Ghrelin resistance occurs in severe heart failure and resolves after heart transplantation

Abstract

Aims

Severe heart failure (HF) is often associated with cachexia that reverses post-heart transplantation (HTx) with frequent development of obesity. Ghrelin is a novel appetite-stimulating hormone. The aim was to determine the role of ghrelin in regulating appetite, food intake, and body composition in HF and post-HTx.

Methods and results

We measured serial ghrelin, hunger sensation, caloric intake, and body composition in 12 HF patients awaiting HTx, 12 patients 12.7 ± 8.6 months post-HTx, and 7 controls. Seven of 12 HF patients were followed for longitudinal analysis post-HTx. Body mass index was 23.1 ± 3.1 in HF and 31.5 ± 5.5 post-HTx (P < 0.001). Heart transplantation patients had gained 18.0 ± 7.7 kg since HTx. Ghrelin area under the curve between controlled meals (control: 186 ± 39; HF: 264 ± 71; HTx: 194 ± 47 ng min/mL, P < 0.007) was higher in HF, but test meal caloric intake (control: 1185 ± 650; HF: 391 ± 103; HTx: 831 ± 309 kcal, P < 0.008) was lower in HF. The longitudinal analysis confirmed these findings.

Conclusion

Heart failure may be associated with resistance to the appetite-stimulating effects of ghrelin, which may contribute to cachexia. Heart transplantation may be associated with resolution of ghrelin resistance, which may contribute to weight gain. These findings are preliminary and should be confirmed in larger trials.

Introduction

Cardiac cachexia is well described in severe heart failure (HF).^1,2 Heart transplantation (HTx) is associated with weight gain and the frequent development of obesity. Although this weight gain has been attributed to glucocorticoid therapy, it is more dramatic than after other solid organ transplants, despite similar steroid regimens, and also unrelated to steroid dose,³ suggesting that other mechanisms are responsible.

Ghrelin is an appetite-stimulating hormone that is released from the stomach in response to fasting and weight loss and is inhibited by food intake.^4,5 Ghrelin was originally identified as the endogenous ligand for the growth hormone (GH) secretagogue receptor and partially acts by stimulating GH release.⁴ Ghrelin receptors are widely distributed in the heart and vessels,⁶ and its administration may improve left ventricular function, decrease muscle wasting, and improve exercise capacity in HF.⁷ In one study, ghrelin was elevated in cachectic HF.⁸ Ghrelin has not been studied post-HTx.

We hypothesized that cachexia in severe HF may be in part due to resistance to the appetite-stimulating and anabolic effects of ghrelin and that weight gain post-HTx may be in part due to resolution of this resistance. Accordingly, we examined the hormonal regulation of appetite, caloric intake, and body composition in HF, HTx, and matched healthy controls.

Methods

We performed a cross-sectional study in 12 patients with New York Heart Association functional class IV HF awaiting HTx, 12 patients 12.7 ± 8.6 months post-HTx, and 7 control subjects. We then performed a longitudinal study of 7 of 12 HF patients 7.4 ± 4.4 months after HTx. Of the remaining five HF patients, three patients declined repeat testing after HTx, one died before HTx, and one became too ill for HTx. Subjects were age and gender matched. Heart failure and HTx patients were matched by aetiology, and controls were matched by body mass index (BMI) to HTx (Table 1). The protocol complied with the Declaration of Helsinki and was approved by the institutional review board of Columbia University Medical Center, and written informed consent was obtained from all patients.

Table 1.

Baseline characteristics of the cross-sectional cohort

Group	Control (n = 7)	HF (n = 12)	HTx (n = 12)	P-value (overall)
Age (years)	42 ± 14	52 ± 16	47 ± 18	>0.2
Gender n (%)
Male	5 (71%)	10 (83%)	11 (92%)	>0.2
Female	2 (29%)	2 (17%)	1 (8%)
BMI (kg/m2)	27.8 ± 4.2	23.1 ± 3.1	31.5 ± 5.5	<0.001
Cachexia presenta	NA	7 (58%)	NA	NA
Diabetes	0	3 (25%)	2 (17%)	NA
Weight gain since HTx (kg)	NA	NA	18.0 ± 7.7	NA
Time since HTx (months)	NA	NA	12.7 ± 8.6	NA
Aetiology
Ischaemic	NA	4 (33%)	6 (50%)	>0.2
Dilated		6 (50%)	5 (42%)
Valvular		2 (17%)	1 (8%)
HF drug therapy
Loop diuretic	NA	12 (100%)	NA	NA
ACE-inhibitor		12 (100%)
Beta-blocker		5 (42%)
Inotropes		7 (58%)
HTx drug therapy
Prednisone (mg)	NA	NA	6 ± 4	NA
Cyclosporine A			9 (75%)
Tacrolimus			3 (25%)
Mycophenolate mofetil			12 (100%)

NA, not applicable.

^aCachexia was defined as non-oedematous weight loss ≥7.5% over the preceding 6 months.^8,9

Bold text denotes P ≤ 0.05.

DEXA scanning to analyse body composition was performed with a QDR 4500 A Delphi W densitometer (Hologic Inc., Bedford, USA). Waist circumference was measured in duplicate at the level of the umbilicus. Measurement of resting energy expenditure (REE) and cardiopulmonary exercise testing with measurement of peak oxygen consumption (peak VO₂) was performed using a metabolic cart (Medical Graphics, Minneapolis, USA).

Subjects reported to the laboratory at 8 a.m. in the fasting state and underwent blood sampling, completed a visual analogue (0–10) assessing hunger, and then consumed a yoghurt-based breakfast test meal where they were encouraged to eat until absolute satiety. Eating until absolute satiety ensured that post-prandial assessment would not be affected by differences in baseline food intake and satiety. Post-prandial serial blood testing and hunger assessment was performed hourly for 5 h. Subjects were then given a lunch meal of regular food of their choosing and were neither encouraged nor dissuaded to eat. Caloric intake at each meal was estimated using computerized software NDS-R 4.05-33 (University of Minnesota, Minneapolis, USA).

Blood was collected in chilled EDTA tubes; 2 mL blood was mixed with 0.1 mL aprotonin (for ghrelin but not GH) and centrifuged at 2500 g for 15 min at 4°C. Plasma was stored at −70°C. Ghrelin was measured by radioimmunoassay (Phoenix Pharmaceuticals, Belmont, USA) and GH was measured by a two-site immunoradiometric assay (Diagnostic Systems Laboratories, Inc., Webster, USA).

Continuous variables were compared by one-way analysis of variance. Pair-wise comparisons between cross-sectional groups were done by Fisher's least significance test. Paired comparisons between longitudinal groups were done by Student's t-test. Serial data over time were used to calculate area under the curve (AUC) for ghrelin. Ghrelin and GH were plotted against each other and Pearson's correlation calculated. A P < 0.05 was considered statistically significant. All results are reported as mean ± standard deviation.

Results

Clinical characteristics and medical regimens of the cross-sectional population are summarized in Table 1. Cachexia, defined as non-oedematous weight loss ≥7.5% over the preceding 6 months,^8,9 was present in 7 of 12 HF patients. Heart transplantation patients had gained 18.0 ± 7.7 kg over 12.7 ± 8.6 months since HTx. The mean prednisone dose in the HTx patients was 6 ± 4 mg/day (range 2–15). The seven HF patients followed longitudinally (Table 3) had similar age, gender, and medical regimens compared with the 12 cross-sectional HTx patients (Table 1).

Table 3.

Baseline characteristics of the longitudinal cohort

Group	HF long (n = 7)	HTx long (n = 7)	P-value
Age at time of respective study (years)	47 ± 19	48 ± 19	NA
Gender n (%)
Male	5 (71%)	Same	NA
Female	2 (29%)
BMI (kg/m2)	23.4 ± 3.3	26.4 ± 4.9	<0.009
Diabetes n (%)	2 (29%)	2 (29%)	NA
Weight gain since HTx (kg)	NA	9.6 ± 6.2	NA
Time since HTx (months)	NA	7.4 ± 4.4	NA
Aetiology
Ischaemic	1 (14%)	Same	NA
Dilated	5 (71%)
Valvular	1 (14%)
HF drug therapy
Loop diuretic	7 (100%)	NA	NA
ACE-inhibitor	7 (100%)
Beta-blocker	3 (43%)
Inotropes	5 (71%)
HTx drug therapy
Prednisone (mg)	NA	6 ± 4	NA
Cyclosporine A		3 (43%)
Tacrolimus		4 (57%)
Mycophenolate mofetil		6 (86%)

NA, not applicable.

Bold text denotes P ≤ 0.05.

Results of the cross-sectional study are depicted in Table 2. Heart failure had the lowest BMI, total body mass, lean mass, and fat mass (P < 0.01 for all except lean mass). The per cent body fat, waist circumference, and abdominal fat mass were greater in HTx than in HF and controls (all P < 0.01).

Table 2.

Body composition, resting energy expenditure, exercise tolerance, hormones, appetite, and caloric intake in the cross-sectional cohort

Group	Control (n = 7)	HF cross-sectional (n = 12)	HTx cross-sectional (n = 12)	P-value (overall)	P-value (control vs. HF)	P-value (control vs. HTx)	P-value (HF vs. HTx)
BMI (kg/m2)	27.8 ± 4.2	23.1 ± 3.1	31.5 ± 5.5	<0.001	0.03	0.08	<0.001
Total mass (kg)	87 ± 15	69 ± 6	92 ± 15	<0.001	0.004	>0.2	<0.001
Fat mass (kg)a	24 ± 9	17 ± 2	31 ± 9	0.006	<0.13	0.10	<0.002
Lean mass (kg)a	66 ± 8	55 ± 4	61 ± 11	0.12	0.04	>0.2	0.17
Fat%	26 ± 7	23 ± 2	33 ± 7	<0.009	>0.2	<0.03	<0.004
Abdominal fat mass (kg)	11.7 ± 5	8.4 ± 2	16.8 ± 6	<0.01	>0.2	<0.06	<0.004
Waist circumference (cm)	99 ± 10	86 ± 4	111 ± 13	<0.001	<0.10	<0.05	<0.001
REE (kcal/day)	1710 ± 384	1812 ± 352	2036 ± 528	>0.2	>0.2	<0.13	>0.2
REE (kcal/m2/h)	35 ± 5	41 ± 9	41 ± 10	>0.2	<0.14	<0.15	>0.2
Peak VO2 (mL/kg/min)a	27 ± 9	11 ± 2	15 ± 3	<0.001	<0.001	<0.001	0.18
GH (ng/mL)	0.21 ± 0.25	1.13 ± 1.19	0.07 ± 0.13	<0.006	<0.02	>0.2	<0.003
Ghrelin (pg/mL)	669 ± 85	917 ± 336	690 ± 194	0.05	0.04	>0.2	<0.04
Hunger fasting (0–10)	5.8 ± 2.5	5.0 ± 2.4	3.4 ± 1.3	<0.05	>0.2	0.02	0.07
Breakfast (kcal)	907 ± 260	832 ± 364	1238 ± 421	0.03	>0.2	<0.07	0.01
Ghrelin AUC (ng min/mL)	186 ± 39	264 ± 71	194 ± 47	<0.007	<0.007	>0.2	<0.006
Hunger before lunch (0–10)	7.3 ± 1.6	6.6 ± 2.8	6.1 ± 2.7	>0.2	>0.2	>0.2	>0.2
Lunch (kcal)	1185 ± 650	391 ± 103	831 ± 309	<0.008	<0.002	0.13	<0.07

^aPeak VO₂ and DEXA scans were not performed in the seven HF patients on inotropic therapy.

Bold text denotes P ≤ 0.05.

Fasting ghrelin and GH were substantially greater in HF (P = 0.05 and P < 0.006, respectively). Ghrelin correlated with GH in HF (Pearson's R = 0.76, P = 0.01) and overall (Pearson's R = 0.74, P < 0.001). At the encouraged breakfast, HF and controls had similar but HTx had higher caloric intake (P = 0.03). After eating to satiety, HF still released more ghrelin than did HTx and control, as judged by the AUC of ghrelin over a 5 h period (P < 0.007, Table 2 and Figure 1). Despite this increased exposure to ghrelin in HF, 5 h after the test meal the sensation of hunger was similar in all groups (Table 2 and Figure 1). Despite higher ghrelin and similar hunger levels, HF patients ingested dramatically less calories at the ad lib meal than HTx and control (P < 0.008, Table 2 and Figure 1). The REE was similar between the groups crudely and after adjusting for body surface area.

Figure 1.

Ghrelin area under the curve, hunger, and caloric intake in the longitudinal and cross-sectional cohorts. All subjects consumed a breakfast until absolute satiety to ensure that post-prandial assessments were not affected by differences in baseline food intake or satiety. For 5 h thereafter, the ghrelin area under the curve (AUC, ng min/mL) was significantly higher in HF than in control or HTx. Despite higher ghrelin, hunger was no different in HF compared with control or HTx. At the ad lib lunch meal, when subjects were neither encouraged nor dissuaded to eat, caloric intake was significantly and dramatically lower in HF than in control or HTx, despite significantly higher ghrelin. Hunger is measured on a 0–10 visual analogue scale. x-sect, cross-sectional; long, longitudinal.

The seven longitudinal patients were studied before and 7.4 ± 4.4 months after HTx and had gained 9.6 ± 6.2 kg since HTx. The longitudinal (Tables 3 and 4) exhibited the same patterns as the cross-sectional (Tables 1 and 2) study. Weight gain after HTx and BMI exhibited the same pattern, although the changes were smaller in the longitudinal cohort, as the time since HTx was less (Tables 1 and 3).

Table 4.

Hormones, appetite, and caloric intake in the longitudinal cohort

Group	HF long (n = 7)	HTx long (n = 7)	P-value
BMI (kg/m2)	23.4 ± 3.3	26.4 ± 4.9	<0.009
GH (ng/mL)	1.21 ± 0.64	0.17 ± 0.15	0.03
Ghrelin fasting (pg/mL)	792 ± 188	683 ± 125	0.15
Hunger fasting (0–10)	5.4 ± 2.1	6.9 ± 2.0	>0.2
Breakfast (kcal)	750 ± 345	773 ± 279	>0.2
Ghrelin AUC (ng min/mL)	239 ± 48	200 ± 32	<0.04
Hunger before lunch (0–10)	6.6 ± 2.5	5.8 ± 2.1	>0.2
Lunch (kcal)	428 ± 166	1058 ± 71	<0.17
REE (kcal/day)	1669 ± 354	1701 ± 122	>0.2
REE (kcal/m2/h)	38 ± 9	37 ± 4	>0.2

Bold text denotes P ≤ 0.05.

Discussion

Cardiac cachexia, defined as a weight loss of ≥7.5% over the preceding 6 months,^8,9 is harmful regardless of BMI.² All HF patients had a BMI less than 30 and 7 of 12 fulfilled criteria for cardiac cachexia, similar to reported prevalences of cardiac cachexia of 16–68%.^1,10,11 Heart transplantation patients had gained 18 kg since HTx, both lean tissue and fat, consistent with previous reports.^3,12 Peak VO₂ was dramatically impaired in HF, consistent with the severity of the HF. The mechanisms underlying cachexia in HF have been suggested to be partially due to neurohormonal and cytokine activation and GH resistance,^9,13 but other mechanisms are possible.

Ghrelin has received attention as an appetite stimulant, but receptors are found in heart and vessels⁶ and numerous cardiovascular effects have been described.^14,15 In human HF, ghrelin administration may improve left ventricular function and cardiac output, decrease muscle wasting, and improve exercise capacity.^7,14,15 Growth hormone may decrease ghrelin in a negative-feedback loop.¹⁶ One study of ghrelin in HF demonstrated elevated ghrelin in cachectic but not in non-cachectic HF, as well as a correlation between GH and ghrelin.⁸ Here, we show elevated ghrelin in HF regardless of BMI and confirm the correlation between GH and ghrelin in both HF and the overall cohort. Ghrelin post-HTx has not previously been studied. We show that fasting ghrelin levels and ghrelin AUC normalize after HTx.

The mechanism for elevated ghrelin levels in HF as in other weight reduced states^5,17,18 is not clear. It may be a physiological compensation for reduced weight in an effort to increase appetite and caloric intake.^17,18 Despite elevated fasting ghrelin, HF patients had no higher hunger than the other groups and actually had lower caloric intake than HTx. Ghrelin levels are known to fall after food intake.⁵ Following the test breakfast meal, ghrelin levels fell more slowly and remained higher, as reflected by a higher AUC, in HF than in control or HTx, suggesting that food intake does not appropriately feedback on ghrelin secretion in HF. Though the AUC of ghrelin was greatest in HF, hunger before lunch was equal among the groups, and ad lib caloric intake at lunch was significantly lower in HF than in HTx and control. At breakfast, all subjects were actively encouraged to consume as much as possible, which may be responsible for the caloric intake in HF being similar to control and only one-third less than HTx. In contrast, the lunch meal was neither encouraged nor un-encouraged, and HF patients consumed considerably less. Resting energy expenditure was similar in all groups and thus does not explain different caloric intakes. These data suggest that HF subjects are resistant to the orexigenic effects of ghrelin, similar to patients with anorexia nervosa who have high ghrelin levels without appetite stimulation or weight gain.^17,18 These data also suggest that ghrelin resistance resolves after HTx.

However, since ghrelin was also elevated in non-cachectic HF and did not correlate with BMI in this study, other explanations for elevated ghrelin are possible. The beneficial cardiovascular effects of ghrelin suggest that elevations in ghrelin could be a compensatory response to low-cardiac output and maladaptive neurohormonal features of HF. Ghrelin's action as a GH secretagogue raises the possibility that ghrelin could be elevated in response to GH resistance, which occurs in severe HF.^13,19 The fall in ghrelin coupled with increased caloric intake post-HTx may reflect resolution of resistance to the orexigenic effects of ghrelin. Again, however, other explanations are possible, including resolution of HF and its associated neurohormonal activation and/or resolution of GH resistance.¹³

The small patient population is the major limitation of the study. However, the consistency of findings and magnitude of differences in the cross-sectional and longitudinal analysis support the conclusions. The complexity of the measurements made it difficult to involve other HTx centres in the study. We cannot rule out that steroids affect weight gain and ghrelin regulation post-HTx, but steroid doses were relatively low and we have previously shown that weight gain post-HTx is unrelated to steroid dose.³ Our findings need to be confirmed in a larger study population in a multi-centre trial. The potential relationship between appetite and ghrelin should be analysed with caution. Recent reports suggest that ghrelin exists both in proforms and in active (acylated) and inactive (non-acylated) forms.²⁰ We measured total ghrelin, which reflects both. However, measurement of active ghrelin is difficult and current assays may be inaccurate.

Conclusion

In this pilot study, we show that HF may be associated with resistance to the appetite-stimulating effects of ghrelin and that post-HTx, ghrelin levels fall and caloric intake increases, suggesting resolution of ghrelin resistance. These states may contribute to cachexia in HF and weight gain post-HTx. These findings are preliminary and should be confirmed in larger trials.

Funding

This work was supported by Stockholms Läns Landsting, Astrid och David Hageléns Stiftelse, Magnus Bergvalls Stiftelse, and the Swedish Heart-Lung Foundation, all in Stockholm, Sweden, in the form of grants to L.H.L., and by the Division of Research Resources, General Clinical Research Centres Program, National Institutes of Health [5 MO1 RR00645], Bethesda, MD, the Foundation for Cardiac Therapies (FACT Fund), New, York, NY and the Altman Fund, New York, NY, in the form of grants to D.M.M.

Conflict of interest: none declared.