL in the Faculty of Nursing, College of Health Science, University of Alberta, 3-045/11405 87 Ave NW, Edmonton, Alberta T6G IC9, Canada

References

Author notes