| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Clinical Research Centers |
Department of Medicine, Columbia College of Physicians and Surgeons, New York, New York 10032, and the Department of Neurosurgery, Mount Sinai Medical Center, New York, New York 10029
Address all correspondence and requests for reprints to: Dr. Pamela U. Freda, Department of Medicine, Columbia College of Physicians and Surgeons, 630 West 168th Street, New York, New York 10032.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Sixty postoperative subjects with acromegaly and 25 age-matched healthy subjects were evaluated with nadir GH levels after 100 g oral glucose as well as baseline IGF-I and IGFBP-3 levels. GH was assayed by polyclonal RIA, sensitive immunoradiometric assay (IRMA), and highly sensitive enzyme-linked immunosorbent assay. The mean nadir GH determined by IRMA was 0.09 ± 0.004 µg/L in the healthy subjects, with the upper limit of the normal nadir being 0.14 µg/L (mean + 2 SD). Subjects with acromegaly were divided into those with active disease (n = 22), defined by elevated IGF-I levels, and those in remission (n = 38), defined by normal IGF-I levels. GH determined by IRMA failed to suppress into the normal range defined by our healthy subjects in all patients with active disease; nadir GH determined by IRMA ranged from 0.33–5.0 µg/L in this group. In 50% of the active group, nadir GH levels determined by IRMA were less than 1.0 µg/L, a GH nadir previously considered normal by strict criteria. When nadir GH levels in the subjects with active disease were measured by polyclonal RIA, there was overlap with the range of RIA values in the healthy subjects. Thus, the IRMA was superior to the RIA in that the overlap between these two groups was eliminated. Subjects with acromegaly in remission included those with normal GH suppression (n = 23; mean nadir GH by IRMA, 0.10 ± 0.006 µg/L) and others with abnormal GH suppression by IRMA (n = 15; mean nadir GH by IRMA, 0.35 ± 0.07 µg/L). The latter group may have persistent GH dysregulation detected by the sensitive IRMA. GH levels measured by enzyme-linked immunosorbent assay confirmed the IRMA results. IGFBP-3 levels were significantly higher in subjects with active acromegaly (4940 ± 301 µg/L) vs. those in healthy subjects (2887 ± 153 µg/L; P < 0.0001) and those in the subjects in remission (2966 µg/L; P < 0.0001). IGFBP-3 levels correlated overall with IGF-I levels (r = 0.765; P < 0.0001), but IGFBP-3 levels were not predictive of disease status because 32% of the subjects with active acromegaly had normal IGFBP-3 levels. In addition, failure of GH to suppress adequately was not associated with a higher IGFBP-3 level among the subjects in remission.
These data indicate that the IRMA is superior to the RIA in distinguishing between patients with active disease (defined by elevated IGF-I levels) and healthy subjects. We also show that GH levels after oral glucose measured with highly sensitive GH assays can be much lower in subjects with active disease than previously believed; values less than 1.0 µg/L may be found in up to 50% of patients. In addition, in 39% of patients in apparent remission with normal IGF-I levels, GH determined by highly sensitive assays fails to suppress normally; it remains to be determined whether these patients are at higher risk for recurrence of active disease.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We studied 60 patients, of a cohort of 150, who underwent transsphenoidal surgery (by K.D.P.) for acromegaly between 1981 and 1997. Subjects were selected by their availability and willingness to participate in the testing. All subjects were studied at least 6 months after surgery, with a mean period from surgery to evaluation of 5.8 yr (range, 6 months to 15.7 yr). All patients had pathological confirmation of a GH-secreting tumor at surgery. There were 34 men and 26 women (mean age, 48 yr; range, 28–77 yr). The mean body mass index (BMI) of the acromegalic subjects was 26.7 kg/m2. Thirteen patients had undergone radiotherapy from 3–11 yr before testing. Two of the subjects who were classified with active disease and one subject in remission group II (see Results) were being treated with octreotide injections, except on the morning of testing. Five subjects had previously been treated with bromocriptine, which had been stopped by their physicians for various reasons; 2 subjects had been treated with octreotide in the past, but had discontinued it because of side-effects. Neither medication had been taken for at least 3 months in any of these 7 subjects. All patients were ambulatory, and no patients had active hepatic or renal disease, glucose intolerance, or diabetes mellitus.
Twenty-five healthy subjects (14 men and 11 women; mean age, 47 yr; range, 27–71 yr) were also studied. The healthy subjects were selected based on health status (no significant medical problems) and age so that they would very closely match the distribution and mean age of our subjects with acromegaly. The mean BMI of the healthy subjects was 23.4 kg/m2. Healthy subjects had no history of endocrine disease, including diabetes mellitus, and were taking no medications, aside from multivitamins and daily aspirin in 2 male subjects.
Premenopausal healthy women and women with acromegaly were studied in the early follicular phase of the menstrual cycle (days 1–5). Women receiving hormone replacement therapy were studied just before a cycle of therapy.
Study procedures
After an overnight fast, all subjects underwent oral glucose tolerance testing. While seated, subjects had blood sampled at baseline and then 60, 90, and 120 min after drinking a 100-g dextrose drink (Trutol 100). Blood was allowed to clot at room temperature for 15 min and was then centrifuged; the serum was frozen at -20 C in multiple aliquots. Baseline blood samples were assayed for IGF-I and IGFBP-3, and blood samples at all time points were assessed for GH by polyclonal RIA and immunoradiometric assay (IRMA). Nadir GH samples measured by IRMA were also assayed for GH by a highly sensitive enzyme-linked immunosorbent assay (ELISA). All samples from each patient were run in the same assay and in duplicate. Serum glucose was measured at baseline and 2 h postdextrose administration by the glucose hexokinase method with a Hitachi 704 autoanalyzer (Hitachi, Tokyo, Japan). Serum glucose levels were less than 120 mg/dL at baseline and less than 200 mg/dL at 2 h after oral glucose in all subjects.
This protocol was approved by the institutional review board of Columbia-Presbyterian Medical Center (New York, NY), and informed consent was obtained from all subjects.
Assays
GH. Polyclonal RIA: GH was measured by RIA using
a polyclonal antiserum raised in rabbits immunized with purified human
GH. Purified human GH was obtained from the National Hormone and
Pituitary Program (AFP-4793B). The human GH reference preparation used
for standards in this RIA is equal in potency to the International
Reference Preparation of human GH (code 88/624), the most recent
calibrator from the WHO. Highly purified human GH (NIDDK h-GH-I-3) was
labeled with 125I by the lactoperoxidase method for use as
tracer. Bound and free hormone were separated by precipitation with a
sheep antiserum to rabbit 
-globulin. The intraassay coefficient of
variation was 6%, and the interassay coefficient of variation was 8%.
The assay had a sensitivity of 0.5 µg/L (17).
IRMA: GH was measured by a two-site IRMA obtained from Diagnostic Systems Laboratories, Inc. (Webster, TX). This IRMA assay used two highly specific anti-hGH mouse monoclonal antibodies, one immobilized to the inside wall of the assay tube, and the other in solution, labeled with 125I. The standards contained 22K recombinant hGH and were calibrated to the WHO International Reference Preparation of human GH (code 88/624), the most recent calibrator from the WHO. There was no cross-reactivity with other human pituitary hormones, including hPRL, or with other species of GH. The intraassay coefficient of variation was 3.1%, and the interassay coefficient of variation was 5.9%. The assay sensitivity in our laboratory was 0.05 µg/L.
ELISA: GH was measured by an ultrasensitive human GH ELISA obtained from Diagnostic Systems Laboratories. The ELISA is an enzymatically amplified sandwich-type immunoassay involving two highly specific mouse monoclonal anti-hGH antibodies and the biotin-streptovidin bridging detection system. The standards contained 22K recombinant hGH and were calibrated to the WHO International Reference Preparation of human GH (code 88/624), the most recent calibrator from WHO. The intraassay coefficient of variation was 4%, and the interassay coefficient of variation was 6%. The assay in our laboratory had a sensitivity of 0.01 µg/L.
IGF-I. IGF-I was measured by RIA using a polyclonal rabbit antibody generated against human IGF-I obtained from Nichols Institute Diagnostics (San Juan Capistrano, CA). In this assay, soluble IGF-I was separated from its binding proteins by extraction with acid-ethanol. Recombinant human IGF-I was used for the standards and labeled with 125I for the tracer. The antiserum for IGF-I showed virtually no cross-reactivity with other peptide hormones, IGF-II, or GH. The standard was calibrated against WHO First International Reference Reagent 1988, IGF-I 87/518. The intraassay coefficient of variation was 4%, and the interassay coefficient of variation was 11%. The assay sensitivity was 13.5 µg/L serum. The normal ranges for this assay are: age 16–24 yr, 182–780 µg/L; 25–39 yr, 114–492; 40–54 yr, 90–360; and more than 55 yr, 71–290. IGF-I levels in all healthy subjects and acromegalic subjects were compared to their age-appropriate normal ranges.
IGFBP-3. IGFBP-3 was measured with a two-site IRMA from Diagnostic Systems Laboratories. Serum samples were added to tubes precoated with goat anti-IGFBP-3 polyclonal antibody to which goat anti-IGFBP-3 125I-labeled polyclonal antibody was added. The intraassay coefficient of variation was 7%, and the interassay coefficient of variation was 9%. The assay sensitivity was 50 µg/L serum. Normal values for IGFBP-3 compiled by the manufacturer were not normally distributed. Therefore, normal data were first log normalized and then transformed to produce a normogram provided by the manufacturer. Our subjects were compared to this normogram and also to our healthy subjects. An elevated value was considered to be more than 2 SD above the mean for that age. Grouping ages together, the upper limits of normal are: age 28 yr, 6470 µg/L; 29 yr, 6052 µg/L; 30–34 yr, 6122 µg/L; 35–39 yr, 4263 µg/L; 40–44 yr, 4298 µg/L; 45–49 yr, 5171 µg/L; 50–59 yr, 4061 µg/L; and more than 60 yr, 3795 µg/L.
Statistical analysis
Nadir GH was defined as lowest value at any time point after oral glucose administration. The upper limit of normal nadir GH was defined as the GH value 2 SD above the mean nadir of the healthy subjects as measured by IRMA. Mean values for nadir GH, baseline IGF-I, and IGFBP-3 levels were calculated for each group of subjects defined below and are expressed ±SE. Mean nadir GH values in healthy subjects, subjects with active acromegaly, and subjects in remission were compared by factorial ANOVA with post-hoc testing using the Bonferroni-Dunn procedure. Nadir GH values were normally distributed in the study population, excluding data from subjects with active disease, which, when included, produced a somewhat upward skew of the data. Therefore, nonparametric analysis of nadir GH values among the groups was also performed with the Kuskal-Wallis test to confirm the results of the ANOVA. IGFBP-3 levels were compared among groups by ANOVA with post-hoc testing. IGFBP-3 levels were normally distributed overall and in each group of our subjects. Because the normal range data for IGFBP-3 reported by the manufacturer did not normally distribute, we confirmed the results of the IGFBP-3 analysis by ANOVA by nonparametric testing with the Kuskal-Wallis test. Correlation of nadir GH values by RIA vs. IRMA and IRMA vs. ELISA as well as IGF-I vs. IGFBP-3 and IGFBP-3 vs. nadir GH were performed with simple regression analysis and confirmed by Spearman rank correlation. The relationship among IGFBP-3, IGF-I, and nadir GH levels was compared with a multiple regression model.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
IGF-I levels. The mean IGF-I level in healthy subjects was
258 ± 18 µg/L (range, 106–449 µg/L); levels for all subjects
were within their age-appropriate normal ranges (Table 1
).
|
GH by IRMA. The mean GH nadir measured by IRMA was 0.09 ± 0.004 µg/L in the healthy subjects, with a range of 0.06–0.13 µg/L. The upper limit of the normal GH nadir was defined as the value that was 2 SD above the mean nadir of the healthy subjects by IRMA, or 0.14 µg/L.
GH by ELISA. The mean GH nadir measured by ELISA was 0.04 ± 0.01 µg/L, with a range of 0.01–0.13 µg/L in the healthy subjects.
Subjects with acromegaly
The 60 postoperative patients with acromegaly were divided based on IGF-I levels. Described first are the subjects with active disease, defined by IGF-I levels above their age-appropriate normal ranges. Described next are the subjects in remission, defined by IGF-I levels within the normal range for age. Two subjects in the active group (patients 6 and 22) were being regularly treated with octreotide, which was not given on the morning of the test.
Subjects with active acromegaly: elevated IGF-I levels
IGF-I levels. Twenty-two subjects with acromegaly had elevated
IGF-I levels. The mean IGF-I level was 742 ± 63 µg/L, with a
range of 490-1356 µg/L (Table 1
).
GH by RIA. The mean GH nadir by RIA in this group of subjects was 3.9 ± 0.54 µg/L, with a range of 1.1–11.7 µg/L. Four subjects with active disease had GH nadirs of 2.0 µg/L or less by RIA, overlapping with the nadir GH levels in the healthy subjects. Thus, when IGF-I level was used to define disease status, the GH nadir by RIA was unable in some cases to distinguish subjects with active disease from healthy subjects.
GH by IRMA. The mean GH nadir measured by IRMA failed to
suppress into the normal range, as defined by our healthy subjects, in
all patients with active acromegaly. The mean GH nadir was 1.3 ±
0.25 µg/L in this group, with a range of 0.33–5.0 µg/L (Table 2). In this group with active disease,
nadir GH values were less than 1.0 µg/L in 50% (11 of 22) of the
subjects. There was no overlap with nadir GH in healthy subjects. GH
measured by IRMA helped clarify disease status in the 4 subjects whose
GH by RIA appeared to adequately suppress to 2.0 ng/mL or less; when
measured by IRMA, their nadir GH values remained above the normal range
established by our healthy subjects, consistent with the elevated IGF-I
level and active disease status.
|
Subjects with acromegaly in remission: normal IGF-I levels
IGF-I levels. Thirty-eight of the subjects with acromegaly
were in remission, defined by normal IGF-I levels; the mean IGF-I level
in these subjects was 280 ± 12 µg/L, with a range of 148–470
µg/L (Table I).
GH by RIA. The mean nadir GH by RIA in the subjects in remission was 1.1 ± 0.11 µg/L, with a range of 0.5–2.4 µg/L.
GH by IRMA. The mean nadir GH by IRMA in these subjects was 0.19 ± 0.03 µg/L, with a range of 0.05–0.9 µg/L.
GH by ELISA. The mean nadir GH by ELISA was 0.23 ± 0.04 µg/L, with a range of 0.02–1.0 µg/L.
The mean nadir GH by IRMA in the remission group was significantly
lower than that in the active disease group (0.19 ± 0.03
vs. 1.3 ± 0.25 µg/L; P < 0.0001)
and was not significantly different from the healthy subjects. The
majority of subjects in the remission group had nadir GH values that
fell within the range of the healthy subjects (i.e. <0.14
µg/L), but 15 subjects (39%) had nadir GH values above this level.
Thus, this latter group of 15 subjects had abnormal GH suppression
compared to our healthy subjects. To better define the status of
disease in the subjects with acromegaly considered to be in remission,
they were further divided based on the nadir GH by IRMA into two groups
(Table 3).
|
GH by IRMA. Twenty-three subjects in the remission group had nadir GH by IRMA results within the normal range defined by our healthy subjects. In this group, nadir GH by IRMA was 0.10 ± 0.006 µg/L, with a range of 0.05–0.14 µg/L, virtually identical to the range in our healthy subjects and not overlapping the range in the active subjects.
GH by ELISA. The GH nadir by ELISA in these subjects was similar at 0.09 ± 0.01 µg/L, with a range of 0.02–0.14 µg/L. The mean IGF-I level was 281 ± 22 µg/L in group I.
Group II: normal IGF-I and abnormal GH suppression by IRMA
GH by IRMA. A subgroup of the subjects in remission was identified who, despite normal IGF-I levels, did not adequately suppress GH as measured by IRMA compared to our healthy subjects. This subgroup (group II) had a mean GH nadir by IRMA of 0.35 ± 0.07 µg/L, with a range of 0.15–1.2 µg/L. All subjects in group II had nadir GH levels greater than the normal range in the healthy subjects (i.e. >0.14 µg/L), and four subjects had nadir GH values that overlapped those of the subjects with active disease (range, 0.36–1.2 µg/L).
GH by ELISA. The mean GH nadir by ELISA in group II was 0.50 ± 0.11 µg/L, with a range of 0.15–1.8 µg/L.
The mean IGF-I level was 270 ± 16 µg/L in group II, which was not significantly different from that in group I.
Comparison of GH levels by RIA, IRMA, and ELISA
Both basal and glucose-suppressed GH values were higher by RIA
than by IRMA in all samples. Mean basal GH levels by RIA vs.
IRMA were 2.0 ± 0.49 vs. 0.56 ± 0.25 µg/L in
healthy subjects, 5.1 ± 0.80 vs. 1.8 ± 0.33
µg/L in the active group, and 2.1 ± 0.29 vs.
0.55 ± 0.13 µg/L in the remission group. When a ratio of basal
RIA/IRMA GH was calculated in each subject, the mean ratio of basal
RIA/IRMA GH levels in subjects overall was 6. Similarly, the mean ratio
of nadir RIA/IRMA GH values was 7. There was variability in the ratios
of GH measured by RIA vs. IRMA in all groups; as expected,
the difference between values in the two assays was greatest at lower
GH values in both basal and nadir samples. For example, the mean ratios
of nadir RIA/IRMA were 9.8 in healthy subjects, 9.4 in group I, 6.0 in
group II, and 3.5 in the active group. Similarly, regression analysis
of nadir GH by RIA vs. IRMA (Fig. 1) found that although overall there was
a good correlation between the assays (r = 0.891;
P < 0.0001), at low values, GH measurements by these
two methods were actually quite discrepant. For example, the GH nadir
as measured in the IRMA assay correlated well with nadir GH values
obtained by RIA at GH values more than 0.14 µg/L (r = 0.872),
but not below this level (r = 0.264). By contrast, nadir GH by
IRMA assay correlated with nadir GH by ELISA assay throughout the range
of GH levels, at GH values less than 0.14 µg/L (r = 0.84;
P = 0.005) and at GH values more than 0.14 µg/L
(r = 0.919; P = 0.0001). Although in our
laboratory the ELISA assay had an approximately 5-fold greater
sensitivity, nadir values (as shown above) were very similar in IRMA
and ELISA assays, and the upper limit of nadir GH after glucose in
healthy subjects was equally detectable in both assays. When the
standards from each of the three GH assays were run in the other
assays, the results were comparable over the range of 1–20 ng/mL.
|
Figure 2 illustrates nadir GH
vs. IGF-I levels in the different groups of subjects. In
healthy subjects and subjects in remission group I, nadir GH by IRMA
and IGF-I did not correlate (r = 0.076). However, in subjects with
active disease combined with those in remission with failure of
adequate GH suppression (group II), the nadir GH level did correlate
with the IGF-I level (r = 0.705; P < 0.0001). The
relationship was clearly stronger in the subjects with active disease.
Within the active group and group II, the relationship between IGF-I
and nadir GH was curvilinear.
|
IGFBP-3 levels were elevated overall in subjects with active
acromegaly, but there was overlap with the normal range; 32% of the
subjects with active disease had IGFBP-3 levels within 2 SD
above the normal mean. IGFBP-3 levels were significantly higher in
subjects with active acromegaly (4940 ± 301 µg/L)
vs. healthy subjects (2887 ± 153 µg/L;
P < 0.0001) and vs. the subjects in
remission (2966 µg/L; P < 0.0001). Mean IGFBP-3
levels were not significantly different among healthy subjects, all
subjects in remission, or subjects in group I (mean, 2838 µg/L) or
group II (mean, 3066 µg/L; Fig. 3). The
distribution of IGFBP-3 levels around the mean was similar in the
healthy subjects and groups I and II of the subjects in remission. In
addition, with multiple regression analysis, failure of adequate GH
suppression in the subjects in remission did not predict a higher level
of IGFBP-3.
|
). Interestingly, the relationship
between IGFBP-3 and IGF-I levels was strongest in group II of the
subjects in remission (r = 0.840; P <
0.0001).
|
The mean BMI was higher in the subjects with acromegaly than in healthy subjects (26.7 ± 0.57 vs. 23.5 ± 0.56; P < 0.001). However, BMI failed to correlate with nadir GH measured by RIA, IRMA, or ELISA; analysis with a multiple regression model confirmed that higher BMI was not predictive of failure of GH suppression. These data suggest that despite a higher BMI, the failure to suppress in some acromegalic subjects is not attributable to a higher BMI. BMI did not correlate with IGF-I levels.
There was no correlation between age and nadir GH level by IRMA or ELISA in all subjects or when each subgroup was examined separately. We were also unable to appreciate a significant difference in nadir GH after glucose administration by IRMA between male and female subjects. Of the 11 healthy women, 5 were premenopausal and 5 were postmenopausal, 1 of whom was receiving oral hormone replacement therapy (HRT). Of the 26 women with acromegaly, 9 were premenopausal with regular menstrual cycles, 2 were premenopausal with irregular cycles, 13 were postmenopausal and not receiving HRT, and 2 were postmenopausal and receiving oral HRT. Nadir GH levels in premenopausal vs. postmenopausal subjects overall or in premenopausal vs. postmenopausal women within each group were not significantly different. However, the number of subjects in each group was very small, making it difficult to draw any firm conclusions about the potential relationship or lack of relationship between menstrual status or HRT and GH suppression in our subjects.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In our evaluation of 60 postoperative patients with acromegaly, we identified a group of 22 subjects who had biochemical evidence of residual disease defined by elevated IGF-I levels. All the subjects in our group with elevated IGF-I levels and thus with active disease failed to adequately suppress GH by IRMA into the normal range defined by our healthy subjects, i.e. all nadir GH values were above 0.33 µg/L in the subjects with active disease compared to less than 0.14 µg/L in the healthy subjects. Clearly, there was no overlap in nadir GH measured by IRMA between the healthy subjects and the subjects with active disease. By contrast, nadir GH values with the RIA assay overlapped between healthy subjects and subjects with active acromegaly. Thus, with the IRMA assay we were able to identify our subjects with active disease as distinct from the healthy subjects based on IGF-I level and nadir GH after glucose suppression.
One of the most striking findings to emerge from our analysis is the extent of GH suppression with oral glucose that can be demonstrated in patients with active acromegaly. Our data show that when subjects with active acromegaly are identified by elevated IGF-I levels, the nadir GH measured by highly sensitive IRMA or ultrasensitive ELISA assays is between 0.33–1.0 µg/L in up to 50% of the subjects. Therefore, as many as half of the subjects with active disease would have been considered cured if criteria previously considered strict, i.e. GH suppression to less than 1.0 µg/L by less sensitive assays, were applied. This finding is of considerable importance in the clinical management of patients with acromegaly, particularly the postoperative patient who may have only moderate GH excess. A suppression of GH to less than 1.0 µg/L measured by a sensitive IRMA assay does not necessarily imply normal GH secretory dynamics and remission in acromegaly. GH levels in the postoperative patient need to be evaluated in detail with attention to the characteristics of the particular GH assay used and in conjunction with a reliable IGF-I measurement.
As suspected, based on our clinical experience with the evaluation of patients with acromegaly after surgery, we found that GH values measured by RIA overlapped considerably among all groups of subjects. The discrimination was especially poor at GH levels of 2.0 µg/L or less. By contrast, the IRMA and ELISA assays correlated well and also discriminated better between GH levels in the low range of less than 2.0 µg/L. Values determined by RIA were higher in all samples than those determined by IRMA, although the degree of difference varied in the groups and was more pronounced at lower levels of GH. It has been known that GH measurements do vary in different assays (30, 31, 32). In general, IRMA assays, such as that used here, are more specific than RIAs. The IRMA assay detects only the 22K form of GH and not the 20K or other circulating GH fragments that are detected by polyclonal RIA. As the 22K standards for each assay compared well in the other, this does not explain differences in the values reported in each assay. Furthermore, we would not expect variations in secreted GH between the groups studied to account for these differences, because monomeric GH circulating in acromegaly has been found to be qualitatively indistinguishable from normal GH (33, 34). The greater discrepancy between GH values at the lower range is probably due to the greater sensitivity of these assays compared to the RIA. This is supported by the finding that a number of samples that had been comparable by RIA were as much as 5- to 10-fold different when examined with IRMA or ELISA. Although IRMA and ELISA compared very well, we found IRMA to be the more useful in evaluating both healthy subjects and subjects with acromegaly because of the greater dynamic range of the IRMA, up to 30 vs. 0.5 µg/L in this ELISA without dilution.
Our data also demonstrate that the upper limit of nadir GH after oral glucose administration in a group of healthy subjects was 0.14 µg/L by IRMA and 0.12 µg/L by ELISA (mean + 2 SD for each assay). By contrast, when nadir GH levels were measured with the less sensitive polyclonal RIA, a distinct normal range could not be identified. Others have shown that when the nadir GH after oral glucose was measured with an ultrasensitive chemiluminescence assay, GH levels were suppressed to a mean of 0.029 µg/L in young healthy men (n = 9) and 0.25 µg/L in young healthy women (n = 6) (9). In another study, the mean nadir GH in healthy subjects (n = 37) was 0.15 µg/L with a highly sensitive enzyme immunoassay for GH (10). Most subjects in these prior studies had nadir GH values below our upper limit of normal, but the range of values in our healthy subjects was smaller, and none of our 25 healthy subjects had a nadir above 0.14 µg/L by IRMA or ELISA. Our healthy population, which is age matched to the population of subjects with acromegaly in our study, is clearly much older than those studied by others; the mean age of our healthy subjects was 47 yr vs. 24.5 yr in a previous study (9). It is possible that the lower mean nadir GH we found in our healthy subjects is due to their older age, but we were unable to demonstrate a significant relationship between age and nadir GH in our healthy subjects. Taken together, our data and these prior studies suggest that older GH assays were insufficiently sensitive to measure the full extent of GH suppression in healthy subjects and many subjects with acromegaly.
Our analysis also revealed a subgroup of those patients in apparent remission who, despite normal IGF-I levels, inadequately suppressed GH after glucose administration, i.e. the GH nadir was more than 0.14 µg/L by IRMA. Failure of adequate suppression of GH after glucose treatment may signify GH dysregulation in these patients. Others have shown that patients with active acromegaly have abnormal pulsatile GH secretion, and although these parameters typically normalize after successful surgery, abnormal GH pulsatility may persist after apparent cure for acromegaly (35, 36, 37, 38, 39). It is possible that the inadequate GH suppression seen in group II reflects persistent mild postoperative GH dysregulation and/or excess. It is not known what anatomical or other factors could also play a role in leading to this inadequate GH suppression. One member of remission group II, who had a normal IGF-I level but a nadir GH of 0.4 µg/L, had been taking octreotide at the time of testing. In this patient, it is possible that changes in the normal GH/IGF-I relationship could have been altered by octreotide therapy. Prior radiotherapy could be associated with disruption of somatostatin tone (40), and thus in part with the mechanisms responsible for glucose-induced suppression of GH secretion. However, although six members of group II with inadequate GH suppression had undergone radiotherapy, three members of group I who had normal GH suppression by our criteria also had had prior radiotherapy. Hypopituitarism did not appear to be associated with failure of GH suppression, as only two members of group II with abnormal GH suppression and three members of group I with normal GH suppression had some degree of hypopituitarism.
If the hypothesis is true that inadequate GH suppression in some patients in group II is due to mild GH excess and dysregulation, suggesting some residual disease, then these patients may be at risk for future recurrence of clinically active disease. Some other work has suggested that failure of adequate GH suppression may herald an early recurrence; five of nine postoperative patients with normal IGF-I levels, but failure of GH to suppress, did recur (41). Our data suggest that use of IRMA, as opposed to RIA, may lead to the identification of more postoperative patients with acromegaly with clearly defined abnormal GH suppression, thereby possibly identifying more patients at risk for recurrence. Clearly, longitudinal follow-up will be needed to determine the true significance of abnormal GH suppression detected with highly sensitive GH assays in patients with apparent remission defined by normal IGF-I levels.
Data conflict regarding the usefulness of IGFBP-3 measurements in the assessment of disease status in patients with acromegaly after surgery. IGFBP-3 levels are usually elevated in untreated acromegaly and, in general, return to normal after successful treatment (42, 43). In recent work, IGF-I and IGFBP-3 levels were highly correlated in 7 of 10 postoperative subjects, and both were also concordant with GH suppressibility, suggesting that IGFBP-3 levels may be useful in evaluating postoperative patients (44). Others have also reported a correlation between IGF-I and IGFBP-3 in patients with active acromegaly, but with an overlap between the values in healthy subjects and those with disease (45, 46, 47, 48). These investigators concluded that the IGFBP-3 level is a less sensitive test for detecting disease than the IGF-I level, because this did not overlap between healthy and acromegalic subjects (45). We, too, found a correlation between IGF-I and IGFBP-3 levels; however, despite inadequate GH suppression and elevated IGF-I levels, 32% of the subjects with active disease had IGFBP-3 levels within the normal range. This finding demonstrates that a normal IGFBP-3 level should not be considered proof of restored GH secretory dynamics in acromegaly. Some have considered it useful to identify a cutoff of nadir GH after glucose suppression below which both IGFBP-3 and IGF-I levels are normal. We did find that in all subjects but one that a nadir GH of less than 0.32 µg/L was associated with normal IGF-I and IGFBP-3 levels. However, the fact that some subjects with active disease had normal IGFBP-3 levels, and some subjects in remission had nadir GH levels above 0.33 calls into question the clinical usefulness of this type of generalization.
It has also been suggested that in some patients whose IGF-I level is normal but whose GH inadequately suppresses, the IGFBP-3 levels may be elevated and could indicate mild GH excess. We were unable to demonstrate that failure of GH suppression could be associated with higher IGFBP-3 levels in patients in remission. However, longitudinal data are still needed to clarify the role of IGFBP-3 measurements in the management of postoperative patients and determine whether failure of adequate GH suppression detected by sensitive GH assays could precede elevation of IGFBP-3 or IGF-I levels.
As GH secretion is influenced by weight (49), our analysis also considered the possible effect of BMI on the degree of GH suppression by glucose. Although baseline GH levels may be lower in obese subjects (50, 51, 52), GH levels after iv glucose administration may not suppress as much in obese subjects as in nonobese subjects (53, 54). Although BMI was higher in the subjects with acromegaly, there was no correlation of nadir GH levels with BMI in our subjects overall or in any subgroup, indicating that we could not explain differences in GH suppression between groups based on the BMI.
Estradiol levels have been shown to correlate with indexes of GH secretion in premenopausal women and postmenopausal women receiving different forms of hormone replacement therapy (51, 55). Chapman et al. found a significant gender difference in the nadir GH responses to glucose, with nadir GH being higher in young healthy women than in men (9). We did not detect a significant gender difference in nadir GH levels, probably because our population is older than those previously studied and contains many postmenopausal women.
Conclusion
The use of highly sensitive GH assays in the evaluation of postoperative patients with acromegaly will have a significant impact on our recognition of residual disease in these patients. Our data demonstrate that whereas by previous criteria normal GH suppression was considered to be less than 2.0 µg/L or, more recently, less than 1.0 µg/L, many patients with persistent disease after surgery may have nadir GH levels after glucose administration of less than 1.0 µg/L as measured by a sensitive IRMA assay. Therefore, we have shown that suppression of GH below 1.0 µg/L after oral glucose administration does not necessarily indicate normal GH secretory dynamics in acromegaly. In addition, we have confirmed other studies by showing that normal nadir GH levels after glucose suppression measured by sensitive assays in subjects of all ages are much lower than previously recognized with our polyclonal RIA or other commercially available RIAs. It remains to be determined on longitudinal follow-up if use of highly sensitive GH assays to evaluate disease status in patients with acromegaly will identify subjects with abnormal GH suppression despite apparent remission with normal IGF-I levels who may be at risk for recurrence of clinically active disease.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received May 27, 1998.
Revised July 30, 1998.
Accepted August 6, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. M. Arafat, M. Mohlig, M. O. Weickert, F. H. Perschel, J. Purschwitz, J. Spranger, C. J. Strasburger, C. Schofl, and A. F. H. Pfeiffer Growth Hormone Response during Oral Glucose Tolerance Test: The Impact of Assay Method on the Estimation of Reference Values in Patients with Acromegaly and in Healthy Controls, and the Role of Gender, Age, and Body Mass Index J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1254 - 1262. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Alexopoulou, M. Bex, R. Abs, G. T'Sjoen, B. Velkeniers, and D. Maiter Divergence between Growth Hormone and Insulin-Like Growth Factor-I Concentrations in the Follow-Up of Acromegaly J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1324 - 1330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Misra, J. Cord, R. Prabhakaran, K. K. Miller, and A. Klibanski Growth Hormone Suppression after an Oral Glucose Load in Children J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4623 - 4629. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Kauppinen-Makelin, T. Sane, H. Sintonen, H. Markkanen, M. J. Valimaki, E. Loyttyniemi, L. Niskanen, A. Reunanen, U.-H. Stenman, and and the Finnish Acromegaly Study Group Quality of Life in Treated Patients with Acromegaly J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 3891 - 3896. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Woodhouse, A. Mukherjee, S. M. Shalet, and S. Ezzat The Influence of Growth Hormone Status on Physical Impairments, Functional Limitations, and Health-Related Quality of Life in Adults Endocr. Rev., May 1, 2006; 27(3): 287 - 317. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Markkanen, T. Pekkarinen, M. J. Valimaki, H. Alfthan, R. Kauppinen-Makelin, T. Sane, and U.-H. Stenman Effect of Sex and Assay Method on Serum Concentrations of Growth Hormone in Patients with Acromegaly and in Healthy Controls Clin. Chem., March 1, 2006; 52(3): 468 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Feelders, M. Bidlingmaier, C. J. Strasburger, J. A. M. J. L. Janssen, P. Uitterlinden, L. J. Hofland, S. W. J. Lamberts, A. J. van der Lely, and W. W. de Herder Postoperative Evaluation of Patients with Acromegaly: Clinical Significance and Timing of Oral Glucose Tolerance Testing and Measurement of (Free) Insulin-Like Growth Factor I, Acid-Labile Subunit, and Growth Hormone-Binding Protein Levels J. Clin. Endocrinol. Metab., December 1, 2005; 90(12): 6480 - 6489. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Puder, S. Nilavar, K. D. Post, and P. U. Freda Relationship between Disease-Related Morbidity and Biochemical Markers of Activity in Patients with Acromegaly J. Clin. Endocrinol. Metab., April 1, 2005; 90(4): 1972 - 1978. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Wang, F. Paris, C. S. Sultan, R. X.-D. Song, L. M. Demers, B. Sundaram, J. Settlage, S. Ohorodnik, and R. J. Santen Recombinant Cell Ultrasensitive Bioassay for Measurement of Estrogens in Postmenopausal Women J. Clin. Endocrinol. Metab., March 1, 2005; 90(3): 1407 - 1413. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
), subjects in remission with normal GH
suppression (group I; 
), and subjects in remission with abnormal GH
suppression (group II; 
) in relation to the normal range of GH
suppression (mean ± 2SD of the healthy subjects
responses, 
) vs. the IGF-I level. Two subjects with
active disease had nadir GH levels of 5.0 and 3.9 µg/L and IGF-I
levels of 1356 and 1000 µg/L, respectively, were not included in the
figure.
