Article Text

This article has a correction. Please see:

Download PDFPDF

Original article
Skeletal muscle abnormalities and exercise capacity in adults with a Fontan circulation
  1. Rachael Cordina1,2,
  2. Shamus O'Meagher1,2,
  3. Haslinda Gould3,4,
  4. Caroline Rae5,
  5. Graham Kemp6,
  6. Julie A Pasco3,4,
  7. David S Celermajer1,2
  1. 1Department of Cardiology, Royal Prince Alfred Hospital, Sydney, New South Wales, Australia
  2. 2Department of Medicine, Sydney Medical School, Sydney, New South Wales, Australia
  3. 3School of Medicine, Deakin University, Geelong, Victoria, Australia
  4. 4Department of Medicine, NorthWest Academic Centre, The University of Melbourne, St Albans, Victoria, Australia
  5. 5Neuroscience Research Australia, Sydney, New South Wales, Australia
  6. 6Departments of Musculoskeletal Biology and Magnetic Resonance and Image Analysis Research Centre, University of Liverpool, Liverpool, UK
  1. Correspondence to Dr Rachael L Cordina, GUCH Department, The Heart Hospital, 16-18 Westmoreland St, London W1G 8PH, UK; Rachael.Cordina{at}


Objectives The peripheral muscle pump is key in promoting cardiac filling during exercise, especially in subjects who lack a subpulmonary ventricle (the Fontan circulation). A muscle-wasting syndrome exists in acquired heart failure but has not been assessed in Fontan subjects. We sought to investigate whether adults with the Fontan circulation exhibit reduced skeletal muscle mass and/or metabolic abnormalities.

Design and patients Sixteen New York Heart Association Class I/II Fontan adults (30±2 years) underwent cardiopulmonary exercise testing and lean mass quantification with dual x-ray absorptiometry (DXA); eight had calf muscle 31P magnetic resonance spectroscopy as did eight healthy age-matched and sex-matched controls. DXA results were compared with Australian reference data.

Setting Single tertiary referral centre.

Results Peak VO2 was 1.9±0.1 L/min (66±3% of predicted values). Skeletal muscle mass assessed by relative appendicular lean mass index was significantly reduced compared with age-matched and sex-matched reference values (Z-score −1.46±0.22, p<0.0001). Low skeletal muscle mass correlated with poorer VO2 max (r=0.67, p=0.004). Overall, skeletal muscle mass T-score (derived from comparison with young normal reference mean) was −1.47±0.21; 4/16 Fontan subjects had sarcopenic range muscle wasting (T-score <−2.0) and 9/16 had less marked, but clinically significant wasting (T-score <−1.0 but ≥−2.0). Muscle aerobic capacity, measured by the rate constant (k) of postexercise phosphocreatine resynthesis, was significantly impaired in Fontan adults versus controls (1.48±0.13 vs 2.40±0.33 min−1, p=0.02).

Conclusions Fontan adults have reduced skeletal muscle mass and intrinsic muscle metabolic abnormalities.

View Full Text

Statistics from


Most children born with a ‘single-ventricle’ circulation palliated with a Fontan operation can look forward to significantly improved quality of life, exercise capacity and life expectancy postoperatively1 ,2 but are still faced with impaired physical functioning and increased mortality.2 ,3 The peripheral muscle pump is increasingly recognised as key in promoting cardiac filling during exercise in these subjects who lack a subpulmonary ventricle.4–6 We recently reported that intensive resistance training to augment the peripheral muscle pump improved exercise capacity and cardiac output in Fontan subjects.4 Others have reported that peripheral and respiratory muscle weakness exist in the setting of congenital heart disease and that this might reflect a myopathy similar to that observed in acquired heart failure. We sought to investigate whether muscle wasting and/or metabolic dysfunction exists in adults with a Fontan circulation.



Sixteen consecutively consenting adults with a Fontan circulation were recruited from our coronary heart disease database at Royal Prince Alfred Hospital (RPAH), Sydney, Australia. Exclusion criteria included frequent symptomatic arrhythmias, clinical evidence of heart failure and functionally significant physical or intellectual impairment. Eight healthy non-smoking adults recruited as controls for 31P magnetic resonance spectroscopy (MRS) testing (see below) were required to have no MR contraindications and no significant cardiac or other disease, to be taking no regular medications and to be performing no more than two exercise sessions per week.

Study design

Testing in Fontan adults comprised body composition scanning and cardiopulmonary exercise testing (CPET). A subset of subjects with time availability, no standard contraindications for MRI and oxygen saturations >95% (n=8) also underwent MRS to assess skeletal muscle aerobic capacity. Healthy controls underwent MRS testing only.

Informed written consent was obtained from all subjects and the study was approved by the Sydney Local Health District Ethics Review Committee (RPAH Zone).

Body composition assessment

Total body and appendicular lean mass was assessed by total body dual x-ray absorptiometry (DXA-Lunar Prodigy: GE Healthcare, Milwaukee, USA). An automatically generated template was adjusted manually to ensure the region of interest was accurately positioned to identify the arms, legs and trunk using bony landmarks as described by the manufacturer. Values for lean mass, fat and bone mineral content were obtained and added together for total mass.

Absolute values for total body lean mass and appendicular lean mass (combined lean mass of arms and legs), relative total lean mass index (total lean mass divided by height squared) and relative appendicular lean mass index (appendicular mass divided by height squared) were calculated. Lean mass indices were compared with sex-matched young normal reference data7 to calculate T-scores, as well as age-matched and sex-matched Australian reference data to generate Z-scores. Sarcopenic-range skeletal muscle wasting was defined as skeletal muscle mass >2 SD below the mean measured in young adults (20–39 years old) of the same sex and ethnic background8 using relative appendicular lean mass index for skeletal muscle mass characterisation. This index is the most widely used parameter for such assessments9 and is superior to total body lean mass indices since DXA lean mass quantification includes water and skin, which contribute a relatively small proportion to overall appendicular lean mass.10

Cardiopulmonary exercise testing

CPET comprised a ramp-protocol cycle test on an electrically-braked bicycle ergometer (Sensormedics 800 Computerised ergometer: Sensormedics Corporation, California, USA). The rate of ergometer power increase (Watts/min) was chosen so that the subject's peak workload would be achieved in around 10 min. A detailed description of the methods is given elsewhere.4 In brief, 15-s averages of breath-by-breath data were calculated. Peak VO2 was defined as the maximal averaged value at peak workload. Oxygen pulse was calculated by dividing VO2 by heart rate. Anaerobic threshold was calculated using the V-slope method, by a blinded observer.11 Peak VO2 was compared with predicted normal values.12

Calf muscle 31phosphorus MRS

On a separate day to exercise testing, MRS was used to assess skeletal muscle metabolism at rest, during exercise and during recovery. Subjects lay supine on a custom-built rig within a Phillips Achieva TX 3 T magnet (Philips Medical Systems, Best, The Netherlands). A 10 cm diameter 31P surface coil (Pulseteq Ltd, Wotton-under-Edge, UK) was strapped to the non-dominant calf. The legs were positioned flat and the non-dominant foot was placed on a pedal system. Force production was measured via a pressure transducer (Honeywell, Morristown USA). The exercise protocol is described in detail elsewhere.4 In brief, a pulse-acquired sequence with an adiabatic pulse with the offset positioned 300 Hz to high frequency of the β-ATP resonance was used to collect 8-scan resting 31P spectra (TR=10 000 ms) to assess resting metabolic state. Subjects then performed cycles of moderate-intensity and high-intensity isometric dorsiflexion. Spectra were collected every 8s (TR=2000 ms) in exercise and in recovery.

Spectra were quantified using the AMARES algorithm within the Java-based magnetic resonance user interface (jMRUI V.3.0, EU Project) to obtain relative concentrations of inorganic phosphate, phosphocreatine (PCr) and ATP. Kinetics of postexercise PCr recovery was assessed by least-squares fit of the PCr relative signal intensity time-course to a monoexponential function. Because pH changes during exercise were minimal, the exponential rate constant of PCr recovery, k (=0.693/halftime) can be taken as a measure of overall muscle mitochondrial capacity, an integrated system property depending on intrinsic mitochondrial numbers and function, and on cardiovascular delivery of substrate and oxygen to the muscle.


All descriptive data are expressed as mean±SEM. Data were tested for normality using a Shapiro–Wilk test. Since all data were normally distributed, unpaired Student's t tests were employed for between-group analyses. In this exploratory study, p values have not been adjusted for the multiple comparisons performed. Peak oxygen pulse and VO2 were compared with relative appendicular lean mass index and Z-score to characterise the relationship, if any, between tested variables. Spearman's rank correlation was used due to the small sample size and nonlinear distribution of data (p values <0.1 are reported for correlations). SPSS statistics Data Editor (IBM Corporation, New York, USA) was used for statistical calculations. A two-tailed p value ≤0.05 was considered statistically significant.


Study population

Fontan subject characteristics are summarised in table 1. All were New York Heart Association (NYHA) Class I/II although this was not a specific inclusion criterion. Overall, atriopulmonary connection (APC) subjects were older compared with total cavopulmonary connection (TCPC) subjects (34±3 vs 27±2 years, n=6 vs n=10, p=0.05) and time since repair was longer (22±1 years for APC vs 16±1 years for TCPC, p=0.009) but otherwise the groups were similar (p>0.4 for all) and thus subgroup data are not shown. More detailed cardiac anatomic data are given in online supplementary material. Characteristics for the two men, six women MRS subgroup included mean age of 31±3 years for Fontan subjects and 31±2 years in healthy controls, and body mass index was 25±1 kg/m2 in Fontan subjects and 26±3 kg/m2 for controls (p=0.8 for both).

Table 1

Fontan subject characteristics

Cardiopulmonary exercise testing

Peak VO2 was 1.9±0.1 L/min (24.5±1.6 mL/kg/min or 66±3% predicted). Peak work was 145±10 W (72±4% predicted). VO2 at anaerobic threshold was 1.3±0.1 L/min (45±2% predicted). Peak oxygen pulse indexed to body surface area was 6.5±0.5 mL O2/beat/m2 (59±3% predicted). More detailed CPET are shown in online supplementary table S2. Peak VO2 was 69±3% of predicted values in the Fontan subgroup that had MRS testing.

Body composition assessment

Lean mass results are shown on table 2 and demonstrated in figure 1. Mean relative appendicular lean mass index T-score was −1.47±0.21. Four of 16 subjects had a relative appendicular lean mass index T-score in the sarcopenic range (<−2.0). In addition, nine Fontan subjects had relative appendicular lean mass index T-score <−1 and ≥−2.0 representing a less marked, but still clinically important, reduction in lean mass. Relative appendicular lean mass index Z-score was −1.46±0.22, which was markedly reduced compared with age-matched and sex-matched reference values (where Z-score=0, p<0.0001).

Table 2

Lean mass in Fontan subjects

Figure 1

A T-score represents the number of SDs from the young normal reference mean. A value <−2.0 represents marked muscle wasting, defined as in the sarcopenic range.

Skeletal muscle mass correlation analysis

Relative appendicular lean mass index positively correlated with peak VO2 (R=0.67, p=0.004) but not with percent-predicted peak VO2 values; however, relative appendicular lean mass index Z-score tended to correlate with both peak VO2 (R= 0.40, p=0.062) and percent-predicted peak VO2 (R=0.37, p=0.077). A significant relationship existed between relative appendicular lean mass index and peak oxygen pulse for both body surface area-indexed (R=0.56, p=0.025) and percent-predicted values (R=0.44, p=0.042). Relative appendicular lean mass index Z-score tended to correlate with percent-predicted peak oxygen pulse (R=0.40, p=0.059).

Calf muscle 31phosphorus MRS

The results of MRS are demonstrated in figure 2. Muscle aerobic capacity was impaired in Fontan subjects compared with healthy controls. PCr recovery constant was 1.48±0.13 vs 2.40±0.33 min−1, p=0.02, n=8.

Figure 2

Muscle aerobic capacity as measured by the rate constant (k) of postexercise phosphocreatine resynthesis during calf muscle 31P magnetic resonance spectroscopy.


In relatively well adults with a Fontan circulation, our observations suggest that muscle wasting is common and associated with reduced peak exercise capacity. Given the exquisite dependence of this physiologic arrangement on the peripheral muscle pump for cardiac filling during exercise, these findings may be of particular clinical relevance.

Body composition has not previously been assessed in subjects with a Fontan circulation, nor has the relationship between lean body mass and exercise capacity been characterised. In our group of relatively older adults with a Fontan circulation, peak VO2 measurements indicated moderate exercise limitation, as previously documented in younger Fontan patients.1 ,13–18 A quarter showed muscle wasting in the sarcopenic range. Lean mass positively correlated with peak VO2 and oxygen pulse. Since oxygen pulse reflects both stroke volume and muscle oxygen extracting capability, the positive association between oxygen pulse indexed for body surface area and relative lean mass index might thus reflect both a superior peripheral muscle pump and healthier muscle. The association between VO2 and lean mass may also, in part, be related to superior strength associated with increased muscle bulk and enhanced ergometer performance.

The peripheral muscle pump is the main factor driving blood into the heart at the onset of erect exercise,19 ,20 and in Fontan subjects, who are especially preload sensitive, this effect may be of particular importance. Furthermore, our recent research suggested that the greater peripheral muscle bulk seen after resistance training in adults with a Fontan circulation is associated with improved stroke volume during exertion.4

In chronic heart failure, there is a well-described myopathy,21 and lean muscle mass and strength are independent predictors of survival.22 ,23 Three previous studies have demonstrated skeletal muscle abnormalities in Fontan subjects. Brassard et al24 assessed ergoreceptor function (small afferent receptors sensitive to metabolites during muscle contraction) through their contribution to blood pressure alteration during exercise; they found a significantly impaired response in Fontan subjects compared with controls. Greutmann et al25 found marked weakness of the respiratory and peripheral muscles, and Inai et al26 showed the capacity for skeletal muscles to replenish oxygen saturations postexercise is attenuated in those with Fontan physiology. Studies of heart failure myopathy have shown muscle wasting disproportionate to cardiac dysfunction.27 31P MRS has shown excessive depletion of PCr and increased intracellular acidosis at a lower workload than normal (in general terms, the results of impaired muscle ATP generating pathways), endothelial dysfunction exists28 ,29 as well as an increase in fatigue-prone type II muscle fibres,30 and our results demonstrating muscle wasting and impaired PCr recovery in subjects with a Fontan circulation suggest that similar abnormalities may be present.

The reduction in lean mass that we observed is likely multifactorial. In acquired heart failure, the pathogenesis is still poorly understood; poor nutrition, malabsorption, inactivity, increased sympathetic stimulation and neurohumoral alterations have all been implicated.31 In Fontan subjects, several of these may also be important but are largely uncharacterised. Nutrition and protein loss related to portal hypertension may be of particular relevance in this group. Of note, in this cohort of Fontan adults, body mass index was >25 kg/m2 suggesting that although a reduction in skeletal mass was observed, as a group, they were slightly overweight. We did not perform a comparative analysis for body fat composition as such local reference data do not exist at this time.


We were limited by small subject numbers, for MRS testing in particular, for which we were constrained by practical factors. The size of the cohort likely also reduced our statistical power for correlation analysis and thus we included several results with borderline statistical significance. We were not able to characterise what degree of deconditioning had contributed to reduced lean mass. It is possible that subjects with better stroke volume lead more active lives and thus have a higher lean body mass as a result, although our cohort did not differ vastly in functional status; none of our subjects were highly trained or performing >2 exercise sessions per week and all were NYHA Class I/II.


Young adults with ‘single-ventricle’ hearts palliated with a Fontan operation have markedly impaired muscle aerobic capacity and are at increased risk for reduced skeletal muscle mass. This is associated with impaired peak exercise capacity and oxygen pulse. Greater peripheral muscle mass may improve exercise performance via an augmented peripheral muscle pump, superior strength and better oxygen extraction. Histologic testing and larger-scale studies are warranted to further characterise this probable ‘Fontan myopathy’. Exercise prescription with a focus on skeletal muscle training might be an important aspect of routine care for adults with Fontan physiology.


The authors wish to thank Mrs Julie Hetherington, Department of Endocrinology, Royal Prince Alfred Hospital, Sydney Australia for her assistance with collecting and analysing dual x-ray absorptiometry data as well as Mr Ken Rayner, Mr Daniel Moran and Mr Segar Suppiah, Diagnostic Medical Services, Neuroscience Research Australia, for expert radiography.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors RC designed the project, obtained ethics and supervised the testing. She analysed the majority of the data and was the primary author of the paper. SO assisted with data collection and contributed to paper construction. HG contributed to manuscript preparation and statistical analysis for the body composition section of the manuscript. GK and CR assisted with design and implementation of the MRS testing as well as data analysis and writing for that section of the paper. DC was primarily responsible for the study. He contributed to study design and was the primary editor of the manuscript. He also provided expert input for data interpretation and manuscript construction.

  • Competing interests None.

  • Ethics approval Royal Prince Alfred Hospital Area Ethics Committee, NSW, Australia.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement We will consider sharing our data with interested parties for scientific purposes if they contact the corresponding author with their request.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and British Cardiovascular Society