The Australian Multicenter Trial of Growth Hormone

0021-972X/98/$03.00/0
Journal of Clinical Endocrinology and Metabolism
Copyright © 1998 by The Endocrine Society
Vol. 83, No. 1
Printed in U.S.A.
The Australian Multicenter Trial of Growth Hormone
(GH) Treatment in GH-Deficient Adults*
ROSS C. CUNEO, STEVE JUDD, JENNIFER D. WALLACE, DON PERRY-KEENE,
HENRY BURGER†, SYLVIA LIM-TIO, BOYD STRAUSS, JAN STOCKIGT,
DUNCAN TOPLISS, FRANK ALFORD, L. HEW, HANS BODE, ANN CONWAY,
DAVID HANDELSMAN, STEWART DUNN, STEVE BOYAGES, N. WAH CHEUNG,
AND DAVID HURLEY
Endocrine Unit, Flinders Medical Centre (S.J.), Adelaide; Metabolic Research Unit, University of
Queensland, Princess Alexandra Hospital (R.C.C., J.D.W.) Brisbane; Endocrine Unit, Royal Brisbane
Hospital (D.P-K.), Brisbane; Prince Henry’s Institute of Medical Research, and Body Composition
Laboratory, Monash Medical Centre (H.B., S.L-T., B.S.), Melbourne; Ewen Downie Metabolic Unit,
Alfred Hospital (J.S., D.T.), Melbourne; Endocrine Unit, St. Vincent’s Hospital (F.A., L.H.), Melbourne;
Paediatric Endocrine Unit, Sydney Children’s Hospital (H.B.), Sydney; Andrology Unit, Royal Prince
Alfred Hospital (A.C., D.H.), Sydney; Medical Psychology Unit, University of Sydney (S.D.), Sydney;
Endocrine Unit, Westmead Hospital (S.B., N.W.C.), Sydney; and Department of Diabetes and
Endocrinology, Royal Perth Hospital (D.H.), Perth, Australia
impedance (P , 0.001) or deuterium dilution (P 5 0.002). Fat mass
measured by DEXA (P , 0.001), skinfold thicknesses (P , 0.001), and
waist/hip ratio (P 5 0.001) decreased in the first 6 months. Most
changes in body composition were complete by 3 months of treatment
and maintained to 12 months. Whole-body bone mineral density
(BMD) (by DEXA) was unaffected by GH treatment. Self-reported
quality of life was considered good before treatment, and beneficial
treatment effects were observed for energy, pain, and emotional reaction as assessed by the Nottingham Health Profile. In the initial 6
months, adverse effects were reported by 84% of patients in the GH
and 75% in the placebo group, with more symptoms relating to fluid
retention in the GH group (48% vs. 30%; P 5 0.016). Such symptoms
were mild and resolved in 70% of patients despite continued treatment. Resting blood pressure did not change over the initial 6 months.
In summary, GH treatment in adults with GH deficiency resulted in 1)
prominent increases in serum IGF-I at the doses employed, in some cases
to supraphysiological levels; 2) modest decreases in total- and low-density lipoprotein cholesterol, together with substantial reductions in totalbody and truncal fat mass consistent with an improved cardiovascular
risk profile; 3) substantial increases in lean tissue mass; and 4) modest
improvements in perceived quality of life. The excessive IGF-I response
and side-effect profile suggest that lower doses of GH may be required
for prolonged GH treatment in adults with severe GH deficiency. (J Clin
Endocrinol Metab 83: 107–116, 1998)
ABSTRACT
GH treatment in adults with GH deficiency has numerous beneficial effects, but most studies have been small. We report the results
of an Australian multicenter, randomized, double-blind, placebo-controlled trial of the effects of recombinant human GH treatment in
adults with GH deficiency. GH deficiency was defined as a peak serum
GH of ,5 mU/liter in response to insulin-induced hypoglycemia. Patients were randomly assigned to receive either GH (0.125 U/kg per
week for 1 month and 0.25 U/kg per week for 5 months) or placebo.
After 6 months, all patients received GH. The primary end points were
biochemical responses, body composition, quality of life, and safety.
One hundred sixty-six patients (72 females and 91 males) with a mean
age of 40 6 1 yr (6SEM; range 17– 67 yr) were recruited. Serum
insulin-like growth factor-I (IGF-I) increased from a standard deviation score of 22.64 6 0.27 (range 28.8 to 13.82; n 5 78) to 11.08 6
2.87 (range 27.21 to 16.42) at 6 months in the GH/GH group; 38% of
the whole group were above the age-specific reference range following
treatment [17.6% and 68.9% with subnormal (,2 SD) or normal (62
SD) pretreatment levels, respectively]. Fasting total cholesterol (P 5
0.042) and low-density lipoprotein cholesterol (P 5 0.006) decreased
over the first 6 months. Fat-free mass increased in the first 6 months
whether measured by bioelectrical impedance (P , 0.001) or dual
energy x-ray absorptiometry (DEXA; P , 0.001). Total-body water
increased in the first 6 months whether measured by bioelectrical
T
REATMENT of GH deficiency in adult humans became
an option following the development of recombinant
DNA-derived human GH and the early reports on the effect
of GH therapy in such patients (1, 2). The syndrome of GH
deficiency in adults principally comprises abnormalities in
body composition, cardiovascular risk factors, and psychological well-being. In comparison to normal individuals,
these patients have increased total-body fat mass (particularly visceral adiposity), reduced lean body (or fat-free) mass
(FFM), reduced skeletal muscle mass, reduced muscle
strength and exercise performance, and reduced bone mass
(3– 6). Increased mortality rates, particularly relating to cardiovascular diseases, have been described in adults with
hypopituitarism who received conventional hormone replacement with glucocorticoids, T4, and sex steroids (7).
Whether these mortality rates are explained by elevated total
and low-density lipoprotein (LDL) cholesterol and reduced
high-density lipoprotein (HDL) cholesterol concentrations,
central adiposity, or other factors remains unexplained (8).
Received May 6, 1997. Revision received September 3, 1997. Accepted
September 12, 1997.
Address all correspondence and requests for reprints to: Ross C.
Cuneo, University Department of Medicine, Princess Alexandra
Hospital, Wooloongabba, Brisbane, 4102, Australia. E-mail: cuneo2@
gpo.pa.uq.edu.au.
* This work was funded by Pharmacia (Australia) Pty Ltd. Portions
of this data were reported at the Endocrine Society of Australia’s annual
scientific meeting, Melbourne, September 1995.
† Principal investigator.
107
108
JCE & M • 1998
Vol 83 • No 1
CUNEO ET AL.
Reductions in muscle force generation and aerobic exercise
performance may relate to reductions in skeletal muscle mass
and altered myocardial function (9 –12). The psychological
dysfunction comprises self-reported reductions in energy,
mood, and sleep, along with objective reductions in marital
and socioeconomic performance (13, 14).
Treatment with GH has been reported to reverse many of
these abnormalities (1, 2, 15, 16). Uncertainty remains regarding the definition of GH deficiency and the appropriate
dose of GH for replacement. Differing studies have selected
patients of markedly different degrees of GH deficiency,
ranging from essentially no GH response to insulin-induced
hypoglycemia to those with a peak GH of ,10 –20 mU/liter
following oral clonidine (2). Recent data suggest that a peak
GH of ,5 ng/mL in response to insulin-induced hypoglycemia is abnormal for all adults (17). Most early reports used
high daily doses of GH (0.4 – 0.49 U/kg per week), resulting
in side effects caused by fluid retention (18 –21).
We undertook a large trial of GH treatment in adults with
GH deficiency to assess the effects of a moderate dose of GH,
in a group of patients selected on the basis of our current best
definition of GH deficiency. The primary end points were
assessment of biochemical responses, body composition,
quality of life, and safety.
Patients and Methods
Study design
Patients were randomly allocated without stratification to treatment
either with GH (Genotropin, Pharmacia AB, Stockholm, Sweden) or
identically presented placebo for the first 6 months in a double-blind
fashion. Thereafter, all patients received GH in an open study for a
further 6 months. The two groups were labeled GH/GH or placebo/GH.
Randomization to either treatment involved a computer-generated listing, with equal numbers being entered per group at each of the 10
centers.
The dose of GH (or placebo) was 0.125 U/kg per week (maximum
2 IU/day) for the first month, and 0.25 U/kg per week (or 0.55 mg/kg
per week; maximum 4 IU/day) thereafter. Because the study remained
blinded throughout, the dose for all subjects was lowered from month
6 until month 7, reverting to the full dose thereafter. The dose was
reduced in response to the side effects of edema or arthralgia and
remained at or was reduced to 0.125 IU/kg per week and increased only
with resolution of the symptoms. Persistent side effects resulted in
transient cessation of therapy, which was resumed with resolution of
symptoms, with the dose remaining at 0.125 IU/kg per week thereafter.
The dose was self-administered using the KabiPen (Pharmacia AB) from
16 IU/mL vials as sc injections 7 nights/week. The injection site was
either the anterior thigh or abdomen, at the patient’s preference, but
remained constant throughout the study. Compliance was checked by
vial count and injection diary.
Patients were assessed at months 0, 6, and 12 for all end points, at
months 1 and 7 for safety checks, and at months 3 and 9 for selected
end-point assessment.
The study protocol was approved by the ethics committees of each
participating unit. Center 1 participated only in the 6-month, doubleblind phase of the study. All patients received written information and
gave written consent according to the Declaration of Helsinki (1964 and
later revisions) and the National Health and Medical Research Council
(Australia) Ethics in Medical Research (1983). The study was conducted
according to the Australian Guidelines of Good Clinical Research Practice (1991).
those ,19 yr; 2) GH deficiency was documented as peak GH of ,5
mU/liter following insulin-induced hypoglycemia (blood glucose ,2.2
mmol/liter or with symptoms); and 3) other pituitary hormone deficiencies were replaced, with stable replacement therapy for at least 6
months before entry. Daily glucocorticoid doses were not to exceed 30
mg hydrocortisone, 37.5 mg cortisone acetate, 7.5 mg prednisolone, or
0.5 mg dexamethasone. T4 replacement was to result in normal serumfree T4 (fT4) and/or free T3 (fT3) concentrations. Parenteral or oral testosterone administration was to be used in men and estrogen (with/or
without progesterone) administration in women under age 50 yr.
Women of child-bearing potential were to be using adequate contraception (oral contraceptives or intrauterine device) and were required to
have a negative pregnancy test at entry.
Preentry insulin-hypoglycemia testing was required within the last 5
yr after the age of 20 yr or after epiphysial closure unless: 1) the patient
had had hypophysectomy with subsequent testing at any time; 2) the
patient had radiotherapy. Testing was also specified to be 3) at least 6
months after normalization of serum PRL in a patient with prolactinoma;
4) at least 6 months after normalization of glucocorticoid status in a
patient with Cushing’s disease; and 5) at least 12 months after cure of
a TSH-secreting pituitary tumor.
Patients were excluded if any of the following were present: 1) GH
treatment in the last 12 months; 2) a history of acromegaly; 3) estrogen
deficiency uncorrected in women below age 50 yr; 4) active Cushing’s
syndrome; 5) any acute severe illness in the last 6 months; 6) pregnancy
or lactation; 7) severe chronic liver disease (defined as persistent elevation of transaminases more than twice the upper limit of normal); 8)
chronic renal impairment (defined as serum creatinine . 0.12 mmol/
liter or persistent hematuria or proteinuria); 9) diabetes mellitus; 10)
history of malignancy other than cranial tumor or leukemia causing GH
deficiency; 11) history of meningioma or currently active intracranial
malignancy; 12) uncontrolled hypertension (diastolic blood pressure
.90 mmHg); 13) overt cardiac dysfunction or ischemia on electrocardiogram; 14) severe chronic airway disease; 15) any medication thought
to alter the response to treatment or the end points, especially high-dose
glucocorticoids and diuretics (cholesterol-lowering agents were permitted, but doses of all medications were held as constant as possible); 16)
history of noncompliance with medications, uncooperativity, or drug or
alcohol abuse; and 17) known sensitivity to m-cresol.
Screening of patients before entry included complete history and
physical examination, resting electrocardiogram, routine biochemistry
and hematology, and urinalysis.
Biochemical responses
Fasting serum concentrations of insulin-like growth factor I (IGF-I)
and IGF-binding protein-3 (IGFBP-3) were measured at months 0, 6, and
12 in all centers. Samples were frozen and stored at 220 C and assayed
in a central laboratory. IGF-I was measured by RIA after acid-ethanol
extraction, using a truncated radiolabeled IGF as tracer (in- house assay,
Pharmacia, Uppsala, Sweden). The detection limit was 20 ng/mL, and
the intra- and interassay coefficients of variation (CVs) at 202 ng/mL
were 3.1 and 10.0%, respectively. The concentration of IGF-I was also
expressed as a standard deviation score (i.e. normal values range
from 12 to 22), expressed in relation to normal age-adjusted adult
values calculated as follows:
IGF-I sds 5 (ln (IGF-I) 2 (5.95 2 0.0197 3 age))/0.282
Selection criteria
The reference population was Swedish, comprised of 83 male and 73
female healthy blood donors, with mean body weights of 80.9 kg (range
60 –115 kg) and 66.1 kg (range 50 –95 kg) for males and females, respectively, and mean body mass indexes of 25 and 23.7 kg/m2, respectively
(Pharmacia, unpublished data). IGFBP-3 was measured by RIA using
125
I-labeled recombinant human IGFBP-3 as tracer (in-house assay, Pharmacia). The assay’s detection limit was 0.9 ng/mL and the intra- and
interassay CVs at 43.2 ng/mL were 4.9% and 7.2%, respectively. The
assay had negligible cross-reactivity with IGFBP-1, -2, and -4 (,0.3%).
The concentration of IGFBP-3 was also expressed as a standard deviation
score from the same Swedish population data as IGF-I, calculated as
follows:
Patients were included if: 1) age was 18 – 65 yr and linear growth was
completed as confirmed by epiphysial closure on wrist radiography for
IGFBP-3 sds 5 (s IGFBP-3 2 exp (2.0466 2 (0.0099 3 age)))/
exp (0.6707 2 (0.0099 3 age))
GH TREATMENT IN ADULTS WITH GHD
Fasting serum was collected for lipid estimation at months 0, 6, and
12 in all centers. Routine assays were performed at individual centers.
Total cholesterol was measured by enzymatic and colorimetric methods
(Boehringer Mannheim, Mannheim, Germany) in all centers except one,
with interassay CVs of ,4.1%. HDL cholesterol was assayed following
polyethylene glycol precipitation of other lipoproteins (phosphotungstic
acid/Mg11 in one center), followed by identical analysis for cholesterol;
CVs were ,6.4%. LDL cholesterol was calculated from the Friedwald
formula. Serum triglyceride was measured by enzymatic and colorimetric methods (Boehringer Mannheim) in all centers except one, with
interassay CVs , 4.8%.
Body composition
All centers measured anthropometric characteristics and bioelectrical
impedance (BIA) at months 0, 3, 6, 9, and 12. Five centers (centers 2– 6)
measured deuterium dilution (D2O) at the same intervals, plus the
specialized body composition measurements at 0, 6, and 12 months [dual
energy x-ray absorptiometry (DEXA) and total-body potassium scanning]. At centers 4 – 6, in vitro neutron activation analysis was also
performed.
Anthropometric measurements included height, weight, skinfold
thicknesses, midarm circumference, and waist/hip circumference ratio
(WHR). A Harpenden caliper was used at four standard sites (biceps,
triceps, subscapular, suprailiac). Training was given to researchers at
each center before the commencement of the study to minimize betweencenter variability, and where possible, one observer performed all measurements. Skinfold measurements are reported as peripheral (sum of
biceps and triceps), truncal (sum of subscapular and suprailiac) and total
(sum of four). Waist circumference was recorded at the narrowest point
or at the umbilicus and hip circumference at the level of the greater
trochanter.
BIA was measured with the SEAC Bioimpedance Meter Model BIM
3.0 (Inderlec, Brisbane, Australia), except in center 4, where a BEI-101
Body Composition Analyzer (RJL Systems, Detroit, MI) was used. FFM
and total-body water (TBW) were calculated according to formulae
derived from healthy individuals (22, 23).
TBW was derived from the D2O method, following oral administration of D2O and spectrophotometric analysis of venous plasma samples
(24). The minimum intraassay CV of the assay was ,5.0% for a low
standard and ,2.2% for a high standard, and interassay CV was ,2.9%
over the duration of the study. DEXA measurements of the whole body
and lumbar spine (results of the latter to be reported separately) were
taken on two Lunar DPX System scanners (Lunar Corp., Madison, WI),
one in both Brisbane and Melbourne. FFM, fat mass (FM), and total-body
calcium were derived from the whole-body scans. Comparability of the
two machines was assessed by scanning phantoms. Total-body potassium (TBK) was measured as 40K content using shadow shield counters
in Brisbane and Melbourne. Subject measurements were calibrated
against a phantom (25). Using the proportion of intracellular potassium
as an index of FFM (64 – 68 mmol/kg) and the two compartment model
of body composition, FFM and FM were derived. For the Melbourne
scanner, the interassay CVs for measurements of the calibration and
anthropometric phantoms were ,3.8%. Comparability of the two machines was assessed by swapping the phantoms of known potassium
content. In vitro neutron activation analysis measured total-body nitrogen, and hence total-body protein (TBP) (24). The interassay CVs for
day-day measurement of the calibration and anthropometric phantoms
were ,3.8% (average 2.5%; within day) and ,10.5% (average 5.7%;
within month) throughout the study period. The mean accuracy was
99.5% for the anthropometric phantom.
Quality of life assessments
Questionnaires were administered at months 0, 3, 6, 9, and 12.
The Nottingham Health Profile (NHP) is a valid and reliable measure
of perceived health, developed in the United Kingdom to assess physical, emotional, and social distress. It is regarded as valid for Australian
populations (26, 27). Part I of the test was administered. Six subscales can
be derived, including emotions, pain, physical mobility, sleep, energy,
and social isolation. Analysis was based on the proportion of positive
answers, with a score higher than 0 indicating a compromised state. A
109
summary score (SUM) was also derived, with a positive score indicating
a favorable quality of life.
The GH Deficiency Questionnaire (GHDQ) was developed for this
trial. It was based on group discussions with adults with GH deficiency
regarding their perceived major areas of disability, i.e. energy, mood,
and sleep. The GHDQ consists of 30 questions relating to these three
subscales or domains, each answered on a 10-cm visual analog scale,
completed on an IBM portable computer, on which mean and range of
responses was recorded. Mean responses are reported. A higher score
represented a more positive status. The GHDQ has been validated in
normal individuals and used as a trial in a small group of adults with
GH deficiency (S. Dunn, unpublished data).
Social history was also assessed at months 0, 6, and 12. Data was
collected regarding marital status, number of children, living arrangements, education level, occupation/employment/retirement status, sick
leave in the past 6 months, physical activity at leisure and work, social
activity and satisfaction, and significant life events.
Safety and adverse events monitoring
Safety assessments included complete physical examination, routine
biochemistry, fT4, glycosylated hemoglobin A1c, full blood examination,
supine blood pressure recorded in triplicate after 5 min rest, and urinalysis at months 0, 6, and 12. All adverse events were recorded, quantitated according to the patient’s estimation of severity, and specific
inquiry was made as to features of fluid retention and arthralgias. The
adverse events were classified according to the WHO codes and preferred terms.
Statistical analysis
The study was analyzed on an intention-to-treat basis. Missing data
points were not estimated. Comparisons of continuous and categorical
baseline data were analyzed by ANOVA and contingency table analysis,
respectively. Treatment effects for NHP data were assessed using the
mean score test, a weighted analysis of covariance (ANCOVA) using
baseline data as the covariate (28); variance was partitioned between
treatment, withdrawal status, baseline data as covariate, month, and
relevant interactions. Treatment effects for other continuous variables
and the GHDQ were assessed using ANCOVA with baseline data as the
covariate. Results of treatment and baseline data were therefore presented as means adjusted for baseline differences (mean 6 sem); sems
for baseline data are presented uncorrected, whereas all posttreatment
sems are adjusted for baseline covariance. For body composition data,
variance was partitioned between treatment, center, gender, month, and
withdrawal status.
After ANCOVA, pairs of means were compared with the least significant difference, allowing comparison between groups at each time
point. Maintenance of a treatment effect beyond the first 6 months was
assigned if there was no significant change in the mean values in the
GH/GH group from 6 –12 months and no significant difference between
the GH/GH and placebo/GH group means at 12 months.
Treatment effects for categorical data were assessed using contingency table analysis and a log linear model (29). Laboratory safety data
were analyzed using Mann-Whitney U or Wilcoxon rank sum W tests.
A P value of ,0.05 was considered significant.
Results
Patient population, study drug, and compliance
One hundred sixty-six patients were entered. Three patients withdrew consent before initiation of treatment. The
demographic characteristics of the patients are shown in
Table 1. The two treatment groups were similar in all aspects
except for the duration of GH deficiency, which was shorter
in the GH/GH group (estimated from the time of abnormal
insulin-hypoglycemia testing or pituitary surgery and/or
radiotherapy; P 5 0.03). In addition to pituitary hormone
replacement, 10% of patients received bromocryptine; 9%
received antidepressants, anxiolytics, or hypnotic medica-
110
JCE & M • 1998
Vol 83 • No 1
CUNEO ET AL.
tions; 6% anticonvulsants; 7% treatment for asthma; and 4%
treatment for hypercholesterolemia. The most frequent
changes in medication during the study were courses of
antibiotics, increased bronchodilators or corticosteroids, and
short increments in glucocorticoid replacement. There were
no differences in medications between the two groups.
The average maintenance dose of GH was 2.6 6 0.8 IU/day
(0.03 6 0.01 U/kg per day or 0.22 6 0.05 IU/kg per week; n 5
74) at month 3 (2.6 6 0.6 IU/day for the placebo group; n 5
78) and 2.4 6 0.8 IU/day (0.22 6 0.06 IU/kg per week; n 5
117) at month 9. Twenty-four patients (29%) of the GH group
had their GH dose reduced or interrupted because of symptoms of fluid retention or arthralgia, and 13 of these patients
withdrew from the study. Two patients in the placebo group
had their placebo dose reduced. High rates of compliance
(.90%) were seen in 70% and 73% of patients who completed
6 and 12 months of treatment, respectively. Poor rates of
compliance (,60%) were seen in 6.9% and 16.2% of patients
who completed 6 and 12 months of treatment, respectively.
TABLE 1. Patient characteristics at baseline
Age (yr)
Gender (F/M)
Weight (kg)
Height (cm)
Body mass index (kg/m2)
Duration of GH deficiency (yr)
Pituitary hormone deficiency (%)
ACTH
TSH
LH/FSH
Antidiuretic hormone
Previous GH treatment (%)
Etiology of GH deficiency (%)
Pituitary tumor
Irradiation
Craniopharyngioma
Idiopathic
Cushing’s disease
Trauma
Septooptic dysplasia
Other
GH/GH group
Placebo/GH group
41.2 6 1.5
33:50
81.4 6 20.1
168.4 6 10.3
28.5 6 6.0
9.3 6 0.8
39.8 6 1.5
39:41
72.5 6 16.9
164.8 6 10.6
26.5 6 5.1
12.1 6 1.0a
67
71
75
28
28
76
78
83
26
35
48
13
6
8
6
5
1
12
39
19
15
13
3
1
0
11
Biochemical responses
Serum IGF-I concentrations. At baseline, there was no difference between treatment groups (Table 2). The percentage of
patients in the entire study group at baseline with IGF-I
sds ,2 and 62 were 56.6% and 43.9%, respectively. There
was a significant increase in serum IGF-I with GH treatment
for months 0 – 6 and 0 –12, and a treatment 3 time interaction
for months 0 –12 (all P , 0.001). The percentage of patients
who developed supraphysiological serum IGF-I concentrations at 12 months (.2 sd) who had subnormal (,2 sd) or
normal (62 sd) pretreatment levels was 17.6% (12/68) and
68.9% (31/45), respectively. Serum IGFBP-3 concentrations
showed similar changes (Table 2).
Serum lipids. At baseline, there were no differences between
treatment groups or centers for fasting serum concentrations
of total, LDL, and HDL cholesterol and triglycerides, except
for HDL cholesterol, where there was a significant difference
between treatment groups (GH 1.13 6 0.04 vs. placebo 1.28 6
0.06 mmol; P 5 0.050) (Table 3). There was a significant
decrease in total cholesterol with GH treatment for months
0 – 6 (P 5 0.040) but not for months 0 –12 (P 5 0.103), with a
significant difference in response between centers (P 5 0.03
and 0.014 for months 0 – 6 and 0 –12, respectively). There was
a significant decrease in LDL cholesterol with GH treatment
for months 0 – 6 and months 0 –12 (P 5 0.005 and 0.019,
respectively), with no difference between centers. For serum
HDL cholesterol concentration, there was no treatment effect
or treatment 3 time interaction, but there was a significant
difference between centers (P 5 0.001 and 0.008 for months
0 – 6 and 0 –12, respectively). For serum triglyceride concentration, there was no treatment effect, treatment 3 time interaction, or difference between centers.
Body composition
Results of body composition are shown in Table 4.
Body weight remained unchanged (GH/GH group: 81.4 6
20.1, 80.0 6 19.4, and 79.7 6 18.2 kg at 0, 6, and 12 months,
respectively; placebo/GH group: 72.5 6 16.9, 72.5 6 16.7, and
72.8 6 16.1 kg, respectively; mean 6 sd).
Anthropometry. At baseline, the total sum of skinfold thicknesses was significantly different between males and females
Data are presented as mean 6 SEM.
, P , 0.05 comparing group means.
a
TABLE 2. Serum IGF-I and IGFBP-3 response
Month
Serum IGF-I (ng/mL)
Serum IGF-I (SDS)
Serum IGFBP-3 (ng/mL)
Serum IGFBP-3 (SDS)
0
6
12
0
6
12
0
6
12
0
6
12
GH/GH group
Placebo/GH group
Mean (6SEM)
n
Mean (6SEM)
n
100
280 6 203,a,b
291 6 243,b
22.46
1.08 6 0.343,a,b
1.17 6 0.373,b
2.97
3.93 6 0.252,a,b
4.08 6 0.251,b
21.66
20.91 6 0.202,a,b
20.84 6 0.193,b
78
71
57
78
71
57
78
71
47
78
71
47
98
90 6 6
255 6 20b,d
22.64
22.94 6 0.26
0.71 6 0.34b,d
2.61
2.66 6 0.18
4.29 6 0.32b,d
21.95
21.94 6 0.14
20.67 6 0.24b,d
79
74
60
79
74
60
79
74
54
79
74
54
Primary analysis for treatment effects was performed with ANCOVA (1, P 5 ,0.05; 2, P , 0.01; 3, P , 0.001; see text for details); secondary
analyses performed with least significant difference analysis (a, difference between groups at each time point; b, c, d, and e, differences within
one group between time point and month 0, 3, 6, and 9, respectively; all P , 0.05).
GH TREATMENT IN ADULTS WITH GHD
111
TABLE 3. Lipid and safety data
Month
Total cholesterol (mmol/liter)
LDL cholesterol (mmol/liter)
HDL cholesterol (mmol/liter)
Triglycerides (mmol/liter)
Fasting serum glucose (mmol/liter)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
0
6
12
0
6
12
0
6
12
0
6
12
0
6
12
0
3
6
9
12
0
3
6
9
12
GH/GH group
Placebo/GH group
Mean (6SEM)
n
Mean (6SEM)
n
5.7
5.5 6 0.11,a,b
5.5 6 0.1b
3.8
3.3 6 0.12,a,b
3.3 6 0.11,b
1.2
1.3 6 0.1a,b
1.3 6 0.0b,d
1.8
1.9 6 0.1
1.8 6 0.1
4.5
4.7 6 0.13,a
4.7 6 0.11
120
118 6 1
118 6 1
115 6 2b
121 6 2e
78
76 6 1
75 6 1
74 6 1b
80 6 1c,d,e
70
67
51
41
36
29
52
48
41
70
67
51
67
64
47
72
72
69
53
53
72
72
69
53
53
5.8
5.7 6 0.1
5.7 6 0.1
3.9
3.8 6 0.1
3.6 6 0.1b
1.2
1.3 6 0.0
1.2 6 0.0
1.8
1.8 6 0.1
1.9 6 0.1
4.5
4.4 6 0.1
4.6 6 0.1
119
120 6 1
121 6 2
119 6 2
123 6 2b,e
76
78 6 1
78 6 1
76 6 1
77 6 1
78
72
60
45
36
35
57
49
47
78
72
60
74
66
52
79
77
77
65
61
79
77
77
65
61
Primary analysis for treatment effects was performed with ANCOVA (1, P 5 ,0.05; 2, P , 0.01; 3, P , 0.001; see text for details); secondary
analyses performed with least significant difference analysis (a, difference between groups at each time point; b, c, d, and e, differences within
one group between time point and month 0, 3, 6, and 9, respectively; all P , 0.05).
(76.2 6 2.7 vs. 93.7 6 3.9 mm; P , 0.001) and between centers
(P , 0.001) but not between treatment groups. There was a
reduction in total sum of skinfold thicknesses with GH treatment for months 0 – 6 and 0 –12 (P , 0.001 and P 5 0.003,
respectively). At baseline, the sum of truncal skinfold thicknesses was significantly different between males and females
(44.7 6 1.6 vs. 50.5 6 2.1 mm; P , 0.001) between centers (P ,
0.001) and between treatment groups (GH 49.3 6 1.9 vs.
placebo 45.1 6 1.8 mm; P 5 0.015). There was a reduction in
truncal skinfold thickness with GH treatment for months 0 – 6
and 0 –12 (P , 0.001 and P 5 0.001, respectively). At baseline,
the sum of peripheral skinfold thicknesses was significantly
different between males and females (31.1 6 1.4 vs. 44.6 6 2.2
mm; P , 0.001) and between centers (P , 0.001) but not
between treatment groups. There was a reduction in peripheral skinfold thickness with GH treatment for months 0 – 6
and 0 –12 (P 5 0.021 and P 5 0.063, respectively). None of
these treatment responses was influenced by gender. At
baseline, WHR was different between centers (P 5 0.007), but
not between treatment groups. There was a significant reduction in WHR with GH treatment for months 0 – 6 and
0 –12 (P 5 0.001 and P 5 0.015, respectively). Parallel reductions in total-body FM as measured by DEXA were also
observed for months 0 – 6 and 0 –12 (both P , 0.001; data not
shown).
FFM derived from BIA. At baseline, there were significant
differences between males and females (57.4 6 1.3 vs. 42.3 6
0.9 kg; P , 0.001) between centers (P 5 0.002) and between
treatment groups (GH/GH group: 53.4 6 1.5 vs. placebo/GH
group: 47.7 6 1.3 kg; P , 0.001). There was an increase in FFM
with GH treatment for months 0 – 6 and 0 –12 (both P , 0.001)
and a treatment 3 time interaction for months 0 –12 (P ,
0.001). This response was influenced by gender (for the 0 – 6and 0 –12-month periods, treatment 3 gender interaction P 5
0.029 and 0.079, respectively) indicating a greater increment
in FFM in males. FFM measured by DEXA, TBK, and TBP
showed similar increments following GH treatment at both
6 and 12 months (P , 0.001 for all; data not shown).
TBW derived from BIA. At baseline, there were significant
differences between males and females (39.1 6 0.8 vs. 27.6 6
0.6 l; P , 0.001), between centers (P 5 0.001), and between
treatment groups (GH/GH group: 35.9 6 1.0 vs. placebo/GH
group: 31.8 6 0.9 l; P , 0.001). There was an increase in TBW
with GH treatment for months 0 – 6 and 0 –12 (P , 0.001 and
P 5 0.005, respectively) and a treatment 3 time interaction
for months 0 –12 (P , 0.001). This response was not influenced by gender (for the 0 – 6- and 0 –12-month periods, treatment 3 gender interaction P 5 0.6 and 0.3, respectively).
TBW derived from D2O showed a similar increment with GH
treatment (P , 0.013; data not shown).
Whole-body BMD. At baseline, there were significant differences between males and females (1.193 6 0.013 vs. 1.097 6
0.017 g/cm2; P , 0.001), between centers (P 5 0.024), but not
between treatment groups (P 5 0.084). At baseline, the BMD
represented 97.4 6 1.3% and 97.4 6 1.6% of young adult
BMD, and 96.3 6 0.35% and 96.4 6 0.27% of age-matched
BMD for the GH/GH and placebo/GH groups, respectively.
There was no treatment effect (for the 0 – 6- and 0 –12-month
periods, P 5 0.30 and 0.19, respectively) or treatment 3 time
interaction. Secondary analysis (by least significant differ-
112
JCE & M • 1998
Vol 83 • No 1
CUNEO ET AL.
TABLE 4. Body composition
Month
Total skinfold thicknesses (mm)
Truncal skinfold thicknesses (mm)
Peripheral skinfold thicknesses (mm)
WHR
FFM (kg) derived from BIA
TBW (liters) derived from BIA
Total-body BMD (g/cm2)
Total-body BMD (% young adult; gender matched)
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
6
12
0
6
12
GH/GH group
Placebo/GH group
Mean (6SEM)
n
84.3 6 3.4
74.0 6 1.6a,b
72.2 6 2.03,a,b
70.7 6 2.7b
69.2 6 3.02,b,c
47.4 6 1.9
40.6 6 1.0a,b
40.6 6 1.43,a,b
39.5 6 1.5b
40.3 6 1.83,b
37.5 6 2.0
34.3 6 1.0b
32.5 6 1.01,a,b
31.8 6 1.5b,c
29.6 6 1.6b,c,d,e
0.894 6 0.009
0.878 6 0.004a,b
0.877 6 0.0043,a,b
0.876 6 0.005b
0.870 6 0.0051,b
50.6 6 1.5
53.5 6 0.4a,b
53.6 6 0.43,a,b
54.1 6 0.5b
53.3 6 0.43,b
34.0 6 1.0
35.8 6 0.2a,b
35.7 6 0.33,a,b
36.1 6 0.3b
35.6 6 0.32,b
1.149 6 0.017
1.141 6 0.003
1.120 6 0.007b,d
97.3 6 1.3
96.9 6 0.4
95.3 6 0.5b,d
72
72
69
54
53
72
72
69
54
53
73
73
70
55
54
73
73
70
54
54
73
72
70
54
53
73
72
70
54
53
45
43
39
34
32
30
Mean (6SEM)
83.5 6 3.3
82.2 6 1.6
81.5 6 1.9
76.1 6 2.0b,c,d
72.7 6 2.4b,c,d,e
46.9 6 1.8
46.6 6 0.9
46.8 6 1.2
43.0 6 1.1b,c,d
42.5 6 1.5b,c,d
36.9 6 1.8
36.3 6 1.0
35.4 6 1.0
33.7 6 1.2b,c
30.8 6 1.2b,c,d,e
0.892 6 0.008
0.895 6 0.003
0.893 6 0.004
0.882 6 0.007b,c,d
0.880 6 0.005b,c,d
50.2 6 1.3
50.3 6 0.3
50.9 6 0.4
54.2 6 0.7b,c,d
53.1 6 0.4b,c,d
33.7 6 0.9
33.8 6 0.2
34.2 6 0.3
36.3 6 0.4c,d
35.5 6 0.3c,d,e
1.148 6 0.016
1.145 6 0.002
1.129 6 0.003b,d
97.3 6 1.6
96.4 6 0.3
95.3 6 0.3b,d
n
79
78
77
63
60
79
78
77
63
60
79
78
77
64
60
79
78
77
65
61
77
75
74
61
58
77
75
74
61
58
43
42
36
35
35
30
Primary analysis for treatment effects was performed with ANCOVA (1, P 5 ,0.05; 2, P , 0.01; 3, P , 0.001; see text for details); secondary
analyses performed with least significant difference analysis (a, difference between groups at each time point; b, c, d, and e, differences within
one group between time point and month 0, 3, 6, and 9, respectively; all P , 0.05).
ence) revealed a significant reduction in BMD in both groups
by 12 months (see Table 4).
Efficacy: quality of life assessment
NHP. At baseline, analysis of the NHP revealed no difference
between the GH/GH and placebo/GH groups with respect
to energy, pain, sleep, and physical mobility (Table 5). The
mean baseline scores were low (indicating little or no impairment). The proportion of patients with 0 scores (no impairment) was: energy 36.5%, pain 77%, emotional reaction
45.5%, sleep 43.5%, social isolation 70.5%, and physical mobility 59.5%. Improvement in quality of life in patients who
score 0 at baseline cannot be measured with the NHP. Analysis of the data for efficacy by assessing only patients with
positive scores at baseline revealed no differences from that
presented (i.e. analyzing all data).
For the energy subscale, there was an initial (months 0 – 6)
improvement in the placebo/GH group vs. the GH/GH
group (P 5 0.016); however, the GH/GH group improved
progressively as shown by a treatment effect over months
0 –12 (P , 0.001) and a treatment 3 time interaction (P 5
0.046). For the pain subscale, there was no treatment effect,
but there was a treatment 3 time interaction for months 0 – 6
(P 5 0.047), suggesting a lessening in perceived pain in the
GH group. For the emotional reaction subscale, there was no
treatment effect for months 0 – 6, but for months 0 –12 there
was a beneficial effect (lessening of emotional reaction) seen
in the GH/GH group vs. the placebo/GH group (P , 0.001).
For the sleep subscale, there was no treatment effect for the
months 0 – 6 or 0 –12, but a significant treatment 3 time
interaction (P 5 0.011) indicated an improvement in the
GH/GH group over the 0 –12-month interval. For the social
isolation subscale, there was no treatment effect. For physical
mobility subscale, there was no treatment effect, but there
was a significant treatment 3 time interaction for the months
0 –12 (P 5 0.001), indicating an initial deterioration in mobility shortly after the introduction of GH treatment in both
groups followed by improvement. There was no treatment
effect for the SUM, but there was a treatment 3 time interaction from months 0 – 6 and 0 –12 (P 5 0.056 and 0.037,
respectively), indicating an overall improvement in the
GH/GH group.
GHDQ. There were no differences between the groups at
baseline, and there were no significant treatment effects for
GH TREATMENT IN ADULTS WITH GHD
113
TABLE 5. NHP results
Month
Energy score
Pain score
Emotional reaction score
Sleep score
Social isolation score
Physical mobility score
SUM score
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
0
3
6
9
12
3
6
9
12
GH/GH group
Placebo/GH group
Mean (6SEM)
n
Mean (6SEM)
n
1.03 6 0.06
0.69 6 0.06
0.55 6 0.051
0.15 6 0.03
0.23 6 0.043
0.77 6 0.10
0.84 6 0.10
0.34 6 0.041
0.28 6 0.05
0.45 6 0.09
1.38 6 0.12
0.87 6 0.09
0.65 6 0.07
0.32 6 0.05
0.31 6 0.063
1.14 6 0.07
0.93 6 0.07
0.85 6 0.07
0.58 6 0.07
0.70 6 0.081
0.48 6 0.06
0.26 6 0.03
0.27 6 0.04
0.13 6 0.03
0.18 6 0.04
0.54 6 0.07
0.84 6 0.07
0.61 6 0.06
0.39 6 0.05
0.58 6 0.073
4.43 6 0.06
4.67 6 0.11
5.24 6 0.09
5.09 6 0.121
73
73
70
52
54
73
73
70
51
54
72
72
70
52
54
73
73
70
52
54
73
72
70
52
54
73
73
70
52
54
70
69
50
53
1.17 6 0.05
0.50 6 0.04
0.56 6 0.04
0.60 6 0.05
0.36 6 0.04
0.23 6 0.06
0.19 6 0.04
0.28 6 0.05
0.44 6 0.06
0.49 6 0.09
0.70 6 0.07
0.43 6 0.05
0.58 6 0.05
0.68 6 0.06
0.55 6 0.06
0.99 6 0.06
0.52 6 0.05
0.55 6 0.05
1.14 6 0.06
0.78 6 0.07
0.24 6 0.03
0.19 6 0.03
0.31 6 0.04
0.10 6 0.02
0.29 6 0.04
0.38 6 0.03
0.26 6 0.04
0.37 6 0.05
0.73 6 0.05
0.54 6 0.06
5.02 6 0.05
4.43 6 0.12
4.78 6 0.12
4.88 6 0.13
76
79
76
64
61
79
76
75
65
60
78
76
76
65
61
79
78
76
65
61
79
78
76
65
61
78
78
76
65
61
69
71
62
57
Primary analysis for treatment effects was performed with mean score test, a weighted ANCOVA, with significant results for either a
treatment effect or a treatment times time interaction being highlighted (1, P 5 ,0.05; 2, P , 0.01; 3, P , 0.001; see text for details).
either mood, energy, or sleep scales throughout the study.
For the sleep scale, there was a significant treatment 3 time
interaction for months 0 –12 (P 5 0.011), indicating an improvement in sleep in the GH/GH group.
Social history. Physical activity at work and at leisure, each
quantitated into five categories, showed no change throughout the study. There was no change for the number of full sick
days taken in the previous 6 months throughout the study,
but the absolute numbers were small in each group (GH/GH
group: 0.45 6 0.11 days sick/patient over 6 months vs. placebo/GH group: 0.41 6 0.09 by month 6). The frequency of
meeting friends and satisfaction with life, quantitated into
five categories or yes/no, respectively, showed no change. At
baseline, 63% of the GH/GH and 73% of the placebo/GH
groups were satisfied with their lives, rising to 85% and 83%,
respectively, by 12 months. There was no change in the
occurrence of significant life events between the groups.
groups, and there was no change in with GH treatment for
months 0 – 6 and 0 –12 (P 5 0.17 and P 5 0.093, respectively).
At baseline, there was no difference in diastolic blood pressure between centers or treatment groups. There was no
change in diastolic blood pressure with GH treatment for
months 0 – 6 and 0 –12 (P 5 0.087 and P 5 0.23, respectively),
but there was a significant treatment 3 month interaction
(P 5 0.002) for the months 0 –12, indicating a progressive fall
in the GH/GH group from 0 –9 months (with an increase at
12 months returning to pretreatment levels).
Alkaline phosphatase and white cell count increased, and
serum urea, potassium, alanine aminotransferase, aspartate
aminotransferase, and g-glutamyltranspeptidase decreased
in the GH/GH compared with the placebo/GH group (data
not shown). The magnitude of such changes rendered these
of doubtful clinical significance. There was no change in red
cell count or glycosylated hemoglobin A1c.
Safety monitoring
Adverse events
At baseline, there was no difference in fasting serum glucose between treatment groups, but serum glucose increased
with GH treatment for months 0 – 6 and 0 –12 (P 5 0.001 and
P 5 0.014, respectively; Table 3). At baseline, there was a no
difference in systolic blood pressure between treatment
Overall, the incidence of reported adverse events was
high, partly relating to the method of documenting separate
but related events. In months 0 – 6, 290 events in 70 of the 83
GH/GH patients (84%) and 219 events in 60 of the 80 placebo/GH patients (75%) were recorded. During the open
114
phase of GH treatment, 411 new events were reported in 99
patients.
Events were considered either to relate to known GH
actions or to be unexpected. Predictable side effects (Table 6)
included edema (which included generalized, peripheral, or
facial edema, carpal tunnel symptoms, and peripheral swelling or tightness; 48% of GH/GH patients vs. 30% placebo/
GH; P 5 0.016), arthralgias, and myalgias (which included
arthritis, arthrosis, myalgia, muscle stiffness, tendonitis, and
muscle weakness; 30% GH/GH vs. 13% placebo/GH; P 5
0.007), paresthesia and anesthesia (12% GH/GH vs. 4% placebo/GH; P 5 0.056), and increased sweating (3.6% GH/GH
vs. 0% placebo/GH; P 5 0.078). Overall, 19 patients from the
GH/GH group and 11 from the placebo/GH group withdrew. The primary reason for withdrawal was a GH-related
adverse event in 40% of these patients.
Events that were not predictably GH related included: 1)
in months 0 – 6 reduced frequency of reported pain (0%
GH/GH vs. 6.3% placebo/GH; P 5 0.02), aggressive reactions (0% GH/GH vs. 3.8% placebo/GH; P 5 0.075), and
moniliasis (0% GH/GH vs. 3.8% placebo/GH; P 5 0.075); 2)
five cases of adrenal insufficiency (all on GH/GH); 3)
two cases of operation for pituitary tumors (one each on
placebo/GH and GH/GH); 4) two episodes of collapse in a
patient with similar past history (on GH/GH); 5) one episode
each of amaurosis fugax and chest pain in one patient (on
GH/GH); and 6) three abdominal surgical procedures. One
patient with pulmonary fibrosis and chronic graft vs. host
disease following childhood acute lymphatic leukemia died
of respiratory failure, which was not considered to be related
to the GH treatment.
Withdrawn patients
Analysis of completed vs. withdrawn patients revealed no
difference with respect to baseline characteristics. Those who
withdrew, compared with those completing the study, were
characterized by 1) a tendency to have lower FFM for the
0 – 6- and 0 –12-month periods (P 5 0.089 and 0.025, respectively); 2) having a progressively worse NHP energy score
TABLE 6. Predictable adverse events
Phase 1
GH
(n 5 83)
Edema
Incidence (%)
Severity (% mild;
moderate; severe)
Duration (days)
Unresolved (%)
Myalgia/Arthralgia
Incidence (%)
Severity (% mild;
moderate; severe)
Duration (days)
Unresolved (%)
Paraestiesie
Incidence (%)
Severity (% mild;
moderate; severe)
Duration (days)
Unresolved (%)
JCE & M • 1998
Vol 83 • No 1
CUNEO ET AL.
Phase 2
Placebo
(n 5 80)
Phase 2 GH
(n 5 130)
48
57; 36; 7
30
67; 30; 2
43
58; 39; 3
36 6 58
30
37 6 57
28
26 6 38
33
30
44; 50; 6
13
64; 36; 0
25
57; 36; 7
101 6 201
36
23 6 49
14
33 6 40
2
12
77; 15; 8
4
80; 20; 0
15
67; 9; 4
43 6 44
39
22 6 36
20
36 6 26
46
for months 0 –12 (P 5 0.003); 3) having a worse NHP pain
score for both the 0 – 6- and 0 –12-month periods (P 5 0.031
and 0.010, respectively); 4) those in the GH/GH group who
withdrew having a worse NHP emotional reaction score than
in the placebo/GH group for months 0 – 6 and 0 –12 (both P ,
0.001); 6) those in the GH/GH group who withdrew having
a worse NHP sleep score than in the placebo/GH group for
months 0 – 6 (P 5 0.006) and 0 –12 (P , 0.001); 7) having a
worse summary NHP score for months 0 – 6 and 0 –12 (P 5
0.037 and 0.005, respectively), and those in the GH/GH
group who withdrew having a worse score than those in the
placebo/GH group (P 5 0.004 and 0.002, respectively).
Maintenance of treatment effects from 6 –12 months
All biochemical end points remained stable (serum IGF-I,
IGFBP-3, total cholesterol, and LDL cholesterol). Significant
improvements in body composition end points were evident
at 3 months (all three skinfold estimates, WHR, FFM by BIA,
and TBW by BIA) and showed a gradual improvement (total
and peripheral skinfolds) or stable course until 12 months.
For FFM by DEXA, FM by DEXA, TBK, and TBN, improvements were evident at 6 months and were maintained to 12
months, with the placebo/GH group still below the GH/GH
group values by 12 months for the last two end points.
Discussion
The definition of GH deficiency used in this study was a
peak GH response to insulin-induced hypoglycemia of less
than 5 mU/liter. Recent studies have shown that insulininduced hypoglycemia is one of the most reliable means of
assessing GH secretion in adults, and that the threshold as
defined in this study represents an optimal distinction between normal adults and patients with severe GH deficiency
(17). The characteristics of the patients selected is representative of previous studies in GH deficiency in adults. The
most frequent cause of the condition was previous pituitary
adenomas and surgical and/or radiotherapy for such tumors
treated during adult life. Three quarters of the group had
panhypopituitarism. The presence of more than one pituitary hormone deficiency is strongly predictive of organic
GH deficiency (28, 30). Approximately one third of the group
received GH treatment for short stature during childhood.
The GH dose during this study was approximately half
that used in the original treatment trials (0.25 IU/kg per week
vs. 0.49 IU/kg per week) (18 –21). Nevertheless, the incidence
of GH-related adverse events was high. In comparing the
results of treatment between this and other studies, it seems
likely that GH-deficient adult patients with more profound
degrees of GH deficiency (i.e. this study group) may be likely
to have a greater response to GH treatment than those with
lesser degrees of GH deficiency (i.e. studies using responses
to clonidine or arginine) (31). Assuming IGF-I can act as a
guide for optimizing GH dose, we found that the a substantial proportion of patients developed a supraphysiological
serum IGF-I concentration following GH treatment, particularly in those with normal pretreatment serum IGF-I concentrations. This strongly suggests that the mean dose is still
too high, and that certain subgroups may require even lower
doses. Given the high rate of reported adverse events in
GH TREATMENT IN ADULTS WITH GHD
patients receiving placebo treatment, it was not possible to
predict those at risk of GH-mediated adverse events.
Adults with GH deficiency have increased risk factors for
premature vascular disease, i.e. increased total and LDL cholesterol concentrations, reduced HDL concentrations, increased apolipoprotein B concentrations, increased number
of small LDL particles, and increased WHRs, visceral adiposity, and varying degrees of insulin resistance (8, 32, 33).
Such abnormalities may explain the increased mortality rates
caused by vascular disease in adults with conventionally
treated hypopituitarism (7, 34). With respect to vascular risk
factors, GH treatment in our study induced small reductions
in total (1.8%) and LDL cholesterol (7.0%) concentrations,
and substantial reductions in truncal adiposity (9.2–14.1%;
see below). The increase in serum glucose with GH treatment
is well described, and appears to represent transient worsening of insulin resistance (35, 36). Previous studies have
shown prominent reductions in intraabdominal or visceral
fat following GH treatment (21), changes that would be expected to improve insulin resistance (37). There were no
changes in other risk factors. Thus, whether continued GH
treatment alters the long-term vascular mortality rate remains unknown.
We were able to confirm the findings that GH treatment
in adults with GH deficiency influenced body composition.
In the 6-month controlled phase, FFM increased 4.5% as
measured by BIA. In a subset of patients, other indices of lean
tissue mass also increased: 4.4% by DEXA, 6.6% by TBK, and
8.6% by TBP. TBW measured by BIA in all patients increased
by 3.2%. FM reflected in total skinfold thickness, truncal, or
peripheral skinfold thicknesses declined 12.0%, 14.1%, and
9.3%, respectively, and when measured by DEXA declined
9.2%. WHR decreased by 2.0%. The different magnitudes of
increase in lean tissue mass between the various methodologies may be partially explained by inaccuracies induced by
concomitant shifts in other body compartments (38), differing specificities of the tools, and that some measurements
were performed on subsets of patients. Our data confirm the
prominent lipolytic action of GH in adults with GH deficiency, and that more subcutaneous fat was lost from the
trunk than the periphery. Overall, the magnitude of such
changes was slightly less than reported in the early trials,
reflecting the lower dose of GH used in the current trial (18,
19, 21). Additionally, we have shown that the GH-induced
changes in body composition were sustained to 12 months.
Males showed greater increments in FFM measured by BIA
than females; the physiological explanation may relate to
interactions between testosterone and GH in stimulating
growth and protein anabolism (39), or in stimulating greater
IGF-I responses (40). Mean whole-body BMD declined 2.5%
in the GH/GH group and 1.7% in the placebo/GH group
over 12 months, with no GH treatment effect. Such findings
are consistent with an early GH-mediated increase in bone
remodeling (41).
Adults with GH deficiency perceive their own quality of
life to be worse than that of normal subjects (5, 42), and have
poor social performances in employment and marriage (43,
44). The NHP questionnaire used in this study has been
validated for a variety of diseased populations and in different cultures. Although we did not assess a control pop-
115
ulation under similar circumstances, between one third and
two thirds of our patients before treatment felt little or no
impairment in their quality of life. Hypothetically, the longterm presence of GH deficiency may have habituated them
to accept their condition. Despite this favorable pretreatment
quality of life, GH treatment still resulted in measurable
beneficial effects for energy, emotional reaction, the overall
score, and possibly pain and sleep. Detrimental effects were
recorded for physical mobility in the 3 months following
initiation of GH treatment, most likely relating to the high
prevalence of adverse effects of peripheral edema or arthralgias. The perceived beneficial effects of GH treatment, the
high compliance rate, and the percent of patients completing
the study is encouraging given the high incidence of adverse
events.
The nature of the adverse events correspond with previous
reports, and can be summarized as those caused by sodium
retention (resulting in edema or carpal tunnel symptoms) or
arthralgias/myalgias (which may relate to articular cartilage
swelling and skeletal muscle growth or edema). Although a
variety of disorders were reported under the heading of
edema, carpal tunnel syndrome was the least common, with
an incidence of 4.8% in the GH/GH group, 1.3% in the
placebo/GH group, and 3.4% during the open phase of treatment. Such effects are clearly dose dependent and in many
individuals self-limiting. One adverse event not previously
reported was glucocorticoid deficiency. Such events were
infrequent (five events in approximately 180 patient years of
treatment), making causality difficult to ascribe to GH. Increased glucocorticoid catabolism has been reported following GH treatment, with reduced 24-h serum cortisol profile
of hypopituitary patients on stable glucocorticoid replacement therapy (45), and induction of cytochrome P-450 enzymes (29).
Adults with GH deficiency treated with GH at a dose of
0.25 IU/kg per week (0.55 mg/kg per week) for 6 –12 months
benefit by means of 1) a doubling in serum IGF-I concentration; 2) substantial benefits in body composition such as
increased FFM and reduced FM; 3) modest reductions in
cardiovascular risk factors with as yet unproven biological
consequences; and 4) improvements in self-perceived quality
of life. Whether these changes translate into improved cardiovascular morbidity and mortality rates and enhanced
physical and psychological function in the long term requires
careful ongoing study. Optimization of the dose, with attendant reductions in adverse events that relate predominantly
to sodium and fluid retention, remains an important objective for the future development of this therapy.
Acknowledgments
We gratefully acknowledge the assistance of Ms. Phillipa Smith and
Mr. George Papadopoulos of Pharmacia (Australia). Without the willingness of the patients and the dedication of the research nurses and
staff, the success of this study would not have been possible. We thank
Dr. Phillip McLeod (Department of Mathematics, Monash University,
Melbourne) and Pharmacia (Sweden; safety data) for performing the
statistical analysis.
References
1. Christiansen JS, Jørgensen JOL. 1991 Beneficial effects of GH replacement
therapy in adults. Acta Endocrinol (Copenh). 125:7–13.
116
CUNEO ET AL.
2. Cuneo RC, Salomon F, McGauley GA, So¨nksen PH. 1992 The growth hormone deficiency syndrome in adults. Clin Endocrinol (Oxf). 37:387–397.
3. Rose´n T, Bengtsson B. 1991 Body composition in patients treated for hypopituitarism. J Endocrinol Invest. 14[Suppl 1]:P50. (Abstract).
4. Cuneo RC, Salomon F, Wiles CM, So¨nksen PH. 1990 Skeletal muscle performance in adults with growth hormone deficiency. Horm Res. 33[Suppl
4]:55– 60.
5. McGauley GA, Cuneo RC, Salomon F, So¨nksen PH. 1990 Psychological
well-being before and after growth hormone treatment in adults with growth
hormone deficiency. Horm Res. 33[Suppl 4]:52–54.
6. Kaufman J-M, Taelman P, Vermeulen A, Vandeweghe M. 1992 Bone mineral
status in growth hormone-deficient males with isolated and multiple pituitary
deficiencies of childhood onset. J Clin Endocrinol Metab. 74:118 –123.
7. Rose´n T, Bengtsson B. 1990 Premature mortality due to cardiovascular disease
in hypopituitarism. Lancet. 336:285–288.
8. Cuneo RC, Salomon F, Watts GF, Hesp R, So¨nksen PH. 1993 Growth hormone
treatment improves serum lipids and lipoproteins in adults with growth hormone deficiency. Metabolism. 42:1519 –1523.
9. Cuneo RC, Salomon F, Wiles CM, Hesp R, So¨nksen PH. 1991 Growth hormone treatment in growth hormone-deficient adults. I. Effects on muscle mass
and strength. J Appl Physiol. 70:688 – 694.
10. Cuneo RC, Salomon F, Wiles CM, Hesp R, So¨nksen PH. 1991 Growth hormone treatment in growth hormone-deficient adults. II. Effects on exercise
performance. J Appl Physiol. 70:695–700.
11. Cuneo RC, Salomon F, Wiles CM, et al. 1992 Histology of skeletal muscle in
adults with GH deficiency: comparison with normal muscle and response to
GH treatment. Horm Res. 37:23–28.
12. Cuneo RC, Salomon F, Wilmshurst P, et al. 1991 Cardiovascular effects of
growth hormone treatment in growth-hormone-deficient adults: stimulation
of the renin-aldosterone system. Clin Sci. 81:587–592.
13. Bjo¨rk S, Jo¨nsson B, Westphal O, Levin J-E. 1989 Quality of life of adults with
growth hormone deficiency: a controlled study. Acta Pædiatr Scand Suppl.
356:55–59.
14. Takano K, Tanaka T, Saito T, Committee for the Study Group of Adult GH
Deficiency. 1994 Psychosocial adjustment in a large cohort of adults with
growth hormone deficiency treated with growth hormone in childhood: summary of a questionnaire survey. Acta Pædiatr Suppl. 399:16 –19.
15. Jørgensen JOL, Pedersen SA, Thuesen L, et al. 1991 Long-term growth hormone treatment in growth hormone deficient adults. Acta Endocrinol
(Copenh). 125:449 – 453.
16. Attanasio AF, Lamberts SW, Matranga AM, et al. 1997 Adult growth hormone
(GH)-deficient patients demonstrate heterogeneity between childhood onset
and adult onset before and during human GH treatment. Adult Growth Hormone Deficiency Study Group. J Clin Endocrinol Metab. 82:82– 88.
17. Hoffman DM, O’Sullivan AJ, Baxter RC, Ho KKY. 1994 Diagnosis of growthhormone deficiency in adults. Lancet. 343:1064 –1068.
18. Jørgensen JOL, Pedersen SA, Thuesen L, et al. 1989 Beneficial effects of
growth hormone treatment in GH-deficient adults. Lancet. 1:1221–1225.
19. Salomon F, Cuneo RC, Hesp R, So¨nksen PH. 1989 The effects of treatment
with recombinant human growth hormone on body composition and metabolism in adults with growth hormone deficiency. N Engl J Med. 321:1797–1803.
20. Whitehead HM, Boreham C, McIlrath EM, et al. 1992 Growth hormone
treatment in adults with growth hormone deficiency: results of a 13-month
placebo controlled cross-over study. Clin Endocrinol. 36:45–52.
21. Bengtsson B, Ede´n S, Lo¨nn L, et al. 1993 Treatment of adults with growth
hormone (GH) deficiency with recombinant human GH. J Clin Endocrinol
Metab. 76:309 –317.
22. Lukaski HC, Bolonchuk WW, Hall CB, Siders WA. 1986 Validation of tetrapolar bioelectrical impedance method to assess human body composition.
J Appl Physiol. 60:1327–1332.
23. Lukaski HC, Bolonchuk WW. 1988 Estimation of body fluid volumes using
tetrapolar bioelectrical impedance measurements. Aviat Space Environ Med.
1163–1169.
24. Forbes GB. 1962 Methods for determining composition of the human body.
Pediatrics. 29:477– 494.
JCE & M • 1998
Vol 83 • No 1
25. Boddy K, King PC, Hume R, Weyers E. 1972 The relationship of total body
potassium to height, weight, and age in normal adults. J Clin Pathol.
25:512–517.
26. Mitchell RA, Imperial E, Zhuo D, et al. 1992 A cross-cultural assessment of
perceived health problems in the elderly. Disabil Rehabil. 14:133–135.
27. Baum FE, Cooke RD. 1989 Community health needs assessment: use of the
Nottingham Health Profile in an Australian study. Med J Aust. 150:587–590.
28. Toogood AA, Beardwell CG, Shalet SM. 1994 The severity of growth hormone
deficiency in adults with pituitary disease is related to the degree of hypopituitarism. Clin Endocrinol (Oxf). 41:511–516. [Published errata appear in Clin
Endocrinol (Oxf) 1994 41:845 and Clin Endocrinol (Oxf) 1995 42:109] [see
comments].
29. Cheung NW, Liddle C, Coverdale S, Lou JC, Boyages SC. 1996 Growth
hormone treatment increases cytochrome P450-mediated antipyrine clearance
in man. J Clin Endocrinol Metab. 81:1999 –2001.
30. Weissberger AJ, Verikiou K, So¨nksen PH. 1994 Diagnosis of growth hormone
deficiency in adults. Lancet. 344:483.
31. Hindmarsh PC, Smith PJ, Pringle PJ, Brook CG. 1988 The relationship between the response to growth hormone therapy and pre-treatment growth
hormone secretory status. Clin Endocrinol (Oxf). 28:559 –563.
32. Russell-Jones DL, Watts GF, Weissberger A, et al. 1994 The effect of growth
hormone replacement on serum lipids, lipoproteins, apolipoproteins and cholesterol precursors in adults growth hormone deficient patients. Clin Endocrinol (Oxf). 41:345–350.
33. O’Neal D, Hew FL, Sikaris K, Ward G, Alford F, Best JD. 1996 Low density
lipoprotein particle size in hypopituitary adults receiving conventional hormone replacement therapy. J Clin Endocrinol Metab. 81:2448 –2454.
34. Wu¨ster C, Slenczka E, Ziegler R. 1991 Increased prevalence of osteoporosis
and arteriosclerosis in patients with conventionally substituted pituitary insufficiency: is there a need for additional growth hormone substitution. Klin
Wochenschr. 69:769 –773.
35. Fowelin J, Attvall S, Lager I, Bengtsson B. 1993 Effects of treatment with
recombinant human growth hormone on insulin sensitivity and glucose metabolism in adults with growth hormone deficiency. Metabolism.
42:1443–1447.
36. O’Neal DN, Kalfas A, Dunning PL, et al. 1994 The effect of 3 months of
recombinant human growth hormone (GH) therapy on insulin and glucosemediated glucose disposal and insulin secretion in GH-deficient adults: a
minimal model analysis. J Clin Endocrinol Metab. 79:975–983.
37. Despres JP, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C. 1990
Regional distribution of body fat, plasma lipoproteins, and cardiovascular
disease. Arteriosclerosis. 10:497–511.
38. De Boer H, Blok G-J, van der Veen EA. 1995 Clinical aspects of growth
hormone deficiency in adults. Endocr Rev. 16:63– 86.
39. Haeusler G, Frisch H, Schmitt K, et al. 1995 Treatment of patients with
Ullrich-Turner syndrome with conventional doses of growth hormone and the
combination with testosterone or oxandrolone: effect on growth, IGF-I and
IGFBP-3 concentrations. Eur J Pediatr. 154:437– 444.
40. Holmes SJ, Shalet SM. 1995 Which adults develop side-effects of growth
hormone replacement? Clin Endocrinol (Oxf). 43:143–149.
41. Vandeweghe M, Taelman P, Kaufman J. 1993 Short and long-term effects of
growth hormone treatment on bone turnover and mineral content in adults
with growth hormone deficiency. Clin Endocrinol (Oxf). 39:409 – 415.
42. Rose´n T, Wire´n L, Wilhelmsen L, Wiklund I, Bengtsson B-Å. 1994 Decreased
psychological well-being in adult patients with growth hormone deficiency.
Clin Endocrinol (Oxf). 40:111–116.
43. Dean HJ, McTaggart TL, Fish DG, Friesen HG. 1985 The educational, vocational, and marital status of growth hormone-deficient adults treated with
growth hormone during childhood. Am J Dis Child. 139:1105–1110.
44. Rikken B, van Busschbach J, le Cessie S, et al. 1995 Impaired social status of
growth hormone deficient adults as compared to controls with short or normal
stature. Clin Endocrinol (Oxf). 43:205–211.
45. Weaver JU, Thaventhiran L, Noonan K, et al. 1994 The effect of growth
hormone replacement on cortisol metabolism and glucocorticoid sensitivity in
hypopituitary adults. Clin Endocrinol (Oxf). 41:639 – 648.