Conditional Cardiovascular Response to Growth Hormone Therapy in Adult Patients with Prader-Willi Syndrome

Abstract

Context: In Prader-Willi syndrome (PWS), an altered GH secretion has been related to reduced cardiac mass and systolic function when compared with controls.

Objectives: The objective of the study was to evaluate the cardiovascular response to GH therapy in adult PWS patients.

Study Participants: Thirteen obese PWS adults (seven males and six females, aged 26.9 ± 1.2 yr, body mass index 46.3 ± 1.6 kg/m²) participated in the study.

Methods: Determination of IGF-I, metabolic parameters, echocardiography, and cardioscintigraphy with dobutamine stimulation was made during 12 months GH therapy, with results analyzed by repeated-measures ANOVA.

Results: GH therapy increased IGF-I (P < 0.0001); decreased C-reactive protein levels (P < 0.05); and improved lean mass (P < 0.001), fat mass (P < 0.05), and visceral fat (P < 0.001). Echocardiography showed that 6- and 12-month GH therapy increased left ventricle mass in 76 and in 61% of patients, respectively (P < 0.05), did not change diastolic function, and slightly decreased the left ventricle ejection fraction (LVEF) (P = 0.054). Cardioscintigraphy documented stable values of LVEF throughout the study, whereas right ventricle ejection fraction decreased significantly (P < 0.05) being normally responsive to dobutamine infusion. A positive association between IGF-I z-scores and LVEF occurred at the 6- and 12-month follow-up (P < 0.05).

Conclusions: In PWS, GH therapy increased cardiac mass devoid of diastolic consequences. The observation of a slight deterioration of right heart function as well as the association between IGF-I and left ventricular function during GH therapy suggest the need for appropriate cardiac and hormonal monitoring in the therapeutic strategy for Prader-Willi syndrome.

OBESITY IS A DISTINGUISHING hallmark for many patients with Prader-Willi syndrome (PWS), which is also comprised of hypotonia, hypothalamopituitary dysfunction, as well as mental and behavioral abnormalities (1, 2). A disproportionate accumulation of body fat develops as early as in childhood (3) and leads progressively to severe obesity by the adult age (4) due to uncontrolled hyperphagia. In combination with muscle hypotonia and respiratory fragility, obesity is acknowledged as a primary cause of morbidity in PWS (1, 2, 5–10).

Constitutive alterations of the GH-IGF axis affect most PWS patients independent of obesity and 40–100% of patients are diagnosed with GH deficiency (GHD) according to different GH-stimulatory tests (11). In PWS, the clinical manifestations suggestive of GHD include low IGF-I levels, impaired longitudinal growth, and reduced lean mass and bone mass when compared with controls (11–14). The cardiovascular relevance of GHD is suggested by studies in hypopituitary populations showing that GHD is associated with increased fat mass, glucolipid abnormalities, and cardiovascular disorders, with most of these alterations being reversed by GH replacement therapy (15–17). In a previous study in adult PWS patients, we have documented cardiovascular features indicative of GHD, which consisted of decreased cardiac mass and lower ejective and chronotropic response to dobutamine when compared with healthy obese controls (18). However, the potential cardiovascular effects of GH administration to patients with PWS are currently unknown, even though GH therapy has been previously shown to benefit body composition, lipid profile, sleep breathing disorders, and pulmonary function in PWS adults (19–21).

To investigate the cardiovascular response to GH therapy in PWS, our study evaluated the effects of a 12-month GH treatment in obese PWS adults with a specific interest in morphological and functional parameters.

Patients and Methods

The patients’ group consisted of 13 consecutive PWS adults [six females and seven males, aged 26.9 ± 1.2 yr, body mass index (BMI) 46.3 ± 1.6 kg/m²] with typical phenotype and suffering from childhood obesity, each admitted to our Institution for evaluation and treatment of obesity by a multidisciplinary team. Due to the rarity of PWS, the selection criteria for the study excluded patients only if they were affected with cardiovascular or kidney illness and if any suffered from uncontrolled arterial hypertension or diabetes mellitus. The baseline evaluation has been described previously (18) and will be briefly summarized. Cytogenetic analysis revealed interstitial deletion of the proximal long arm of chromosome 15 (15q11-q13) in all but two subjects with uniparental disomy. No patient suffered from cardiovascular disorders except for two hypertensive patients treated with angiotensin-converting enzyme inhibitor plus Ca-antagonist (n = 1) and loop diuretics (n = 1) and one patient treated with a low-dose β-blocker due to past episode of ventricular arrhythmia. At the time of the study, two of three women with primary amenorrhea were undergoing sex steroid substitutive therapy, two untreated women suffered from oligomenorrhea, and another suffered from secondary amenorrhea. One hypertensive patient had type 2 diabetes and was treated with insulin. Five patients had previously undergone GH treatment, withdrawn in all cases 1–4 yr before enrollment in the current study. Five patients were receiving treatments with neuroleptics. No patient had previously undergone bariatric surgery or was taking weight-reducing drugs. All subjects were enrolled in the study after approval by the Ethic Committee of the Istituto Auxologico Italiano, and individuals signed consent under parental guidance.

Physical examination included determination of height and weight in fasting conditions and after voiding. BMI was defined as weight in kilograms divided by the square of height in centimeters. Waist was measured as midway between lower ribs and iliac crests in relaxed exhalation, hip circumference was measured as the maximum value over the buttocks, and their ratio was calculated. Dual-energy x-ray absorptiometry (DXA) was used for measurements of fat body mass (percent) (GE-Lunar, Madison, WI). Intraabdominal [visceral (VAT)] and sc abdominal fat (SAT) were measured by 6-mm single-slice L₄-level computed tomography using GE Hi-Speed DX/I with 6.4 computed tomography scanner software as previously described (18). Pituitary GH secretion was evaluated by dynamic testing with GHRH (1 μg/kg) + arginine (ARG; 0.5 g/kg) (18).

After baseline evaluation, the experimental protocol encompassed a 12-month treatment with recombinant human GH (Genotropin; Pfizer, Rome, Italy), with patients admitted to our institution at 0, 6, and 12 months to complete the evaluations. Patients received GH therapy at a mean starting dose of 0.3 ± 0.008 mg/d (0.021 ± 0.001 μg/kg·wk) for the first month. Subsequently the GH dose was adjusted to reach the 50th percentile of normal serum IGF-I for sex and age. At the end of each study period, the mean daily GH dose was 0.96 ± 0.05 mg (0.064 ± 0.004 μg/kg·wk) at 6 months and 0.96 ± 0.04 mg (0.065 ± 0.004 μg/kg·wk) at 12 months. Individual dietary prescriptions consisted of 75% of total daily resting energy expenditure (kilocalories per 24 h), estimated by computed open-circuit indirect calorimetry (Sensormedics 29, Anaheim, CA). At each visit, measurements of gas exchange were made by a ventilated canopy in a thermo-regulated room (22–24 C) at 1-min intervals for 30 min, expressed as 24-h values according to the standard abbreviated Weir’s equation (22):

Resting energy expenditure (kilocalories per day) = [3.941 VO₂ (milliliters per min) + 1.106 VCO₂ (milliliters per min)]·1.44

Hormone assays

All measurements were performed using commercially available kits. GH levels were measured by chemiluminescence (Immulite 2000 analyzer; Diagnostic Products Corp., Los Angeles, CA) calibrated against World Health Organization first international reference preparation 80/50, having a sensitivity of 0.01 μg/liter and intra- and interassay coefficients of variation (CVs) of 2.9–4.2 and 4.2–6.5%, respectively. Total IGF-I levels were assayed by chemiluminescence IGF-I immunoassay by Liaison (Nichols Advantage, San Juan Capistrano, CA), having a sensitivity of 6 μg/liter, intraassay and interassay CVs of 4.8 and 6.7%, respectively. Normal IGF-I ranges for age were established in the laboratory, mean control values being 303 ± 80 μg/liter for ages of 15–20 yr and 284 ± 80 μg/liter for ages of 21–35 yr, and individual IGF-I values were calculated as z-scores. Serum insulin levels were measured by chemiluminescence (Immulite 2000). Enzymatic methods (Roche Molecular Biochemicals, Mannheim, Germany) were used for determination of blood glucose; total, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) cholesterol; and triglycerides. The total to HDL cholesterol ratio was also calculated as an indicator of cardiovascular risk as previously described (18). Insulin resistance and insulin sensitivity were measured by the homeostatic model approach as HOMA-IR [insulin (microunits per milliliter) × blood glucose (millimoles per liter)/22.5] and HOMA-S% [22.5/insulin (microunits per milliliter) × blood glucose (millimoles per liter)] (23). Ultrasensitive C-reactive protein (CRP) was measured by CRP (latex) HS Roche kit, having sensitivity of 0.003 mg/dl, intraassay CVs of 5.35% at 0.05 mg/dl, 2.51% at 0.17 mg/dl, and 4.25% at 0.193 mg/dl; interassay CVs of 5.79% at 0.0481 mg/dl and 4.25% at 0.193 mg/dl. For conversion from metric to SI units: insulin, microunits per milliliter × 7.175 = picomoles per liter; glucose, milligrams per deciliter × 0.05551 = millimoles per liter; cholesterol, milligrams per deciliter × 0.02586 = millimoles per liter.

Cardiovascular examinations

M-mode, two-dimensional, and pulsed Doppler echocardiographic studies were performed as previously described (18) with commercially available ultrasound systems (Sonos 2500; Hewlett-Packard, Andover, MA) using a 2.5-MHz transducer, during three to five consecutive cardiac cycles. The following measurements were recorded on M-mode tracing: interventricular septum thickness (IVST; millimeters), left ventricular (LV) posterior wall thickness (LVPWT; millimeters), and LV end-diastole diameter (LVEDD; millimeters); LV mass (LVM; grams) was calculated using the Devereux’s formula according to the Penn convention with the following regression-corrected cube formula (24): LVM = 1.04 [(IVST + LVEDD + LVPWT)³ − (LVEDD)³] − 14 g as well as after correction for body surface area (LVMi; grams per square meter), height^2.7 (LVM/h^2.7), or percent fat mass (LVM per fat mass, grams). Doppler studies provided indices of ventricular filling that were derived from the mitral flow velocities curves, i.e. maximal early diastolic flow velocity (E; centimeters per second), maximal late diastolic flow velocity (A; centimeters per second), peak E/A wave velocity ratio (normal value ≥ 1), and the deceleration time of early filling (milliseconds). Estimated pulmonary artery systolic pressure (PASP; mm Hg) was derived from the amount of tricuspid regurgitation using the modified Bernoulli equation, in addition to the estimated right atrial pressure (normal ≤ 25 mm Hg).

Equilibrium radionuclide ventriculography was performed as previously described (18), at rest and during infusion of the inotropic β₁-adrenergic agent dobutamine. Acquisitions were obtained with patients in supine position in the left anterior oblique best septal view with a large field-of-view camera (Apex SP6; Elscint, Haifa, Israel) equipped with a parallel-hole high sensitivity collimator. Data were collected in minilist mode to compensate for heart variability during acquisition; 32 frames were acquired in a 64 × 64 array, excluding extrasystolic and postextrasystolic beats. Dobutamine infusion was performed in 5-min steps under electrocardiogram and blood pressure monitoring, image acquisition being obtained during the last 3 min of each step. Indices of LV and right ventricle (RV) function were derived by analysis of the background-corrected time-activity curve, which was constructed by a semiautomated edge-detection method with a variable region of interest. LV and RV ejection fraction (percent) were computed on the basis of relative end-diastolic and end-systolic count and peak filling rate (PFR) was computed from the first derivative of a third-order polynomial function fitted to the first two thirds of the diastolic portion of the LV time-activity curve by a least squares technique, normalized for end-diastolic volume (EDV) and expressed as EDV per second. As normal, the following values were taken in consideration: LV ejection fraction 50% or greater in basal conditions with 5% or greater increments during dobutamine; RV ejection fraction 45% or greater in basal condition with 5% or greater increments during dobutamine; PFR 2.5 or greater EDV per second.

Data analysis

Results are presented as mean ± sem. Two-tailed paired Student’s t test and repeated-measures ANOVA followed by Newman-Keuls multiple comparison test or test for trend were used for comparisons among the different follow-ups. Relationships between variables were analyzed using Pearson’s correlation analysis. Significance was set at P < 0.05. For comparative purposes, baseline results obtained from previously published data (18) in age-, sex- and BMI-matched controls have been included in the tables.

Results

At study entry, peak GH response after GHRH + ARG test was 6.8 ± 0.9 μg/liter and IGF-I levels were 2 or more sd below the age-related mean in 11 patients (85%) (Table 1). IGF-I levels increased significantly (F = 38.3, P < 0.0001) by 2.4- and 2.9-fold at the 6- and 12-month evaluations, respectively, and all but one patient achieved age-related normal IGF-I levels by the end of study period (Table 2). GH therapy led to a significant improvement of body composition, consisting of decreased visceral fat (F = 11.4, P < 0.001), reduced total fat mass (F = 5, P < 0.05) and increased lean body mass (F = 10.3, P < 0.001) (Table 2). These results were paralleled by a significant reduction of CRP levels (F = 4.1, P < 0.05). Blood glucose levels remained within the normal limits during GH therapy despite an increase in fasting glucose (F = 7.5, P < 0.01), insulin (F = 7.3, P < 0.01), and HOMA-IR (F = 7.5, P < 0.01) and decrease in HOMA-S% (F = 7.3, P < 0.01).

TABLE 1.

Individual bioanthropometric and hormonal results obtained in PWS patients at study entry

Patients (no.)	Age (yr)	Weight (kg)	BMI (kg/m2)	GH peak (μg/liter)	IGF-I (μg/liter)	IGF-I z-score	Waist to hip ratio	Fat mass (%)a	Lean mass (kg)a
1	19.6	124.2	48.5	6.3	203.5	−1.3	1.04	57.2	52.1
2	21.1	94	43.5	3.2	64.7	−2.7	0.92	63.8	33.7
3	22	109.1	43.1	5	112	−2.2	1.02	53.5	49.9
4	22.3	115.4	42.1	8.8	120.8	−2.1	0.96	53.8	52.3
5	25.8	110.6	46.8	2.5	71.8	−2.7	0.95	59	43.7
6	26	80.9	37.1	7.3	142.3	−1.8	0.80	56.8	34.9
7	28.3	116.6	48.8	7.5	125.5	−2	0.91	56.1	51
8	28.7	107.7	41.5	3	72.6	−2.6	0.98	52.3	52.9
9	29	95	41.7	3.8	25.2	−3.2	0.83	56.9	39.7
10	29.1	102.5	41.4	12.9	118.4	−2.1	1.03	45	55.2
11	31.1	127.8	55.4	3.4	63	2.8	0.86	59.4	51.7
12	32	88.5	42.5	12.8	73.4	−2.6	0.91	56.5	39.1
13	33.3	107.5	54.7	9.1	40.3	−3	0.86	61.9	39.3

^a

As determined by total-body DXA.

TABLE 2.

Bioanthropometric and biochemical results (mean ± sem) obtained in PWS patients at baseline and after 6 and 12 months of GH treatment

Parameters	Obese controls	PWS patients
		Baseline	GH treatment		P value (ANOVA)
			6 months	12 months
Body weight (kg)	125.4 ± 3.3a	108.5 ± 4	109 ± 4.4	107.7 ± 4.7	NS
BMI (kg/m2)	43 ± 1.2	46.3 ± 1.6	46.4 ± 1.7	45.8 ± 2	NS
Waist to hip ratio	0.91 ± 0.02	0.93 ± 0.02	0.94 ± 0.02	0.93 ± 0.02	NS
IGF-I (μg/liter)	155.3 ± 13b	94.5 ± 13.3	305 ± 39c	328 ± 32c	0.0001
IGF-I z-scores	−1.6 ± 0.2b	−2.4 ± 0.2	0.3 ± 0.5c	0.5 ± 0.4c	0.0001
Plasma glucose (mg/dl)	85.7 ± 3.1	84.2 ± 5.7	90.2 ± 5.6d	91.9 ± 5.6e	0.003
Insulin (mIU/ml)	15.6 ± 2f	6.2 ± 2.4	16.1 ± 2.2d	15.5 ± 3.1e	0.003
HOMA-IR	3.3 ± 0.5f	1.9 ± 0.3	3.6 ± 0.5e	3.6 ± 0.8d	0.003
HOMA-S%	0.39 ± 0.06f	0.78 ± 0.2	0.43 ± 0.13d	0.40 ± 0.06d	0.003
CRP (mg/dl)	1.6 ± 0.6	1.6 ± 0.6	0.9 ± 0.2	0.6 ± 0.2d	0.02
Triglycerides (mg/dl)	172.4 ± 27.8f	102 ± 12.8	110.3 ± 10.8	105.5 ± 10.8	NS
Total cholesterol (mg/dl)	215.5 ± 13.3f	181.1 ± 10.9	175.5 ± 9.3	187 ± 9.7	NS
LDL cholesterol (mg/dl)	148.2 ± 10.6f	118.8 ± 8.4	113.5 ± 8.1	123.5 ± 9.1	NS
HDL cholesterol (mg/dl)	45.6 ± 2.6	50.7 ± 3.7	49.5 ± 2.9	53.8 ± 2.8	NS
Total/HDL cholesterol	4.9 ± 0.3f	3.7 ± 0.3	3.7 ± 0.3	3.6 ± 0.3	NS
Fat mass (%)g	46 ± 1.6b	56.1 ± 1.3	54.3 ± 1.3d	53.9 ± 1.4d	0.01
Lean mass (kg)g	60.5 ± 3.7a	45.8 ± 2.1	48.6 ± 1.8c	48.4 ± 2e	0.0006
SAT (cm2)h	675.5 ± 33.6	707 ± 32.8	701 ± 28.2	671 ± 31.9d	NS
VAT (cm2)h	150.1 ± 24.8	147.3 ± 15.3	117.1 ± 14d	101 ± 14.3c	0.0003
VAT/SATh	0.24 ± 0.06	0.21 ± 0.02	0.16 ± 0.02e	0.15 ± 0.02e	0.0008
Abdominal APD (cm)h	30.1 ± 6	29.3 ± 0.6	30 ± 0.8	29.4 ± 1	NS

Previously published data (18 ) from age-, sex-, and BMI-matched controls have been included for comparative purposes at baseline. For conversion factors, see Patients and Methods; for F values of ANOVA, see Results. APD, Anteroposterior diameter; NS, not significant.

^a

P < 0.001, obese controls vs. PWS patients.

^b

P < 0.01, obese controls vs. PWS patients.

^c

P < 0.001, 6- and 12-month GH treatment vs. baseline in PWS patients.

^d

P < 0.05, 6- and 12-month GH treatment vs. baseline in PWS patients.

^e

P < 0.01, 6- and 12-month GH treatment vs. baseline in PWS patients.

^f

P < 0.05, obese controls vs. PWS patients.

^g

As determined by total-body DXA.

^h

As determined by abdominal computed tomography scans.

At echocardiography, no patient showed structural abnormalities throughout the study period, and derived parameters were comparable between patients who had or had been not previously treated with GH (data not shown). Compared with baseline, LV mass was significantly increased after 6 and 12 months of GH therapy by 14 and 12%, respectively (F = 3.8, P < 0.05) (Figs. 1 and 2² and Table 3). Individual values of LV mass were found to be increased in 10 patients by 1.1–39% after 6 months and 9 patients by 6–48% after 12 months of GH. This trend was confirmed when LV mass was indexed for body surface area (F = 3.9, P < 0.05), height^2.7 (F = 3.5, P < 0.05), or fat mass (F = 6.2, P < 0.01) (Table 3). An increase in interventricular septum thickness occurred specifically after 6 months (F = 4.1, P < 0.05; P < 0.05 vs. baseline); thereafter both interventricular septum thickness and LV posterior wall thickness showed a trend toward a decrease, whereas LV end-diastole diameter increased slightly (Table 3). In contrast, LV end-diastole volume did not change significantly. By Doppler analysis, no significant variation of the diastolic indices, i.e. early-to-late mitral flow velocity and deceleration time, was recorded. Inversely, LV ejection fraction values slightly decreased with GH therapy (F = 3.2, P = 0.054) (Figs. 1 and 2² and Table 3). On average, values of pulmonary artery systolic pressure were unchanged during GH treatment, being however higher than normal in eight (61%) patients at baseline and at 6 months and in nine patients (69%) at 12 months. Derived echocardiographic parameters did not differ between patients with chromosome 15 deletion and those with uniparental disomy (data not shown).

Fig. 1.

Individual cardiovascular results in PWS patients obtained before and after 6 and 12 months of GH treatment. A, Echocardiography-derived LVM (upper panel), LVM indexed by body surface area (LVMi; middle panel), and LV ejection fraction (LVEF; lower panel). B, Cardioscintigraphy-derived LVEF in basal conditions (upper panel), under dobutamine (middle panel), and as rest-to-peak increments (Δ; lower panel). C, Cardioscintigraphy-derived right ventricle ejection fraction (RVEF) in basal conditions (upper panel), under dobutamine (middle panel), and as rest-to-peak increments (Δ; lower panel). D, Cardioscintigraphy-derived LV PFR in basal conditions (upper panel), under dobutamine (middle panel), and as rest-to-peak increments (Δ, lower panel).

Fig. 2.

Cumulative percent variation (Δ) over baseline of echocardiogram-derived LVM and LV ejection fraction (LVEF) in PWS after 6 (white boxes) and 12 months of GH treatment (gray boxes).

TABLE 3.

Results (mean ± sem) obtained at echocardiography and radionuclide angiography in PWS patients at baseline and after 6 and 12 months of GH treatment

Parameters	Obese controls	PWS patients
		Baseline	GH treatment		P value (ANOVA)
			6 months	12 months
Echocardiography
LVPWT (mm)	10.1 ± 0.3a	9 ± 0.4	9.6 ± 0.3	9 ± 0.2	NS
IVST (mm)	9.4 ± 0.3	9.6 ± 0.4	10.4 ± 0.4	9.4 ± 0.4	0.02
LVM (g)	163.5 ± 8.4a	135.7 ± 7.7	155.2 ± 10.7b	148.6 ± 5.7c	0.03
LVMi (g/m2)	70.6 ± 3.1	67.1 ± 3.2	76.4 ± 4.6b	73.9 ± 2.1c	0.03
LVM/h (g/h2.7)	39.5 ± 1.9	42.7 ± 1.9	48.5 ± 2.5b	47.1 ± 1.9c	0.04
LVM/FM (g, %)	3.7 ± 0.3d	2.4 ± 0.2	2.9 ± 0.2b	2.8 ± 0.1b	0.007
LVEDD (mm)	47 ± 1.2	43.6 ± 1.3	44.4 ± 1.5	46.7 ± 1	0.04
LV ejection fraction (%)	60.6 ± 1	63.1 ± 1.4	59.4 ± 2c	60 ± 1.6	(0.054)
E/A	1.5 ± 0.1	1.6 ± 0.1	1.5 ± 01	1.6 ± 0.1	NS
PASP (mm Hg)	31.6 ± 1.5	31.6 ± 1.3	34.1 ± 1.3	31.4 ± 1.8	NS
DT (msec)	175.4 ± 4.5	181 ± 7.6	181 ± 6	174 ± 4.9	NS
Radionuclide angiography
LV ejection fraction (%)	58.6 ± 1.6	56.7 ± 1.7	56.3 ± 1.5	56.8 ± 2	NS
LV ejection fraction Dob (%)	76.3 ± 1.2a	71.9 ± 1.9	72.1 ± 1.4	69.5 ± 1.5	NS
ΔLV ejection fraction (%)	31.2 ± 3.4	27.2 ± 3.5	30.2 ± 2.2	25 ± 5.8	NS
RV ejection fraction (%)	42.4 ± 2.7	43.5 ± 1.4	38.6 ± 1.9b	38.7 ± 1.1d	0.013
RV ejection fraction Dob (%)	57.2 ± 3.3	61.1 ± 1.6	54.6 ± 2.8	57.7 ± 2.2	NS
ΔRV ejection fraction (%)	37.4 ± 8	42.6 ± 4.1	45 ± 7	49.8 ± 4.8	NS
PFR (EDV per second)	2.9 ± 0.2	2.6 ± 0.2	2.6 ± 0.1	2.7 ± 0.2	NS
PFR Dob (EDV per second)	4.5 ± 0.3	4.2 ± 0.3	4.2 ± 0.3	3.5 ± 0.2	(0.077)
ΔPFR (%)	56.2 ± 8	70.2 ± 22	64.6 ± 12.5	37.4 ± 17.4	NS
HR (bpm)	68.3 ± 3	63.2 ± 3.5	67.7 ± 3.4	66.5 ± 2.7	NS
HR Dob (bpm)	128 ± 2.8a	103 ± 6.9	105 ± 5.3	108.5 ± 5.6	NS
ΔHR (%)	92.9 ± 10a	57.3 ± 9.1	51.2 ± 3.1	61.5 ± 7.4	NS
DBP (mm Hg)	78.5 ± 2	75.5 ± 2.3	75.8 ± 1.9	77.5 ± 3.4	NS
DBP Dob (mm Hg)	83.1 ± 3.1	75.9 ± 2.2	80.4 ± 1.8c	80.9 ± 4	NS
SBP (mm Hg)	125.8 ± 3.9	117 ± 2.9	117.1 ± 3	123.3 ± 8.9	NS
SBP Dob (mm Hg)	145.4 ± 9	131.4 ± 7.2	142.9 ± 6.6	145 ± 8.9	NS

Previously published data (18 ) from age-, sex-, and BMI-matched controls have been included for comparative purposes at baseline. For F values of ANOVA, see Results. LVM/h, LVM indexed by height^2.7; LVM/FM, LVM indexed by percent fat mass; DT, deceleration time; HR, heart rate; DBP and SBP, diastolic and systolic blood pressures; Dob, dobutamine; NS, not significant.

^a

P < 0.05, obese controls vs. PWS patients.

^b

P < 0.01, 6- and 12-month GH treatment vs. baseline in PWS patients.

^c

P < 0.05, 6- and 12-month GH treatment vs. baseline in PWS patients.

^d

P < 0.01, obese controls vs. PWS patients.

Radionuclide angiography was performed in all but the one patient treated with β-blockers. Another patient was excluded from the 12-month analysis due to technical drawbacks. During each examination, the dobutamine dose was individually titrated up to a maximum of 40 γ/kg·min, with average doses being similar among the three evaluations. No patient developed clinically relevant alterations necessitating termination of the test. At baseline, derived parameters were comparable between previously GH-treated and GH-naïve patients (data not shown). Diastolic and systolic blood pressure measured at peak dobutamine infusion appeared to increase during GH therapy (Table 3). Normal values of LV ejection fraction were documented in all patients in basal conditions except for two patients before and after 6 and 12 months of treatment (Fig. 1). In all cases, LV ejection fraction values increased significantly (≥5%, compared with baseline) under dobutamine except for one patient before treatment and two others after 12 months of GH. LV ejection fraction and Δ-LV ejection fraction remained stable and were comparable among all time points.

On average, RV ejection fraction decreased during GH therapy, being abnormal (<50%) in 75% of patients at study entry and 100% of patients after 12 months. Interestingly, pulmonary artery systolic pressure was normal in 80% of patients with normal RV ejection fraction at baseline and abnormal in 71% of patients, with reduced RV ejection fraction at the 12-month evaluation. However, the dobutamine-stimulated RV ejection fraction and its rest-to-peak increments were preserved throughout the study period (Fig. 1). LV peak filling rate did not change significantly both in unstimulated and dobutamine-stimulated conditions, whereas rest-to-peak increments decreased during GH therapy without statistical significance, even when tested for linear trend, likely due to the wide distribution of individual changes (Fig. 1 and Table 3). With a gender-based analysis, LV mass differed significantly between males and females at study entry (149.3 ± 9.2 vs. 120.1 ± 9.2 g, P > 0.05) but not at the end of the study period (155.7 ± 6.5 vs. 140.4 ± 8.9 g, ns). Derived cardioscintigraphic parameters did not differ between patients with chromosome 15 deletion and those with uniparental disomy (data not shown).

No correlation was observed between the peak GH response or pretreatment IGF-I z-scores and the echocardiographic or radionuclide parameters during GH therapy. Likewise, GH peak and IGF-I z-scores did not differ between patients when stratified by echocardiography and radionuclide results (data not shown). After 6 and 12 months of GH therapy, a significant correlation occurred between IGF-I z-scores and LV ejection fraction measured by echocardiography (r = 0.74, P = 0.006, and r = 0.64, P = 0.018, respectively) and cardioscintigraphy (r = 0.62, P = 0.025, and r = 0.60, P = 0.028, respectively) (Fig. 3). In addition, there was a negative association between the percent IGF-I change at the 12-month time point and both LV mass (r = −0.57, P = 0.04) and LV mass indexed for body surface area (r = −0.59, P = 0.03).

Fig. 3.

Bivariate correlation analysis between individual IGF-I SDS (z-score) and LV ejection fraction (LVEF) values determined by echocardiography (upper panel) and cardioscintigraphy (lower panel) in PWS patients at baseline (asterisks and dotted line) and after 6 (closed circles and broken line) and 12 months (open circles and solid line) of GH treatment.

By stepwise multivariate regression analysis, IGF-I z-scores constituted an independent predictor of LV ejection fraction after 6 and 12 month of GH, either when determined by echocardiography (β = 0.74, P < 0.01, and β = 0.64, P < 0.05, respectively) or cardioscintigraphy (β = 0.62, P < 0.05, and β = 0.61, P < 0.05, respectively); IGF-I z-scores were also predictors of RV ejection fraction during GH therapy (β = 0.59, P < 0.05, and β =0.66, P < 0.05, respectively). Among the variables related to body composition, fat-free mass independently predicted the response of LV ejection fraction to dobutamine after 12 months of GH therapy (β = 0.62, P < 0.05).

Discussion

To date, this is the only prospective study investigating the cardiovascular effects of GH therapy in adult patients with PWS. The results of our investigation show that a 12-month GH treatment improved the LV mass in most patients without significant effects on cardiac function and lipid profile.

Adult PWS is typically characterized by obesity, behavioral abnormalities, and impaired GH secretion (11). The latter is known to increase the cardiovascular risk in both young and elderly patients with childhood- or adult-onset hypopituitarism (15–17). The clinical impact of GHD involves impaired growth rate of cardiac muscle, reduced cardiac performance on effort, dyslipidemia, and endothelial dysfunction (reviewed in Ref. 17). In adults with PWS, we previously documented structural and functional cardiac alterations suggestive of GHD (18). Our current results show that a 12-month GH therapy decreased CRP levels, a recognized marker of cardiovascular risk (25); reduced total body fat by 4.2% and abdominal visceral fat by 30.8%; and increased lean mass by 6.2 kg. Similar results have been previously obtained in GH-treated PWS children and adults (3, 19, 26–28), and a placebo-controlled study showed that 12-month GH therapy decreased fat mass by 2.5% and increased lean body weight by 2.2 kg (28). Differences between the latter (28) and our results may depend on the varying anthropometric and genetic characteristics of patients, dietary regimens, or GH doses. In our investigation, GH therapy also slightly impaired glucose homeostasis. Reduction of insulin sensitivity is recognized as a drawback of GH therapy (29) and has been previously observed in 30% of GH-treated PWS adults (28). The potential clinical consequences of these effects should be taken into consideration in the adult setting of PWS.

By serial echocardiographic evaluations, an increase of LV mass was documented in most patients during GH therapy. This result was mediated overall by an increase of the end-diastole diameter and septal thickness, likely reflecting the increased size of cardiomyocytes (30). As a whole, LV mass was unrelated to peak GH response to GHRH + ARG as well as pretreatment IGF-I values, whereas the inverse correlation seen between posttreatment IGF-I levels and LV mass likely reflected the degree of cardiac impairment in patients with lower IGF-I levels at study entry. A trend toward a reduction of echocardiogram-derived left ventricle ejection fraction, occurring in eight patients after 6 months and one additional patient after 12 months of GH therapy, was also observed. Nevertheless, LV ejection fraction remained within the normal ranges in all but one patient who did not reach normal IGF-I levels under GH due to dose-related side effects.

When a more accurate, operator-independent methodology was used, in the present investigation cardioscintigraphy, no negative effects of GH on left ventricle hemodynamics were documented. On the other hand, GH therapy was associated with a decrease of resting right ventricle ejection fraction, which was, as illustrated by the individual curves, particularly evident after 6 months of therapy and appeared to become stable thereafter. It is known from experimental studies that GH and IGF-I induce (re)expression of early genes associated with cardiac hypertrophy, i.e. myosin light and heavy chain, atrial natriuretic factor, c-fos, collagen α1 type III, fibronectin, and α2-tubulin (31–34). Cardiomyocyte hypertrophy and increased calcium responsiveness of myofilaments mediate the positive inotropic effects of GH, both in wild-type and GH-transgenic animal models (35, 36). Alternatively, there is no previous evidence that GH administration induced right-heart impairment in either animal studies or clinical studies in GHD patients. It is therefore unclear whether our observations imply a (transient) right heart maladaptation to the hypertrophic effects of GH in the obese setting of PWS, which might hypothetically be related to an increase in septal thickness through the mechanism of ventricular interaction or to overloading due to GH-mediated increase of intravascular volume (31, 37). Additional direct or indirect GH-related effects on cardiac contractility may include hyperinsulinemia, altered peripheral resistance, and interstitial collagen deposition (31). However, because PASP values were stable and rest-to-peak right heart responses were normal throughout the study period, the aforementioned mechanisms seem unlikely to be clinically relevant. The impact of current observations remains to be substantiated in larger study samples and longer observations; however, we consider cardiovascular assessment as a mandatory procedure in the work-up of PWS adults and their follow-up during GH therapy.

Previous investigations in childhood- and adult-onset non-PWS GHD patients showed that GH therapy improves cardiac mass, diastolic dysfunction, and systolic function under physical exercise (38–41). In agreement with a previous investigation (42), we did not observe changes in diastolic function during GH therapy in our PWS patients. Previous polysomnographic investigations suggested a potential deterioration of sleep-breathing disorders during GH treatment, particularly in younger PWS patients (21). This and our findings suggest the need for adequate monitoring of the cardiopulmonary function during GH therapy, although more extensive studies will be needed for an adequate assessment in the condition of critical illnesses.

In conclusion, our results suggest that GH therapy may improve some cardiovascular features of PWS, particularly cardiac mass, body composition, and some markers of cardiovascular risk. GH therapy did not affect left ventricle diastolic and systolic function, but individual signs of deterioration in right ventricle function should be taken into account and warrant an appropriate surveillance. Furthermore, the positive association between IGF-I and left ventricle ejection fraction during GH therapy support the view that appropriate hormonal targeting is needed in PWS.

Acknowledgments

Disclosure Statement: The authors have no potential conflict of interest to disclose.

Abbreviations:

• A,

  Late diastolic flow velocity;

• ARG,

  arginine;

• BMI,

  body mass index;

• CRP,

  C-reactive protein;

• CV,

  coefficient of variation;

• DXA,

  dual-energy x-ray absorptiometry;

• E,

  early diastolic flow velocity;

• EDV,

  end-diastolic volume;

• GHD,

  GH deficiency;

• HDL,

  high-density lipoprotein;

• HOMA-IR,

  homeostatic model of insulin resistance;

• HOMA-S%,

  homeostatic model of insulin sensitivity;

• IVST,

  interventricular septum thickness;

• LDL,

  low-density lipoprotein;

• LV,

  left ventricule;

• LVEDD,

  LV end-diastole diameter;

• LVM,

  LV mass;

• LVMi,

  LVM after correction for body surface area;

• LVPWT,

  LV posterior wall thickness;

• PASP,

  pulmonary artery systolic pressure;

• PFR,

  peak filling rate;

• PWS,

  Prader-Willi syndrome;

• RV,

  right ventricle;

• SAT,

  sc abdominal fat;

• VAT,

  visceral abdominal fat.