High adiponectin concentrations have emerged as an independent risk factor of outcome in patients with CHF (chronic heart failure); however, modification of adiponectin in CHF patients has not been assessed to date. The aim of the present study was to investigate the effect of exercise training on adiponectin levels in CHF patients. A total of 80 patients with CHF due to systolic dysfunction were included. The effect of 4 months exercise training was studied in 46 patients, whereas the remaining 34 untrained CHF patients served as a sedentary control group. Circulating adiponectin concentrations, exercise capacity, anthropometric data and NT-proBNP (N-terminal pro-brain natriuretic peptide) levels were assessed. Adiponectin levels were significantly higher in CHF patients compared with healthy subjects [9.3 (7.1–16.1) and 4.9 (3.9–8.6) mg/l respectively; P=0.015]. Stratification of CHF patients according to tertiles of NT-proBNP revealed an increase in adiponectin with disease severity (P<0.0001). Exercise training reduced circulating adiponectin levels in CHF patients [10.7 (7.2–17.6) mg/l before training to 9.4 (5.9–14.8) mg/l after training; P=0.013], whereas no changes were observed in the sedentary CHF group [9.0 (7.0–13.5) mg/l before training and 10.1 (6.0–15.7) mg/l after a similar time interval]. A significant time×group interaction (P=0.008) was observed for the mean change in adiponectin between the trained and untrained CHF patients. Adiponectin concentrations were positively associated with NT-proBNP and HDL (high-density lipoprotein)-cholesterol and negatively correlated with BMI (body mass index), triacylglycerols and exercise capacity. In conclusion, circulating adiponectin concentrations are higher in CHF patients compared with healthy subjects and increase with disease severity. Exercise training for 4 months lowers circulating adiponectin levels.

INTRODUCTION

CHF (chronic heart failure) is a complex syndrome of haemodynamic and neurohormonal abnormalities. Recently, the impact of insulin resistance on the progression of CHF [1] and the existence of a metabolic vicious cycle in heart failure [2] have been highlighted.

A hyperadrenergic state, characteristic of CHF, initiates an adverse metabolic vicious cycle, whereby aberrant metabolism, such as insulin resistance, further detrimentally affects disease progression [2,3]. In spite of therapeutic advances, mortality and morbidity in CHF are still unacceptably high. Inefficient cardiac energy utilization and metabolic impairment in general represent promising targets for CHF therapy [3]. Physical training has emerged as one of the most efficient ways of improving physical capacity and quality of life in CHF patients. In addition, exercise training provides anti-remodelling effects, and meta-analyses strongly suggest survival benefit [4]. From a pathophysiological point of view, the fact that exercise training modulates inflammation, neuro-endocrine activation and oxidative stress in these patients is highly relevant [58].

Adiponectin is an adipocyte-derived cytokine with insulin-sensitizing, anti-inflammatory and anti-atherogenic properties [9]. Adiponectin levels are reduced in conditions such as Type 2 diabetes mellitus, obesity, the metabolic syndrome, hypertension and ischaemic heart disease, but increase following lifestyle adaptations such as weight loss or restricted energy intake. In contrast with these findings, in CHF increased circulating levels of adiponectin have emerged as an independent risk factor for morbidity and mortality [1012]. The mechanisms underlying this inverse relationship remain unexplained. Several theories have been put forward, including adiponectin as a mere marker of disease severity [10,13], the compensatory rise of adiponectin in response to inflammatory and metabolic disturbances in CHF [14,15], the association between adiponectin and natriuretic peptides [10,14,16] and the existence of adiponectin resistance in CHF [17].

Modification of adiponectin levels in CHF patients after intervention has not been studied previously. In the present study, we assessed circulating adiponectin concentrations in CHF patients, compared with controls, and evaluated the effects of a 4-month exercise training programme.

MATERIALS AND METHODS

Subjects and study design

A total of 80 consecutive patients with CHF and systolic dysfunction due to ischaemic or dilated cardiomyopathy were recruited from the Cardiac Rehabilitation Centre and the outpatient Heart Failure Clinic of the University Hospital of Antwerp. Patients were included in this prospective non-randomized single-centre study if (i) LVEF (left ventricular ejection fraction) was <30% (assessed by echocardiography); (ii) they were in NYHA (New York Heart Association) functional class II or III; and (iii) their symptoms had been stable on medical treatment for at least 1 month prior to inclusion. Exclusion criteria were: (i) recent acute coronary syndrome or revascularization (<3 months), valvular disease requiring surgery, exercise-induced myocardial ischaemia or malignant ventricular arrhythmia, acute myocarditis or pericarditis; (ii) cerebrovascular or musculoskeletal disease preventing exercise testing or training; (iii) acute or chronic infections, allergies, cancer or inflammatory diseases; and (iv) diabetes mellitus. All data were assessed at baseline and after 4 months and consisted of clinical evaluation by a cardiologist (medical history, including medication used, NYHA classification based on patient information and physical examination), echocardiography, collection of fasting blood samples and CPET (cardiopulmonary exercise testing). Staff members were blinded to previous CPET results and the patients' participation and grouping. Medication remained unchanged during the study period.

The effect of exercise was studied in 46 CHF patients enrolled in a 4-months exercise training programme at the Cardiac Rehabilitation Centre. A total of 34 age- and gender-matched CHF patients, attending the outpatient Heart Failure Clinic, served as an untrained CHF control group. They had comparable disease severity, but were unable to attend the training programme because of, for example, logistical problems. The same data were assessed in these patients before and after a similar time interval of 4 months. Ten age- and gender-matched subjects served as a healthy baseline control group (no medication, no chronic underlying condition, no cardiac history and no allergic condition).

The study complies with the Declaration of Helsinki, the research protocol was approved by the local Ethics committee (Antwerp University Hospital), and written informed consent was obtained from all subjects.

Exercise training

Patients trained in hospital on an ambulatory base three times/week for 1 h. The training session started with 5 min of warming-up and concluded with a 5-min cooling down and stretching period. Two types of training programme were in clinical practice at the time of inclusion: ET (endurance training) and CT (combined endurance/resistance training). Both training protocols have been described in detail previously [18]. They are routinely used and are considered standard treatment at our Rehabilitation Centre. Depending on the patients' characteristics, such as severe muscle wasting, dynamic resistive exercises were incorporated into the training programme.

Briefly, for both CT- and ET-trained patients, an endurance training THR (target heart rate) was calculated as 90% of the heart rate achieved at the anaerobic threshold. The initial resistance training intensity was set at 50% of 1RM (1 repetitive maximum) (for the nine different muscle groups), with an increase to 60% after 2 months. Repetitions were slowly increased from 1×10, 1×15, 2×10 to 2×15 repetitions. Between each series of repetitions, rest for 1 min was allowed. The ET group trained for 8 min on five different training devices (treadmill, bicycle, stair or step, arm-cycling, half recumbent or reclined cycling). When changing from one device to another, 2 min of recuperation time was introduced.

During the first 2 months, patients assigned to the CT group trained for almost 40 min on the Fitness equipment (Unica; Technogym), whereas only 10 min were spent on ET. The next 2 months, resistance training was reduced to 30 min (nine muscle groups, 2×15 repetitions each) and ET was increased to 2×8 min.

Measurements

CPET

Symptom-limited CPET was performed on a treadmill (Medical Jaeger) under non-fasting conditions and medication. Depending on age, NYHA class and LVEF, two protocols were used, aiming at an optimal exercise duration of 8–12 min. A ramp protocol starting with an equivalent of 20 or 40W and incremental steps of 10 or 20W/min respectively, was used. Breath-by-breath gas exchange measurements permitted determination of V̇E (minute ventilation), V̇O2 (oxygen uptake) and V̇CO2 (carbon dioxide production) online (Cardiovit CS-200 Ergo-Spiro; Schiller). Patients were monitored continuously with a 12-lead ECG, and automatic blood pressure measurement was performed every 2 min. V̇O2peak (peak V̇O2) was expressed as the highest attained V̇O2. CPET was performed at baseline and after 4 months. Body composition was assessed by bioelectrical impedance analysis (BF300 body fat monitor; Omron) [18].

Biochemical analyses

Fasting blood samples were collected between 08.00 and 0.900 hours at baseline and after 4 months. Creatinine, total cholesterol, TAGs [triacylglycerols (triglycerides)], LDL (low-density lipoprotein)-cholesterol and HDL (high-density lipoprotein)-cholesterol levels were assessed immediately. Plasma was separated by centrifugation and aliquots were stored at −20°C.

Circulating adiponectin concentrations were measured using a commercially available ELISA (Human adiponectin Quantikine® ELISA; R&D Systems). Standards and samples were analysed in duplicate according the manufacturer's instructions. The minimal detection limit of the assay is 0.246 μg/l, with an intra-assay precision of 2.5% for 19.8±0.50 mg/l (n=20) and inter-assay precision of 3.2% for 12.5±0.41 mg/l (n=40). Internal controls were included (Quantikine® ELISA kit Controls).

NT-proBNP (N-terminal pro-brain natriuretic peptide) was determined using a sandwich immunoassay on an Elecsys 2010 (Roche Diagnostics). The analytical range extends from 5 to 35000 pg/ml. The coefficient of variation was 1.3% (n=10) at a level of 221 pg/ml and 1.2% (n=10) at a level of 4091 pg/ml.

Statistical analyses

Results are expressed as means±S.D. or medians (interquartile range) for non-Gaussian distributed parameters. The data were tested for normal distributions using the Kolmogorov–Smirnov test. Baseline characteristics of the groups were compared using a Student's t test or Mann–Witney U and Kruskal–Wallis test for continuous variables, as appropriate. A χ2 test was used for categorical variables.

Pairwise comparisons (baseline compared with 4 months) were carried out using a paired samples Student's t test or Wilcoxon signed-rank test. Differences between the two groups (trained compared with untrained) and changes over time within each group (time effect), as well as any interaction (different trends over time between groups), were assessed using two-way repeated measures ANOVA. Linear regression analysis for the change in adiponectin was performed to identify possible relationships with other variables.

Because circulating levels of adiponectin, NT-proBNP TAG, CRP (C-reactive protein) and the measurements reflecting exercise capacity and LVEF were not normally distributed, correlations at baseline were determined with the non-parametric Spearman's correlation test. Multivariable linear regression analyses examining the correlates of circulating adiponectin levels included baseline variables that were associated with adiponectin at the P<0.05 level in univariate analyses. P value <0.05 was considered statistically significant. All statistical analyses were performed using the software package SPSS version 16.0.

RESULTS

Clinical characteristics

A total of 80 CHF patients were recruited in the present study. From the total group of 80 CHF patients, 46 consecutive patients were enrolled in a 4-month training programme, whereas 34 CHF patients served as an untrained CHF control group. As shown in Table 1, both groups of CHF patient were well matched at baseline with respect to demographic characteristics, LVEF, functional NYHA class, aetiology, co-morbidities, medical treatment, V̇O2peak and laboratory measurements. HDL-cholesterol values at baseline were significantly higher in the untrained CHF group.

Table 1
Baseline characteristics of CHF patients

Values are means (S.D.) or medians (interquartile range). ATII, angiotensin II antagonists. Differences between the trained and untrained CHF groups were determined using a Student's t test, Mann–Whitney U test or χ2 test, as appropriate.

Characteristic CHF patients (n=80) Trained CHF (n=46) Untrained CHF (n=34) P value 
Demographics     
 Age (years) 59.0 (11.2) 57.5 (10.8) 61.1 (11.6) 0.158 
 Gender (% male) 65 70 59 0.319 
 BMI (kg/m225.3 (4.4) 25.1 (4.4) 25.1 (4.3) 0.957 
Serum creatinine (mg/dl) 1.2 (1.0–1.6) 1.2 (1.0–1.5) 1.25 (1.0–1.7) 0.551 
Heart failure measurements     
 LVEF (%) 18 (15–23) 17 (14–22) 19 (15–24) 0.115 
 NT-proBNP (pg/ml) 1088 (482–2757) 1216 (530–2887) 833 (373–2477) 0.154 
 NYHA class 2/3 (%) 77.5/22.5 74/26 82/18 0.372 
 Aetiology (% ischaemic) 55 52 59 0.555 
V̇O2peak (ml·kg−1 of body weight·min−119 (16–24) 19 (15–24) 20 (16–24) 0.381 
Medication (%)     
 β-Blocker 75 70 82 0.192 
 ACEI and/or ATII 99 100 97 0.242 
 Diuretics 84 80 88 0.350 
 Statins 30 22 41 0.061 
Lipid status (mmol/l)     
 Total cholesterol 5.2 (1.1) 5.1 (1.2) 5.3 (1.1) 0.725 
 HDL-cholesterol 1.3 (0.4) 1.2 (0.3) 1.4 (0.4) 0.045 
 LDL-cholesterol 3.3 (1.0) 3.4 (1.1) 3.2 (0.9) 0.353 
 TAG 128 (100–199) 126 (94–202) 130 (107–198) 0.931 
Adiponectin (mg/l) 9.3 (7.1–16.1) 10.7 (7.2–17.6) 9.0 (7.0–13.5) 0.450 
Characteristic CHF patients (n=80) Trained CHF (n=46) Untrained CHF (n=34) P value 
Demographics     
 Age (years) 59.0 (11.2) 57.5 (10.8) 61.1 (11.6) 0.158 
 Gender (% male) 65 70 59 0.319 
 BMI (kg/m225.3 (4.4) 25.1 (4.4) 25.1 (4.3) 0.957 
Serum creatinine (mg/dl) 1.2 (1.0–1.6) 1.2 (1.0–1.5) 1.25 (1.0–1.7) 0.551 
Heart failure measurements     
 LVEF (%) 18 (15–23) 17 (14–22) 19 (15–24) 0.115 
 NT-proBNP (pg/ml) 1088 (482–2757) 1216 (530–2887) 833 (373–2477) 0.154 
 NYHA class 2/3 (%) 77.5/22.5 74/26 82/18 0.372 
 Aetiology (% ischaemic) 55 52 59 0.555 
V̇O2peak (ml·kg−1 of body weight·min−119 (16–24) 19 (15–24) 20 (16–24) 0.381 
Medication (%)     
 β-Blocker 75 70 82 0.192 
 ACEI and/or ATII 99 100 97 0.242 
 Diuretics 84 80 88 0.350 
 Statins 30 22 41 0.061 
Lipid status (mmol/l)     
 Total cholesterol 5.2 (1.1) 5.1 (1.2) 5.3 (1.1) 0.725 
 HDL-cholesterol 1.3 (0.4) 1.2 (0.3) 1.4 (0.4) 0.045 
 LDL-cholesterol 3.3 (1.0) 3.4 (1.1) 3.2 (0.9) 0.353 
 TAG 128 (100–199) 126 (94–202) 130 (107–198) 0.931 
Adiponectin (mg/l) 9.3 (7.1–16.1) 10.7 (7.2–17.6) 9.0 (7.0–13.5) 0.450 

Circulating adiponectin levels

For the total CHF population (n=80), adiponectin levels were significantly higher compared with healthy subjects (n=10; mean age, 60.8±5.0 years; and 60% males) [9.3 (7.1–16.1) compared with 4.9 (3.9–8.6) mg/l; P=0.015]. Stratification of CHF patients according to the tertiles of NT-proBNP revealed a significant increase in adiponectin with disease severity (P<0.0001; Figure 1).

Circulating adiponectin levels in healthy subjects and CHF patients according to severity of disease

Figure 1
Circulating adiponectin levels in healthy subjects and CHF patients according to severity of disease

Baseline circulating adiponectin levels in CHF patients (n=80) and healthy control subjects (n=10) are shown. CHF patients are divided into tertiles of baseline NT-proBNP levels, indicating increasing disease severity. Values are medians (interquartile range) and 5–95% range. Differences among the groups were analysed using a Kruskal–Wallis test.

Figure 1
Circulating adiponectin levels in healthy subjects and CHF patients according to severity of disease

Baseline circulating adiponectin levels in CHF patients (n=80) and healthy control subjects (n=10) are shown. CHF patients are divided into tertiles of baseline NT-proBNP levels, indicating increasing disease severity. Values are medians (interquartile range) and 5–95% range. Differences among the groups were analysed using a Kruskal–Wallis test.

Adiponectin levels were significantly higher in female compared with male CHF patients [13.6 (9.5–19.3) and 8.2 (5.1–13.9) mg/l; P<0.001]. The presence of chronic renal failure did not influence circulating adiponectin levels in our CHF patients (P>0.05). After stratification according to disease aetiology (ischaemic or idiopathic), no significant differences were found between either group (P>0.05).

Effect of exercise

Changes in exercise capacity and prognostic markers

Table 2 shows the effect of 4 months of exercise training. A significant time×group interaction was observed between the trained and untrained CHF patients for the mean change in maximal workload (+33.4±4.0 compared with −0.4±1.7; P<0.001) and the mean change in V̇O2peak (+1.8±0.6 compared with −0.5±0.4; P=0.003). Furthermore, exercise training resulted in a beneficial effect on work efficiency and NT-proBNP levels, and improved NYHA functional class significantly (P<0.05 for the time×group interaction)

Table 2
Effect of exercise training in patients with CHF

Values are expressed as means (S.D.) or medians (interquartile range). *P<0.05, **P<0.01 and ***P<0.001 compared with baseline within each group, as determined using a paired samples Student's t test or Wilcoxon signed-rank test. †Differences in changes between groups as determined by ANOVA.

 Trained group (n=46) Untrained group (n=34)  
Parameter Baseline 4 months Baseline 4 months P value† 
Prognostic markers      
NT-proBNP (pg/ml) 1216 (530–2887) 1186 (385–2496)* 833 (373–2477) 976 (420–2082) 0.122 
LVEF (%) 17 (14–22) 17 (14–23) 19 (15–24) 19 (15–27) 0.276 
NYHA class I/II/III (%) 0/74/26 47/46/7*** 0/82/18 6/85/9 <0.001 
Anthropometric data      
 BMI (kg/m225.1 (4.4) 25.9 (4.1)* 25.1 (4.3) 25.3 (4.3) 0.594 
 Waist circumference (cm) 90 (15) 90 (14) 86 (16) 88 (18) 0.553 
 Lean body mass (kg) 52 (11) 52 (11) 47 (13) 48 (13) 0.471 
Maximal exercise capacity      
V̇O2peak (ml−1·kg−1 of body weight min−119 (15–24) 21 (16–25)** 20 (16–24) 19 (15–24) 0.003 
 Maximal workload (W) 90 (70–133) 130 (100–160)*** 100 (80–140) 100 (80–145) <0.001 
 Work efficiency (W/V̇O2peak) (W·ml−1·kg−1 of body weight min−15.0 (4.5–5.7) 6.4 (5.4–7.5)*** 5.4 (4.9–6.2) 5.7 (4.9–6.4) <0.001 
Biochemical data (mmol/l)      
 Total cholesterol 5.2 (1.1) 5.4 (1.0) 5.3 (1.1) 4.9 (0.9)* 0.014 
 HDL-cholesterol 1.2 (0.3) 1.3 (0.4)** 1.4 (0.4) 1.4 (0.3) 0.088 
 LDL-cholesterol 3.3 (1.0) 3.5 (1.0) 3.2 (0.9) 3.0 (0.8) 0.059 
 TAG 126 (94–202) 137 (95–174) 130 (107–198) 160 (88–233) 0.437 
Adiponectin (mg/l) 10.7 (7.2–17.6) 9.4 (5.9–14.8)* 9.0 (7.0–13.5) 10.1 (6.0–15.7) 0.008 
 Trained group (n=46) Untrained group (n=34)  
Parameter Baseline 4 months Baseline 4 months P value† 
Prognostic markers      
NT-proBNP (pg/ml) 1216 (530–2887) 1186 (385–2496)* 833 (373–2477) 976 (420–2082) 0.122 
LVEF (%) 17 (14–22) 17 (14–23) 19 (15–24) 19 (15–27) 0.276 
NYHA class I/II/III (%) 0/74/26 47/46/7*** 0/82/18 6/85/9 <0.001 
Anthropometric data      
 BMI (kg/m225.1 (4.4) 25.9 (4.1)* 25.1 (4.3) 25.3 (4.3) 0.594 
 Waist circumference (cm) 90 (15) 90 (14) 86 (16) 88 (18) 0.553 
 Lean body mass (kg) 52 (11) 52 (11) 47 (13) 48 (13) 0.471 
Maximal exercise capacity      
V̇O2peak (ml−1·kg−1 of body weight min−119 (15–24) 21 (16–25)** 20 (16–24) 19 (15–24) 0.003 
 Maximal workload (W) 90 (70–133) 130 (100–160)*** 100 (80–140) 100 (80–145) <0.001 
 Work efficiency (W/V̇O2peak) (W·ml−1·kg−1 of body weight min−15.0 (4.5–5.7) 6.4 (5.4–7.5)*** 5.4 (4.9–6.2) 5.7 (4.9–6.4) <0.001 
Biochemical data (mmol/l)      
 Total cholesterol 5.2 (1.1) 5.4 (1.0) 5.3 (1.1) 4.9 (0.9)* 0.014 
 HDL-cholesterol 1.2 (0.3) 1.3 (0.4)** 1.4 (0.4) 1.4 (0.3) 0.088 
 LDL-cholesterol 3.3 (1.0) 3.5 (1.0) 3.2 (0.9) 3.0 (0.8) 0.059 
 TAG 126 (94–202) 137 (95–174) 130 (107–198) 160 (88–233) 0.437 
Adiponectin (mg/l) 10.7 (7.2–17.6) 9.4 (5.9–14.8)* 9.0 (7.0–13.5) 10.1 (6.0–15.7) 0.008 

Changes in anthropometric data and biochemical characteristics

Training did not lead to a significant change in body weight or any other anthropometric data (fat mass, lean body mass, and waist and hip circumference). Lipoproteins remained unchanged, except for total cholesterol values which were significantly lower after 4 months of follow up in the untrained group (Table 2).

Changes in adiponectin levels

Exercise training significantly reduced circulating adiponectin levels in CHF patients (P=0.013; Table 2), whereas no significant changes were observed in the untrained control CHF group (Table 2). A significant time×group interaction (P=0.008) was observed for the mean change in adiponectin between the trained and untrained CHF patients (Figure 2).

Change in adiponectin levels before and after 4 months in the trained and untrained CHF groups
Figure 2
Change in adiponectin levels before and after 4 months in the trained and untrained CHF groups

Values are means±S.E.M. P value indicates the time×group interaction.

Figure 2
Change in adiponectin levels before and after 4 months in the trained and untrained CHF groups

Values are means±S.E.M. P value indicates the time×group interaction.

Linear regression analysis for the change in adiponectin confirmed the independent effect of exercise training on the decrease in circulating adiponectin level, after adjustment for associated baseline variables (NT-proBNP, HDL-cholesterol, BMI (body mass index) and parameters reflecting exercise capacity) and for the change in variables with a significant difference after exercise training (Table 2).

In the present study, there was no difference between the type of exercise training (ET or CT) with respect to the decrease in adiponectin levels (P=0.284 for time×group interaction).

Correlation with adiponectin

Adiponectin concentrations measured at baseline in the total CHF group (n=80) were positively associated with NT-proBNP (Figure 3) and were inversely correlated with an atherogenic lipid profile (TAGs, r=−0.493, P<0.0001; HDL-cholesterol, r=0.480, P<0.0001). As shown in Table 3, a strong negative relationship was found with the anthropometric results, with the exception of fat mass.

Baseline correlation of NT-proBNP with adiponectin

Figure 3
Baseline correlation of NT-proBNP with adiponectin

Scatterplot showing the association between plasma adiponectin and NT-proBNP levels.

Figure 3
Baseline correlation of NT-proBNP with adiponectin

Scatterplot showing the association between plasma adiponectin and NT-proBNP levels.

Table 3
Baseline correlations with adiponectin
Vairable Univariate correlation coefficient P value Multivariate β-coefficient P value 
Gender (male=1, female=2) 0.401 <0.0001   
BMI (kg/m2−0.345 0.002   
Lean body mass (kg) −0.465 0.001   
Waist circumference (cm) −0.487 <0.0001   
Waist/hip ratio −0.414 0.003   
HDL-cholesterol (mmol/l) 0.480 <0.0001 0.663 <0.0001 
TAG (mmol/l) −0.493 <0.0001   
NT-proBNP (pg/ml) 0.481 <0.0001 0.344 0.001 
V̇O2peak (ml−1·kg−1 of body weight·min−1−0.392 <0.0001   
Maximal workload (W) −0.442 <0.0001   
Vairable Univariate correlation coefficient P value Multivariate β-coefficient P value 
Gender (male=1, female=2) 0.401 <0.0001   
BMI (kg/m2−0.345 0.002   
Lean body mass (kg) −0.465 0.001   
Waist circumference (cm) −0.487 <0.0001   
Waist/hip ratio −0.414 0.003   
HDL-cholesterol (mmol/l) 0.480 <0.0001 0.663 <0.0001 
TAG (mmol/l) −0.493 <0.0001   
NT-proBNP (pg/ml) 0.481 <0.0001 0.344 0.001 
V̇O2peak (ml−1·kg−1 of body weight·min−1−0.392 <0.0001   
Maximal workload (W) −0.442 <0.0001   

Adiponectin correlated negatively with exercise capacity (V̇O2peak and maximal workload; Table 3). There were no significant correlations between adiponectin and LVEF, inflammatory parameters (CRP and sedimentation) or renal function.

On stepwise multivariable analyses, two parameters (i.e. high NT-proBNP and HDL-cholesterol) were significant independent predictors of baseline circulating adiponectin levels (Table 3).

DISCUSSION

Although previous studies have already demonstrated increased adiponectin concentrations in patients with CHF, our present study is the first to evaluate the effect of exercise training on circulating adiponectin levels in CHF patients with systolic dysfunction. Several interesting findings emerged from the present study. (i) We have demonstrated for the first time that adiponectin levels in CHF can be modified by exercise training. Modulation of adiponectin levels in CHF patients occurs independently of possible confounders, such as baseline exercise capacity, LVEF, lipid profile and NT-proBNP levels. (ii) Circulating adiponectin levels were related to several baseline clinical and laboratory parameters. There was a clear positive relationship with NT-proBNP concentrations and HDL-cholesterol, and a negative association with TAGs and BMI. The present study extends these findings by including exercise-testing-derived prognostic markers and anthropometric parameters. (iii) Our results confirm that CHF patients are characterized by elevated adiponectin levels, which increase with disease severity. These findings support a role for adiponectin as a biomarker in CHF [11,14].

Adiponectin is an adipocytokine with known anti-inflammatory, anti-atherogenic and insulin-sensitizing properties [9]. Adiponectin plasma levels have been found to be reduced in patients with Type 2 diabetes mellitus, obese subjects and patients with coronary artery disease. Moreover, lifestyle improvements, such as weight loss, increase adiponectin concentration. Paradoxically, in CHF patients, high adiponectin levels are associated with adverse outcome. Kistorp et al. [10] showed for the first time in 195 CHF patients that increased adiponectin is associated with poor outcome. This finding has been confirmed by others [11,15], demonstrating that adiponectin is an independent prognostic predictor of mortality and morbidity in CHF.

Despite all of the observational findings published recently concerning adiponectin and CHF, the mechanisms for an increase in the concentration of adiponectin in CHF remain to be elucidated.

Causes of increased adiponectin levels in CHF

Several hypotheses have been proposed to explain elevated adiponectin levels in CHF. CHF patients are characterized by numerous metabolic disturbances, among which insulin resistance and increased concentrations of NEFAs [non-esterified fatty acids (‘free fatty acids’)] have been repeatedly demonstrated [19,20]. The existence of a hyperadrenergic state that increases the concentration of NEFAs in plasma in CHF patients had been identified a number of years ago [20]; however, the crucial role of NEFAs in insulin resistance and its complications in the setting of CHF [2,16,21,22] have only gained attention more recently. As adiponectin regulates glucose and fatty acid metabolism, increased adiponectin secretion might be a compensatory mechanism to attenuate NEFA-induced insulin resistance [14,15].

CHF patients are marked by a chronic low-grade inflammatory state, characterized by increased levels of circulating pro-inflammatory cytokines. The well-known anti-inflammatory effects of adiponectin point to a possible physiological response in an attempt to suppress inflammation [9,14].

In accordance with previous studies [10,1214], we have also shown in the present study a strong positive relationship between adiponectin levels and NT-proBNP concentrations. Adiponectin and natriuretic peptides behave similarly in CHF: both factors increase with disease severity, which makes these circulating peptides of great potential interest in serving as biomarkers, and it might be hypothesized that their increase is a compensatory mechanism. Sengenes et al. [23] identified a lipolytic and potential lipid-mobilizing effect of natriuretic peptides, which was mediated by specific adipocyte membrane receptors. It has been hypothesized that natriuretic peptides, through an increase in lipid mobilization, indirectly stimulate adiponectin levels [10]. A recent study [16] has demonstrated that exogenous ANP (atrial natriuretic peptide; carperitide) infusion for 7 days increased plasma total and high-molecular-mass adiponectin, thereby supporting previous statements that natriuretic peptides increase adiponectin levels via a functional GC-A receptor. In addition, two recently published studies, dealing with the relationship between adiponectin and BNP (brain natriuretic peptide) [24,25], endorse the fact that natriuretic peptides might have lipid-mobilizing effects.

Another explanation implies the existence of a functional adiponectin resistance [17]. This hypothesis is supported by a study published previously showing a decrease in mRNA and protein expression of the adiponectin receptor AdipoR1 in the left ventricle of infarcted mouse hearts compared with normal hearts [26]. Hyperadiponectinaemia may reflect defective compensatory mechanisms to overcome this adiponectin resistance.

Contrary with the suggested compensatory mechanisms, Nørrelund et al. [15] have suggested that high adiponectin levels might not be beneficial in patients with established CHF. On the basis of the proposed role of adiponectin to increase energy expenditure and induce weight loss through a direct effect on the brain [27], previous studies consider high adiponectin levels as a marker of the wasting process that characterizes severely ill CHF patients [10,15].

Adiponectin and lipid profile

In spite of the paradoxical behaviour of adiponectin in CHF in terms of prognostic relevance, the anti-atherogenic association of adiponectin appears to be preserved. We found a positive and independent association between HDL-cholesterol and adiponectin, and a negative association with TAG. Our present results are consistent with the hypothesis of von Eynatten et al. [28], who postulated that adiponectin may mediate part of its proposed anti-atherogenic properties by influencing HDL-cholesterol concentrations.

Although an atherogenic lipid profile is a well-known risk factor for the development of coronary artery disease and, hence, of CHF, it has indeed been shown that once CHF is established, these factors may provide protection against further progression, a concept known as reverse epidemiology [29]. The remaining anti-atherogenic property of adiponectin reflects the dyslipidaemic state in CHF. Adiponectin might be increased to overcome this dyslipidaemic state and thereby even worsening outcome in CHF patients. It might also be that adiponectin levels and the dyslipidaemic profile are not aetiologically linked to disease progression in CHF, but that they are simply markers of disease severity.

Adiponectin and the effect of exercise training

In the present study, there was a clear reduction in adiponectin levels in CHF patients following exercise training, independent of possible confounders such as gender, lean body mass, exercise capacity, lipid profile, LVEF and NT-proBNP. The clinical relevance of this decrease is reflected in parallel and favourable changes in terms of exercise capacity as well as in a significant improvement in NYHA functional class. In addition, we found an important association of adiponectin with anthropometric data and exercise capacity, indicating the involvement of adiponectin in the regulation of whole-body energy metabolism.

The present study is the first to investigate the modulation of adiponectin levels in CHF. Several hypotheses can be put forward to explain the observed effect. First, the role of adiponectin in mitochondrial dysfunction might be relevant. The known impact of adiponectin on mitochondrial biogenesis in skeletal muscle [30], which is severely disrupted in CHF [31], might be affected by exercise training. Physical exercise is known to improve aerobic metabolism in CHF [32] and to partly reverse mitochondrial abnormalities. A second hypothesis relies on the role of adiponectin in the metabolic impairment in CHF, characterized by insulin resistance and increased circulating levels of fatty acids [15]. Exercise training might break through this metabolic vicious cycle and conquer adiponectin resistance. Improvement in insulin resistance, which is a well-known effect of exercise training in healthy subjects and diabetics, might subsequently result in lower adiponectin levels. In addition, the anti-inflammatory role of exercise training, as demonstrated by our group [5] and others [6], could lower adiponectin levels. Finally, the close relationship demonstrated between circulating adiponectin and natriuretic peptides on the one hand and the fact that exercise training lowers NT-proBNP levels [7] supports the proposed co-regulatory role of NEFA metabolism and natriuretic peptides [14,33].

Limitations

First, the present study was designed as a prospective single-centre study comparing an exercise-trained CHF group with an untrained CHF control group. Randomization between the two treatment options would have increased the strength of the conclusions; however, considering the clear benefit of exercise training in CHF and the fact that patients were specifically referred to a tertiary Cardiac Rehabilitation Centre for exercise training, we considered randomization ethically unacceptable. Although the difference was not statistically significant, it may appear that our trained patients are more severely diseased. This notion is based on increased NT-proBNP levels and lower LVEF; however, another important prognostic parameter and indicator of disease severity, V̇O2peak, was similar in both groups.

Secondly, two different training protocols were applied: ET and CT. Both training modalities were routinely used and are considered as standard treatment at our Rehabilitation Centre at the time of inclusion. Depending on the patients' characteristics, such as severe muscle wasting, dynamic resistive exercises were incorporated into the training programme. The aim of the present study was to investigate the effect of an intervention, i.e. a training programme, on circulating adiponectin levels in the CHF population, without differentiation according to training modalities. A subanalysis performed in the present study population did not detect any difference between the type of exercise training (ET compared with CT) on circulating adiponectin levels. However, we admit that this might be due to the relatively small numbers of patients in each training group. The elaboration of a possible differential effect of the type of exercise training will be the subject of future studies.

Thirdly, the possible effect of drugs on circulating adiponectin levels should be taken into account. ACEIs (angiotensin-converting enzyme inhibitors) have been reported to increase circulating adiponectin levels, whereas β-blockers decrease adiponectin levels [34]. These drugs were, however, routinely prescribed to all CHF patients, with no significant differences in the use between the trained and untrained CHF patients. In addition, medical treatment was not changed during the study period. The effect of statins on circulating adiponectin levels is dependent on the population studied and remains a controversial topic. The untrained CHF group in our present study appears to have a higher percentage of statin intake. The treatment, however, was initiated well before enrolment in the study and remained stable during the study period. Therefore it appears unlikely that this issue explains the evolution of adiponectin.

Conclusions

In conclusion, the present study has demonstrated for the first time that high adiponectin levels are reduced following exercise training in patients with CHF. The importance of adiponectin in CHF might be extended from its role as a biomarker with proven prognostic value to a pathophysiological link between the energetic (mitochondrial dysfunction) and metabolic (insulin resistance) abnormalities in CHF. Our present findings, together with those from other studies, suggest that dysregulation of the adiponectin pathway contributes to the observed metabolic impairment in CHF. Further exploration of the adiponectin signalling pathway downstream the receptor in skeletal muscle might allow potential target-specific pharmacotherapy and might provide a fundamental insight in the modulatory role of exercise training.

FUNDING

This work is supported by the Fund for Scientific Research (FWO) – Flanders (Belgium). A. M. V. B. is supported by a Ph.D. Fellowship and P. B. is the recipient of a special Ph.D. Fellowship from the Fund for Scientific Research (FWO) – Flanders (Belgium). V. M. C. is a Senior Clinical Investigator supported by the Fund for Scientific Research (FWO) – Flanders (Belgium).

We thank Viviane Van Hoof, Manou Martin and co-workers for assistance with the NT-proBNP assays. We also thank Kurt Wuyts for assistance in CPET, peripheral muscle strength measurements and the Cardiac Rehabilitation training programme.

Abbreviations

     
  • ACEI

    angiotensin-converting enzyme inhibitor

  •  
  • BMI

    body mass index

  •  
  • CHF

    chronic heart failure

  •  
  • CPET

    cardiopulmonary exercise testing

  •  
  • CRP

    C-reactive protein

  •  
  • CT

    combined endurance/resistance training

  •  
  • ET

    endurance training

  •  
  • HDL

    high-density lipoprotein

  •  
  • LDL

    low-density lipoprotein

  •  
  • LVEF

    left ventricular ejection fraction

  •  
  • NEFA

    non-esterified fatty acids fatty acid (‘free fatty acid’)

  •  
  • NT-proBNP

    N-terminal pro-brain natriuretic peptide

  •  
  • NYHA

    New York Heart Association

  •  
  • TAG

    triacylglycerol (triglyceride)

  •  
  • V̇O2

    oxygen uptake

  •  
  • V̇O2peak

    peak V̇O2

References

References
1
Doehner
W.
von Haehling
S.
Anker
S. D.
Insulin resistance in chronic heart failure
J. Am. Coll. Cardiol.
2008
, vol. 
52
 (pg. 
239
-
240
)
2
Opie
L. H.
The metabolic vicious cycle in heart failure
Lancet
2004
, vol. 
364
 (pg. 
1733
-
1734
)
3
Ashrafian
H.
Frenneaux
M. P.
Opie
L. H.
Metabolic mechanisms in heart failure
Circulation
2008
, vol. 
116
 (pg. 
434
-
448
)
4
Piepoli
M. F.
Davos
C.
Francis
D. P.
Coats
A. J.
Exercise training meta-analysis of trails in patients with chronic heart failure (ExTraMATCH)
Br. Med. J.
2004
, vol. 
328
 pg. 
189
 
5
Conraads
V. M.
Beckers
P.
Bosmans
J.
De Clerck
L. S.
Stevens
W. J.
Vrints
C. J.
Brutsaert
D. L.
Combined endurance/resistance training reduces plasma TNF-α receptor levels in patients with chronic heart failure and coronary artery disease
Eur. Heart J.
2002
, vol. 
23
 (pg. 
1854
-
1860
)
6
Adamopoulos
S.
Parissis
J.
Karatzas
D.
Kroupis
C.
Georgiadis
M.
Karavolias
G.
Paraskevaidis
J.
Koniavitou
K.
Coats
A. J.
Kremastinos
D. T.
Physical training modulates proinflammatory cytokines and the soluble Fas/soluble Fas ligand system in patients with chronic heart failure
J. Am. Coll. Cardiol.
2002
, vol. 
39
 (pg. 
653
-
663
)
7
Conraads
V. M.
Beckers
P.
Vaes
J.
Martin
M.
Van Hoof
V.
De Maeyer
C.
Possemiers
N.
Wuyts
F. L.
Vrints
C. J.
Combined endurance/resistance training reduces NT-pro-BNP levels in patients with chronic heart failure
Eur. Heart J.
2004
, vol. 
25
 (pg. 
1797
-
1805
)
8
Linke
A.
Adams
V.
Schulze
P. C.
Erbs
S.
Gielen
S.
Fiehn
E.
Möbius-Winkler
S.
Schubert
A.
Schuler
G.
Hambrecht
R.
Antioxidative effects of exercise training in patients with chronic heart failure: increase in radical scavenger enzyme activity in skeletal muscle
Circulation
2005
, vol. 
111
 (pg. 
1763
-
1770
)
9
Kadowaki
T.
Yamauchi
T.
Adiponectin and adiponectin receptors
Endocr. Rev.
2005
, vol. 
26
 (pg. 
439
-
451
)
10
Kistorp
C.
Faber
J.
Galatius
S.
Gustafsson
F.
Frystyk
J.
Flyvbjerg
A.
Hildebrandt
P.
Plasma adiponectin, body mass index, and mortality in patients with chronic heart failure
Circulation
2005
, vol. 
112
 (pg. 
1756
-
1762
)
11
George
J.
Patal
S.
Wexler
D.
Sharabi
Y.
Peleg
E.
Kamari
Y.
Grossman
E.
Sheps
D.
Keren
G.
Roth
A.
Circulating adiponectin concentrations in patients with congestive heart failure
Heart
2008
, vol. 
92
 (pg. 
1420
-
1424
)
12
Tamura
T.
Furukawa
Y.
Taniguchi
R.
Sato
Y.
Ono
K.
Horiuchi
H.
Nakagawa
Y.
Kita
T.
Kimura
T.
Serum adiponectin level as an independent predictor of mortality in patients with congestive heart failure
Circ. J.
2008
, vol. 
71
 (pg. 
623
-
630
)
13
Tsutamoto
T.
Tanaka
T.
Sakai
H.
Ishikawa
C.
Fujii
M.
Yamamoto
T.
Horie
M.
Total and high molecular weight adiponectin, haemodynamics, and mortality in patients with chronic heart failure
Eur. Heart J.
2007
, vol. 
28
 (pg. 
1723
-
1730
)
14
McEntegart
M. B.
Awede
B.
Petrie
M. C.
Sattar
N.
Dunn
F. G.
MacFarlane
N. G.
McMurray
J. J.
Increase in serum adiponectin concentrations in patients with heart failure and cachexia: relationship with leptin, other cytokines, and B-type natriuretic peptide
Eur. Heart J.
2007
, vol. 
28
 (pg. 
829
-
835
)
15
Nørrelund
H.
Wiggers
H.
Halbirk
M.
Frystyk
J.
Flyvbjerg
A.
Bøtker
H. E.
Schmitz
O.
Jørgensen
J. O.
Christiansen
J. S.
Møller
N.
Abnormalities of whole body protein turnover, muscle metabolism and levels of metabolic hormones in patients with chronic heart failure
J. Intern. Med.
2006
, vol. 
260
 (pg. 
11
-
26
)
16
Tanaka
T.
Tsutamoto
T.
Sakai
H.
Nishiyama
K.
Fujii
M.
Yamamoto
T.
Horie
M.
Effect of atrial natriuretic peptide on adiponectin in patients with heart failure
Eur. J. Heart Failure
2008
, vol. 
10
 (pg. 
360
-
366
)
17
Kintscher
U.
Does adiponectin resistance exist in chronic heart failure?
Eur. Heart J.
2007
, vol. 
28
 (pg. 
1676
-
1677
)
18
Beckers
P. J.
Denollet
J.
Possemiers
N. M.
Wuyts
F. L.
Vrints
C. J.
Conraads
V. M.
Combined endurance-resistance training vs. endurance training in patients with chronic heart failure: a prospective randomized study
Eur. Heart J.
2008
, vol. 
29
 (pg. 
1858
-
1866
)
19
Doehner
W.
Rauchhaus
M.
Ponikowski
P.
Godsland
I. F.
von Haehling
S.
Okonko
D. O.
Leyva
F.
Proudler
A. J.
Coats
A. J.
Anker
S. D.
Impaired insulin sensitivity as an independent risk factor for mortality in patients with chronic heart failure
J. Am. Coll. Cardiol.
2005
, vol. 
46
 (pg. 
1019
-
1026
)
20
Lommi
J.
Kupari
M.
Yki-järvinen
H.
Free-fatty acid kinetics and oxidation in congestive heart failure
Am. J. Cardiol.
1998
, vol. 
81
 (pg. 
45
-
50
)
21
Shulman
G. I.
Cellular mechanisms of insulin resistance
J. Clin. Invest.
2000
, vol. 
106
 (pg. 
171
-
176
)
22
Murray
A. J.
Anderson
R. E.
Watson
G. C.
Radda
G. K.
Clarke
K.
Uncoupling proteins in human heart
Lancet
2004
, vol. 
364
 (pg. 
1786
-
1788
)
23
Sengenes
C.
Berlan
M.
De Glisezinski
I.
Lafontan
M.
Galitzky
J.
Natriuretic peptides: a new lipolytic pathway in human adipocytes
FASEB J.
2000
, vol. 
14
 (pg. 
1345
-
1351
)
24
Ang
D. S.
Welsh
P.
Watt
P.
Nelson
S. M.
Struthers
A.
Sattar
N.
Serial changes in adiponectin and BNP in ACS patients: paradoxical associations with each other and prognosis
Clin. Sci.
2009
, vol. 
117
 (pg. 
41
-
48
)
25
Tsukamoto
O.
Fujita
M.
Kato
M.
Yamazaki
S.
Asano
Y.
Ogai
A.
Okazaki
H.
Asai
M.
Nagamachi
Y.
Maeda
N.
, et al. 
Natriuretic peptides enhance the production of adiponectin in human adipocytes and in patients with chronic heart failure
J. Am. Coll. Cardiol.
2009
, vol. 
53
 (pg. 
2070
-
2077
)
26
Von Haehling
S.
Doehner
W.
Anker
S. D.
Nutrition, metabolism, and the complex pathophysiology of cachexia in chronic heart failure
Cardiovasc. Res.
2007
, vol. 
73
 (pg. 
298
-
309
)
27
Qi
Y.
Takahashi
N.
Hileman
S. M.
Patel
H. R.
Berg
A. H.
Pajvani
U. B.
Scherer
P. E.
Ahima
R. S.
Adiponectin acts in the brain to decrease body weight
Nat. Med.
2004
, vol. 
10
 (pg. 
524
-
529
)
28
von Eynatten
M.
Hamann
A.
Twardella
D.
Nawroth
P. P.
Brenner
H.
Rothenbacher
D.
Relationship of adiponectin with markers of systemic inflammation, atherogenic dyslipidemia, and heart failure in patients with coronary heart disease
Clin. Chem.
2006
, vol. 
52
 (pg. 
853
-
859
)
29
Kalantar-Zadeh
K.
Block
G.
Horwich
T.
Fonarow
G. C.
Reverse epidemiology of conventional cardiovascular risk factors in patients with chronic heart failure
J. Am. Coll. Cardiol.
2004
, vol. 
43
 (pg. 
1439
-
1444
)
30
Civitarese
A. E.
Ukropcova
B.
Carling
S.
Hulver
M.
DeFronzo
R. A.
Mandarino
L.
Ravussin
E.
Smith
S. R.
Role of adiponectin in human skeletal muscle bioenergetics
Cell Metab.
2006
, vol. 
4
 (pg. 
75
-
87
)
31
Mudd
J. O.
Kass
D. A.
Tackling heart failure in the twenty-first century
Nature
2008
, vol. 
451
 (pg. 
919
-
928
)
32
Ventura-Clapier
R.
Mettauer
B.
Bigard
X.
Beneficial effects of endurance training on cardiac and skeletal muscle energy metabolism in heart failure
Cardiovasc. Res.
2007
, vol. 
73
 (pg. 
10
-
18
)
33
Haugen
E.
Furukwawa
Y.
Isic
A.
Fu
M.
Increased adiponectin level in parallel with increased NT-pro BNP in patients with severe heart failure in the elderly: A hospital cohort study
Int. J. Cardiol.
2008
, vol. 
125
 (pg. 
216
-
219
)
34
Yamaji
M.
Tsutamoto
T.
Tanaka
T.
Kawahara
C.
Nishiyama
K.
Yamamoto
T.
Fujii
M.
Horie
M.
Effect of carvedilol on plasma adiponectin concentration in patients with chronic heart failure
Circ. J.
2009
, vol. 
73
 (pg. 
1067
-
1073
)