Author + information
- Received March 23, 2013
- Accepted April 1, 2013
- Published online August 1, 2013.
- Antonio Cittadini, MD∗∗ (, )
- Alberto M. Marra, MD∗,
- Michele Arcopinto, MD∗,
- Emanuele Bobbio, MD∗,
- Andrea Salzano, MD∗,
- Domenico Sirico, MD∗,
- Raffaele Napoli, MD∗,
- Annamaria Colao, MD†,
- Salvatore Longobardi, MD‡,
- Ragavendra R. Baliga, MD§,
- Eduardo Bossone, MD⋮ and
- Luigi Saccà, MD∗
- ∗Department of Internal Medicine, Cardiovascular and Immunological Sciences, University “Federico II,”, Naples, Italy
- †Department of Molecular and Clinical Endocrinology and Oncology, University “Federico II,”, Naples, Italy
- ‡Medical Liaison Office, Merck Serono S.p.A., Rome, Italy
- §Division of Cardiovascular Medicine, The Ohio State University, Columbus, Ohio
- ⋮Cardiology Division, Cava dei Tirreni-Amalfi Coast Hospital, Heart Department, University of Salerno, Salerno, Italy
- ↵∗Reprint requests and correspondence:
Dr. Antonio Cittadini, Department of Internal Medicine, Cardiovascular and Immunological Sciences, University of Naples Federico II, Via Pansini, 5, 80131 Naples, Italy.
Objectives This study sought to evaluate the efficacy and safety of long-term growth hormone (GH) replacement therapy in GH-deficient patients with chronic heart failure (CHF).
Background Recent evidence indicates that growth hormone deficiency (GHD) affects as many as 40% of patients with CHF, and short-term GH replacement causes functional benefit. Whether long-term GH replacement also affects CHF progression is unknown.
Methods The study is an extension of a previous randomized, controlled single-blind trial that screened 158 consecutive CHF patients (New York Heart Association classes II to IV) and identified 63 who had GHD by the growth hormone releasing hormone plus arginine test. Fifty-six patients were randomized to receive either GH therapy or standard CHF therapy. Patients were evaluated at baseline and after a 4-year follow-up. The primary endpoint was peak oxygen consumption (VO2). Secondary endpoints included left ventricular (LV) ejection fraction and volumes, serum amino terminal fragment of the pro-hormone brain-type natriuretic peptide, quality of life, and safety.
Results Seventeen patients in the GH group and 14 in the control group completed the study. In the GH group, peak VO2 improved over the 4-year follow-up. The treatment effect was 7.1 ± 0.7 ml/kg/min versus −1.8 ± 0.5 ml/kg/min in the GH and control groups, respectively. At 4 years, LV ejection fraction increased by 10 ± 3% in the GH group, whereas it decreased by 2 ± 5% in control patients. The treatment effect on LV end-systolic volume index was −22 ± 6 ml and 8 ± 3 ml/m2 in the GH and control groups, respectively (all p < 0.001). No major adverse events were reported in the patients who received GH.
Conclusions Although this is a preliminary study, the finding suggests a new therapeutic approach to a large proportion of GHD patients with CHF.
A recent model of chronic heart failure (CHF) predicts that deficiency of the main anabolic forces plays a central role, not only with regard to the functional abnormalities, but also in terms of disease progression and mortality (1). In this context, the growth hormone (GH)/insulin-like growth factor (IGF)-1 axis is of the utmost importance because: 1) it is the most effective anabolic system in nature; 2) between one-third and two-thirds of CHF patients are also affected by GH or IGF-1 deficiency (1,2); and 3) GH and IGF-1 play an essential physiological role in supporting cardiac growth and performance (3). On this ground, it is reasonable to ask whether correction of growth hormone deficiency (GHD) in CHF patients improves their functional capacity, and more importantly, whether it delays CHF progression.
We recently reported positive effects of 6-month GH replacement on physical and cardiac performance in CHF patients with coexisting GHD (4). Notwithstanding the encouraging results, several critical issues remain unsettled. Because GH is an anti-insulin hormone, its long-term administration may cause insulin resistance or clinical diabetes, both of which are strong predictors of CHF progression (5,6). Another critical issue is whether the beneficial effects documented after a brief course of GH replacement tend to vanish with time or persist, or even beget further benefit. Consequently, we designed this study to determine the effects of a very long-term replacement therapy in patients with CHF and concomitant GHD.
This is an extension of a previous controlled, single-blind single-center study (4), in which 158 consecutive CHF patients (New York Heart Association [NYHA] classes II to IV) were screened with regard to their GH/IGF-1 status by the growth hormone releasing hormone (GHRH) plus arginine stimulation test. As described elsewhere (4), according to national guidelines, GHD in adults was diagnosed if peak GH circulating levels were <9 μg/l after the GHRH plus arginine stimulation test. Of the 63 patients who satisfied the criteria for GHD, 56 of them were enrolled in the trial. Twenty-eight patients were randomly assigned to receive subcutaneous somatropin (rDNA origin) for injection GH (Saizen, Merck-Serono International, Geneva, Switzerland, or NutropinAq, Ipsen, Paris, France) at a dose of 0.012 mg/kg every second day, on top of standard therapy for CHF. GH was kept refrigerated and administered using sterile disposable syringes and needles. Small volume syringes were used so the prescribed dose could be drawn from the vial with reasonable accuracy. The other 28 patients only received standard therapy for CHF and served as a control group. The inclusion criteria were as follows: patients of either sex affected by CHF NYHA classes II to IV secondary to ischemic or idiopathic-dilated cardiomyopathy; ages 18 to 80 years; stable medications for at least 1 month, including beta-blockade that had to be started at least 5 months before entering the study; left ventricular (LV) ejection fraction 40% or less and LV end-diastolic dimension 60 mm or more; and informed consent. Exclusion criteria included active proliferative or severe nonproliferative diabetic retinopathy, active malignancy, any evidence of progression or recurrence of an underlying intracranial tumor, unstable angina or recent myocardial infarction, severe liver disease, and serum creatinine levels >2.5 mg/dl. Racial distribution was homogenous in our study population because all patients were Caucasian, which was strongly representative of the epidemiological features of our region. Because this 6-month course of GH administration exerted positive effects on exercise capacity and cardiac function, we hypothesized that very long-term GH treatment could have an effect on the progression of CHF. Therefore, we amended the initial protocol by setting a 4-year follow-up and peak oxygen consumption (VO2) as the study primary endpoint. Written informed consent was obtained from each patient, and the study protocol was approved by the Ethics Committee of the University Federico II (NCT01576861).
Standard 12-lead electrocardiography was performed in each patient before and after 6 months of GH treatment to assess heart rate, conduction, and repolarization abnormalities. Ambulatory electrocardiographic monitoring was performed at baseline and after 6 months. Quality of life was evaluated with the Minnesota-Living with Heart Failure Questionnaire (MLHFQ) (7).
All patients underwent an incremental symptom-limited cardiopulmonary exercise test (CPET) on a bicycle ergometer. After a 1-min warmup period at 0-W workload, a ramp protocol of 10 W/min was started and continued until limiting symptoms or other indications for exercise termination appeared (4,8). Respiratory gas exchange measurements were obtained breath-by-breath using a commercially available system (Vmax 29C, Sensormedics, Yorba Linda, California). VO2 was recorded as the mean value of VO2 during the last 20 s of the test. The ventilatory anaerobic threshold was detected using the V-slope method. The ventilation per min (VE) versus carbon dioxide production (VCO2) relationship was measured by plotting ventilation against VCO2 obtained every 10 s of exercise (VE/VCO2 slope). The VE/VCO2 slope was calculated as a linear regression function, excluding the nonlinear part of the relationship after the onset of acidotic drive to ventilation (4).
An ultrasound system equipped with a 2.5-MHz multifrequency transducer (Aplio SSA-770A, Toshiba, Japan) was used for complete M-mode, 2-dimensional, and Doppler echocardiographic analyses. M-mode and 2-dimensional recordings were made with the patients in the lateral recumbent position, according to previously published methods (9). Intraobserver and interobserver variabilities were 9.2% and 16.4% for LV mass, 2.4% and 3.1% for LV dimensions, and 3.9% and 4.6% for ejection fraction, respectively. Details of the variability coefficients of the measurements of diastolic function are reported elsewhere (9).
Serum GH was assayed with an immunoradiometric assay method (Pharmacia & Upjohn Diagnostic AB, Uppsala, Sweden) and serum IGF-1 with a radioimmunoassay using a monoclonal antibody after acid-ethanol extraction. Serum amino terminal fragment of the pro-hormone brain-type natriuretic peptide (NT-proBNP) concentrations were measured by an electrochemiluminescence immunoassay (Roche Diagnostics, Indianapolis, Indiana) using a Roche Elecsys analyzer. The sensitivity of the assay was 5 pg/ml. We used core laboratories blinded to treatment allocation for all measurements.
The results are expressed as means ± SEM. Treatment effect (Δ changes in GH group vs. Δ changes in control group) was evaluated by an unpaired t-test. The primary variable of the study was peak VO2. Previous data showed that peak VO2 increased by 1.8 ml/kg/min over the control group after 6-month GH replacement (4). In a recent study that demonstrated the beneficial effects of long-term physical training in CHF patients, the difference in peak VO2 between treated and untreated patients at 4 years was 4 to 4.5 ml/kg/min (10). Based on these observations, we reasoned that the effect of long-term GH replacement could be considered clinically relevant in terms of disease progression, if the 4-year Δ change in peak VO2 between GH and control patients was at least 4 ml/kg/min. By setting a significance level of 5%, a study power of 80%, and a SD of 4 U in the VO2 measurements, a total sample of 34 patients was required to reach the endpoint. Therefore, we were confident that the initial sample of 58 patients was adequate to test the hypothesis, even assuming a 40% dropout rate during follow-up because of adverse events or noncompliance. Considering the high dropout rate, the statistical plan for dealing with missing data was a sensitivity analysis that counted missing data as no changes in values.
Baseline characteristics of the 2 groups of patients that reached 4 years of follow-up are shown in Table 1. In the GH group, peak VO2 increased remarkably, and at 4 years, reached the value of 21 ± 1 ml/kg/min, which was 2-fold higher than the corresponding value of the control group (11.8 ± 0.2 ml/kg/min) (Table 2). Comparison of the Δ changes (baseline 4 years) in peak VO2 revealed that the treatment effect was 7.1 ± 0.7 ml/kg/min in the GH group (n = 17) and −1.8 ± 0.5 in the control group (n = 14) (p < 0.001). The VE/VCO2 slope decreased in the GH group from 32 ± 1 to 29 ± 2, indicating improved ventilatory efficiency. However, the treatment effect did not reach statistical significance. GH replacement induced LV reverse remodeling, as documented by the significant reductions of both LV end-diastolic and end-systolic volumes indexes (−23% and −37%, respectively) and circumferential wall stress (−46%). Accordingly, LV ejection fraction rose to 42 ± 2% at 4 years, whereas in the control group, it decreased slightly. The treatment effect was 10 ± 3% versus −2 ± 5% in GH and control patients, respectively (p < 0.001). The slight reduction of LV mass following GH replacement therapy, although in theory in contrast to the known GH-induced, growth-promoting effects, was accounted for by the remarkable reduction of LV cavity size during follow-up.
The MLHF questionnaire score decreased by 26% in the GH group, and this was paralleled by a reduction of the NYHA class. Holter electrocardiography, performed every 6 months, showed no significant differences between the 2 groups in terms of arrhythmic events. Lown class ranged from 1A to 5 in all CHF patients. Specifically, in the GH-treated group, there was a reduction from Lown class 4A to 1A in 2 patients, 4B to 2 in 4 patients, and in the other 11, the class remained unchanged. In the control group, 2 patients had a decrease in Lown class from 4B to 4A, 7 patients shifted from 2 to 4A, whereas in the other 5, Lown class remained unchanged. The diuretic dose was decreased in 8 patients of the GH group and increased in 3 patients of the control group. In the GH group, 5 patients with an ejection fraction of less than 35% at the study start, in whom implantable cardioverter-defibrillation was suggested, had an increase in ejection fraction to more than 35% during the treatment period; therefore, implantation was withheld.
Serum IGF-1 increased by 84% from baseline at 24 months and remained stable at 48 months (77%). NT-proBNP rose from 3,940 ± 1,050 pg/ml to 4,909 ± 432 pg/ml in the control group, whereas it decreased from 3,201 ± 900 pg/ml to 2,794 ± 432 pg/ml in the GH group (p < 0.001 vs. control group). A significant correlation was found between the IGF-1 levels at 4 years and the values of peak VO2 (r = 0.59; p < 0.05) and LV ejection fraction (r = 0.55; p < 0.05).
Of the 56 patients originally enrolled, 17 and 14 patients in the GH and control groups, respectively, completed the 4-year follow-up. Six patients in the GH group and 10 patients in the control group died because of cardiac events. Of the remaining 5 patients in the GH group, 2 patients withdrew from the study because of arthralgia, 2 patients for personal reasons, and 1 patient underwent cardiac transplantation. In the control group, 1 patient died because of melanoma, and 2 withdrew for personal reasons. Hospitalizations for worsening CHF were lower in the GH group than in the control group (11 patients vs. 20 patients). Although the study was not designed for hard clinical endpoints, it was noteworthy that there was a marked difference in the aggregate of death and hospitalization for worsening CHF (17 and 31 events in the GH and control groups, respectively).
In view of the considerable amount of missing data, we performed a sensitivity analysis that counted the lost to follow-up patients as no change in values. After this correction, the improvements of all endpoints studied remained statistically significant (Table 2). The finding was not unexpected because although the treatment effect was lower after the adjustment, this was paralleled by the concomitant increase of the number of patients included in the analysis. As a prototype, the final adjusted value of peak VO2 in the GH group was 17.3 ± 1 ml/kg/min (p = 0.008 from baseline) and the adjusted treatment effect on the same parameter was 4.3 ± 0.8 ml/kg/min, whereas in the control group these values were 12.2 ± 0.5 ml/kg/min and −0.9 ± 0.3 ml/kg/min, respectively. In the active treatment group, the adjusted value of ejection fraction after 4-year follow-up was 36 ± 1.8% (p < 0.01 from baseline) and the corrected treatment effect was 6.24 ± 1.76%, whereas in the control group, the values were 29.8 ± 1.7% and 1.35 ± 2.8%, respectively.
Safety and tolerability
Except for 2 patients who complained of arthralgia, no other side effects of GH treatment were reported. GH induced no significant changes in the main biochemical and hormonal parameters. In particular, thyroid hormones, testosterone, and glycosylated hemoglobin (HbA1c) were unaffected by GH administration. Specifically, thyroid-stimulating hormone ranged from 2.2 ± 1 mU/ml at baseline to 1.8 ± 0.8 mU/ml after 4 years (p = NS), free testosterone ranged from 435 ± 12 ng/ml at baseline to 425 ± 9 ng/ml after 4 years, and HbA1c ranged from 7.3 ± 0.9% at baseline to 7.5 ± 1.2% after 4 years.
The present study suggests that correction of GH deficit improves exercise and cardiac performance of CHF patients in an enduring way. Peak VO2, measured during CPET, was chosen as the study primary endpoint because it is a strong, independent predictor of CHF progression (11).
Against the small decrease observed in control patients, peak VO2 rose remarkably in the GH group, and the treatment effect was well above the predicted value of 4 ml/kg/min. The improvements in LV end-systolic volume and ejection fraction were equally relevant; both are consolidated predictors of survival. The value of peak VO2 above 20 ml/kg/min and that of ejection fraction above 40%, which was observed in the GH group at 4 years, are hardly compatible with clinical CHF syndrome. Accordingly, the average NYHA class in the GH patients dropped to 1.6, which is midway between asymptomatic and mild CHF. The data on hospitalization for CHF and death, although inconclusive in view of the small sample size, are well in agreement with the functional response observed in the GH group. Altogether, the present study suggests that the sustained functional benefit of long-term GH replacement must involve radical changes in the intimate mechanisms that mediate the progression of CHF.
In this regard, several experimental and human studies provide the pathophysiological underpinnings for GH attenuation of pathological remodeling. Activation of the GH/IGF-1 axis promotes a unique pattern of adaptive myocardial growth characterized by preserved capillary density, absence of fibrosis and changes of calcium regulatory proteins, and cardioprotection against ischemia-reperfusion and mechanical stretch (12–14). Furthermore, IGF-1 increases myocardial contractility mainly through calcium sensitization of the myofilaments, which represent a downstream mechanism not associated with arrhythmias (15). Reduction of apoptosis is another molecular hallmark of GH/IGF-1 activation and has been described by a number of in vivo and in vitro studies (16,17). Endothelial nitric oxide release and vasodilation are further markers of the activation of the GH/IGF-1 pathway and may account for the salutary vasodilatory actions of GH therapy (18). Importantly, GH also beneficially influences respiratory and skeletal muscle, which are both impaired in CHF (19–21). Although it is difficult to dissect the pleiotropic actions of enhanced GH/IGF-1 signaling, it is likely that the large improvements of cardiopulmonary indexes could be accounted for by the well-described GH actions on peripheral muscle in addition to the benefits on LV architecture and function. Taken together, GH and its mediator IGF-1 affect a broad array of relevant targets of pathological remodeling that may have induced the remarkable reverse remodeling observed in the present study.
In the present study, we did record very few adverse effects due to GH replacement therapy (i.e., 2 cases of arthralgia). This is a well-known side effect of GH replacement, also reported by other investigators (22,23). The relatively low prevalence of adverse effects is likely due to the replacement and not the pharmacological doses of GH employed in the present study, and is congruent with the largest database available, which indicates that GH replacement in adults is well tolerated (24).
A few additional comments are warranted. First, at variance with the progressive improvement of exercise capacity in the GH group, LV volumes and ejection fraction showed the highest response at 2 years and then tended to stabilize. The explanation may be that exercise capacity integrates the multifaceted activity of GH/IGF-1 that goes beyond the heart and encompasses a variety of targets, including lung, skeletal muscle, endothelium, and energetic metabolism (3,25,26). Second, previous studies recently reviewed by Arcopinto et al. (27) using pharmacological doses of GH in CHF patients have shown variable results. It must be stressed that our data are not directly comparable to those reported by these clinical trials because we used a relatively low replacement GH therapy in GHD patients with CHF, whereas previous data were obtained from CHF patients treated with pharmacological doses of GH, independent of their GH/IGF-1 status. Although congruent with most of the 12 published studies (26), our data are not congruent with the 2 largest double-blind, placebo-controlled trials (28,29). Potential explanations include differences in the study design, duration of GH administration, and the lack of preliminary assessment of the GH/IGF-1 status. In this regard, it has been postulated that differences in the IGF-1 increment in response to GH may explain the variable results. This was confirmed by a meta-analysis of 12 studies in which the investigators showed significant relationships between the IGF-1 response and GH treatment effects (30). The outstanding issue remains which CHF patients would display higher benefits after GH administration. Taken together, the results of the present study suggest that noncachectic CHF patients with low GH/IGF-1 activity may benefit better from GH therapy because they display higher IGF-1 increases, paralleling improved outcome measures. We observed a significant correlation between the IGF-1 levels at 4 years and the values of peak VO2 (r = 0.59; p < 0.05) and LV ejection fraction (r = 0.55; p < 0.05). These findings support the concept that patients with coexisting GHD and CHF represent a subset of patients who are highly responsive to replacement therapy, whereas GH therapy in unselected CHF patients yields unpredictable results.
Our population had a prevalence of GHD of almost 40%. Despite several studies that have addressed the prevalence of IGF-1 deficiency (2,31,32), little evidence is available regarding GHD prevalence in the CHF population. Our data are congruent with data reported by Broglio et al. (33), who showed a variable reduction of somatotroph responsiveness in 53% of patients with dilated cardiomyopathy.
Although the results of the present study are encouraging, it is important to underline that this was a small, single-center, single-blind study. In addition to the lack of a placebo arm, the large patient dropout represents another limitation. However, care was taken that the investigators who performed all measurements were fully blinded to the study protocol, and many parameters were objective and not easily biased. Specifically, all biochemical, imaging, and exercise tests were performed in facilities different from the coordinating center by investigators unaware of the treatment protocol. Moreover, to deal with the high dropout rate, we performed a sensitivity analysis that counted those lost to follow-up as no change in values, and the improvements of all endpoints studied remained statistically significant.
The results from this preliminary single-blind study suggest that more research is warranted to test the hypothesis that long-term GH replacement therapy could delay the progression of CHF combined with GHD in a large multicenter placebo-controlled double-blind trial.
This study has been supported in part by unrestricted grants from Merck Serono and Ipsen. Dr. Longobardi is a Merck Serono employee. All other authors reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- chronic heart failure
- cardiopulmonary exercise test
- growth hormone
- growth hormone deficiency
- growth hormone releasing hormone
- glycosylated hemoglobin
- insulin-like growth factor
- left ventricular
- Minnesota-Living with Heart Failure Questionnaire
- serum amino terminal fragment of the pro-hormone brain-type natriuretic peptide
- New York Heart Association
- carbon dioxide production
- ventilation per minute
- oxygen consumption
- Received March 23, 2013.
- Accepted April 1, 2013.
- 2013 American College of Cardiology Foundation
- Sacca L.
- Jankowska E.A.,
- Biel B.,
- Majda J.,
- et al.
- Doehner W.,
- Rauchhaus M.,
- Ponikowski P.,
- et al.
- Pasquali D.,
- Arcopinto M.,
- Renzullo A.,
- et al.
- Belardinelli R.,
- Georgiou D.,
- Cianci G.,
- Purcaro A.
- Cittadini A.,
- Stromer H.,
- Katz S.E.,
- et al.
- Cittadini A.,
- Ishiguro Y.,
- Stromer H.,
- et al.
- Cittadini A.,
- Grossman J.D.,
- Napoli R.,
- et al.
- Merola B.,
- Longobardi S.,
- Sofia M.,
- et al.
- Napoli R.,
- Guardasole V.,
- Matarazzo M.,
- et al.
- Arcopinto M.,
- Bobbio E.,
- Bossone E.,
- et al.
- Broglio F.,
- Benso A.,
- Gottero C.,
- et al.