help button home button Endocrine Society JCEM JCEM Call for Nominations for EIC
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Journal of Clinical Endocrinology & Metabolism, doi:10.1210/jc.2005-0304
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
90/11/6048    most recent
Author Manuscript (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Björk, J.
Right arrow Articles by Erfurth, E. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Björk, J.
Right arrow Articles by Erfurth, E. M.
Related Collections
Right arrow Endocrine Oncology
Right arrow Neuroendocrinology and Pituitary
Right arrow Pediatric Endocrinology
The Journal of Clinical Endocrinology & Metabolism Vol. 90, No. 11 6048-6054
Copyright © 2005 by The Endocrine Society

The Utility of the Growth Hormone (GH) Releasing Hormone-Arginine Test for Diagnosing GH Deficiency in Adults with Childhood Acute Lymphoblastic Leukemia Treated with Cranial Irradiation

Jonas Björk, Katarina Link and Eva Marie Erfurth

Competence Center for Clinical Research (J.B.), Department of Endocrinology (K.L., E.M.E.), Lund University Hospital, SE 221 85 Lund, Sweden

Address all correspondence and requests for reprints to: Dr. Eva Marie Erfurth, Department of Endocrinology, Lund University Hospital, SE-221 85 Lund, Sweden. E-mail: eva-marie.erfurth{at}med.lu.se.


   Abstract
Top
 Abstract
Introduction
 Patients and Methods
 Results
 Discussion
 References
 
Context: The insulin tolerance test (ITT) is the current standard diagnostic test for the diagnosis of adult GH deficiency (GHD), but alternative tests, such as the GHRH-arginine test, have been proposed.

Objective: We investigated the sensitivity and specificity of the GHRH-arginine test using ITT as the gold standard in diagnosing GHD in a group of young adults treated with cranial irradiation (CRT) for childhood acute lymphoblastic leukemia (ALL). We estimated the positive and negative predictive values of the GHRH-arginine test among patients as well as a number of individual characteristics and therapy-related factors during both the GHRH-arginine test and ITT.

Design: Forty-three young adults, treated for childhood ALL with 18–30 Gy CRT and chemotherapy, were studied, and comparison was made with matched controls.

Results and Conclusions: We evaluated four different cutoff levels for GHD in the GHRH-arginine test: 5, 7.5, 9, and 16.5 µg/liter. Using 7.5 µg/liter as the cutoff yielded high specificity (94%), but at the same time the sensitivity was only 66%, which leads to a low negative predictive value (27%). In contrast, a failed GH response to the GHRH-arginine test accurately reflects the presence of radiation-induced GHD, illustrated by a high positive predictive value (95% at 7.5 µg/liter). Only age at CRT and body mass index remained significant predictors of the peak GH during the GHRH-arginine test. Because a high proportion of GHD patients show a normal response to the GHRH-arginine test, it cannot be used reliably to exclude GHD in these patients. Complementary ITT is also warranted to confirm GHD in obese patients.


   Introduction
Top
 Abstract
 Introduction
Patients and Methods
 Results
 Discussion
 References
 
CRANIAL RADIOTHERAPY (CRT) is a potent cause of hypopituitarism, and the severity is related to dose, the fraction schedule, and the postirradiation time interval (1, 2). CRT will affect the hypothalamus, which is more sensitive to irradiation than is the anterior pituitary (3, 4) and GH deficiency (GHD) is usually the first and often the only established endocrine sequel of CRT (1, 5). The insulin tolerance test (ITT) is the current standard diagnostic test for the diagnosis of adult GHD (6). However, this test has been questioned due to its poor reproducibility and the facts that it is labor intensive and there are common contraindications. In contrast, the test has the advantage that it allows evaluation of the complete hypothalamic-somatotroph axis (7). The need for accurate alternative tests has been emphasized (6), and the GHRH-arginine test has been proposed as a useful, reproducible, and age-independent alternative provocative test (6, 8, 9, 10). This test has been validated profoundly in patients with presumed direct pituitary damage from a tumor and/or surgery (11). However, the majority of patients with CRT-induced GHD remain responsive to GHRH analogs (12, 13). Very recently, a comparison between these two tests showed that the GHRH-arginine test resulted in false negative diagnosis in the early years, i.e. within 5 yr after CRT (14). In contrast, the ITT was more accurate within the first 5 yr after CRT (14). Furthermore, the poor utility of the GHRH-arginine test in irradiated patients was illustrated by only a 50% sensitivity (14). The vast majority of these patients were exposed to high doses of both focal and whole brain CRT (median, 58.3 Gy) and were treated in both childhood and adulthood (14). Thus, more questions need to be answered regarding the sensitivity and specificity of the GHRH-arginine test compared with the ITT in patients treated with a lower CRT dose, e.g. 18–30 Gy.

The most common childhood cancer is acute lymphoblastic leukemia (ALL), with an annual incidence of three or four per 100,000 children (15). Treatment of childhood ALL, including multiagent chemotherapy and prophylactic CRT, has markedly improved the survival rate for these children (16). However, CRT in children frequently causes abnormal hypothalamic-pituitary function later in life (4), and GHD has been shown in childhood ALL after low doses of CRT (17, 18). In adults, the sensitivity of serum IGF-I in diagnosing GHD is low (19, 20); it is particularly low in adults with GHD and a history of CRT (18, 21, 22). Previously, it was shown that only years since treatment, but neither age at treatment nor dose of CRT, was of significant importance for the peak GH during an ITT in adults with childhood ALL (18). Furthermore, the impact of obesity on the GH response of the GHRH-arginine test has also been emphasized lately (23, 24). The corresponding impact of these variables on the peak GH during a GHRH-arginine test has not been validated previously in former ALL patients.

One purpose of the present study was to investigate the sensitivity and specificity of the GHRH-arginine test using the ITT as the gold standard in diagnosing GHD in a group of young adults treated with chemotherapy and CRT for childhood ALL. The ultimate aim was to estimate the positive and negative predictive values of the GHRH-arginine test in this group of patients. Another purpose was to investigate the impact of a number of individual characteristics and therapy-related factors during both the GHRH-arginine test and ITT in former childhood ALL patients. A homogenous group of 43 young adults, treated for childhood ALL with 18–30 Gy CRT and chemotherapy, was considered an appropriate study group. Patients were compared with controls randomly selected from the general population and individually matched for sex, age, smoking habits, and residence.


   Patients and Methods
Top
 Abstract
 Introduction
 Patients and Methods
Results
 Discussion
 References
 
Patients

A consecutive series of 58 patients treated for childhood ALL with chemotherapy and CRT during 1971–1992 at the Children’s Hospital (Lund, Sweden) and at least 18 yr of age were invited to participate. Exclusion criteria were presented previously (22). Characteristics, including radio- and chemotherapy, for the 43 patients are shown in Table 1Go. All patients had been off all chemotherapy for a median of 16.7 yr (range, 6.3–23.9 yr).


View this table:
[in this window]
[in a new window]
  		TABLE 1. Characteristics of the 43 patients and matched controls that performed the GHRH-arginine test

 
Initially, the children were treated according to the protocols of the Swedish Child Leukemia Group (25); since 1981, they have been treated according to the common protocols in the five Nordic countries (26, 27). The details of previous treatments were presented previously (22). Based on prognostic factors, the patients were allocated to different risk groups with varying treatment intensities. Central nervous system therapy included intrathecal methotrexate (MTX) and prophylactic CRT with 18–24 Gy; in cases of central nervous system involvement, the dose was 30 Gy. Maintenance therapy included 6-mercaptopurine and MTX; in some patients, cyclophosphamide was also given. In higher risk groups, reinductions with prednisolone and vincristine, and/or doxorubicin and intermediate dose MTX were applied. Five patients had received CRT in combination with spinal radiation (median, 23.0 Gy; range, 14–30 Gy). Twelve boys were given radiation to the testes. Five patients had been treated for relapse, and one patient had been treated for more than one relapse. One patient was treated with high dose chemotherapy with autologous stem cell rescue.

Four patients had previously received GH treatment in childhood, but not for at least 5 yr. All females had spontaneous and regular menstrual cycles, except one patient with primary amenorrhea and five patients using estrogen-containing contraceptives. Thus, all but one female had isolated GHD or GH insufficiency. Ten of the 12 males who had received radiation to the testes were given substitution therapy with testosterone as injections (n = 7), tablets (n = 1), or patches (n = 2). Two patients subjected to testes radiation had no androgen substitution; one had not been properly irradiated due to undescended testes and subsequently had normal testosterone levels, and the other was not given sufficient substitution and for the purpose of this study was excluded. Thus, all but 11 males, who were given sufficient testosterone substitution, had isolated GHD or GH insufficiency. One patient suffered from heart failure and hypertension, and was treated with digoxin and enalapril maleate; another had hypertension and was treated with atenolol and enalapril maleate; and two female patients received sertralin hydrochloride for depression. Three men and two women were smokers.

The GHRH-arginine test and the ITT were performed in all 43 former ALL patients. The ITT results from six of the patients were excluded, however; one patient due to insufficient substitution of testosterone, another patient due to technical problems during the ITT, and four patients because their ITTs were inconsistent, because acceptable hypoglycemia was not accomplished despite repetition of the test.

Control subjects

The aim was to select one control subject for each patient enrolled in the study. To achieve this, 10 potential control subjects, matched for age, gender, and residence (rural/nonrural), were selected randomly from a computerized register of the population in the catchment area of the patients (Southern Swedish Medical Region). Details of matching procedures and the outcome of the recruitment process were presented previously (22). Two controls were taking medication for depressive disorders. All patients and controls had normal thyroid, sex steroid, and cortisol levels (data not shown). All 43 controls performed the GHRH-arginine test only. Characteristics of the control subjects are given in Table 1Go.

Test procedures

The clinical investigation was performed as a cross-sectional study without any change in medication. In the morning after an overnight fast, all patients and control subjects underwent a GHRH-arginine testing with GHRH (Ferring, Malmo, Sweden; 1 µg/kg, iv, at 0 min) and arginine-hydrochloride (0.5 g/kg, iv, during the first 30 min) (10). The ITT was performed, at least 1 month after the GHRH-arginine test, after an overnight fast with 0.1–0.2 U/kg soluble insulin (Actrapid, Novo Nordisk A/S, Bagsvaerd, Denmark), iv. Hypoglycemia was confirmed when a blood glucose level below 2.2 mmol/liter was reached. Blood samples for analysis of serum GH for both tests were taken every 15 min from –15 to 90 min (10). The ethics committee of Lund University approved the protocol, and all subjects gave their written informed consent.

Cutoff levels for peak GH responses

Previously defined cutoff levels for GHD were defined as a peak GH response less than 3 µg/liter to an ITT (6, 19) and a peak response of 9 µg/liter to the GHRH-arginine test (10). Partial GHD or GH insufficiency is defined as a peak GH response between 3–5 µg/liter during the ITT (14) and between 9–16.5 µg/liter during the GHRH-arginine test (7, 9, 11). Normal GH levels are defined as a peak GH response of more than 5 µg/liter during the ITT (14) and more than 16.5 µg/liter during the GHRH-arginine test (10). For the GHRH arginine test, we also evaluated three other cutoff levels for GHD: 5, 7.5, and 16.5 µg/liter.

Biochemical assays

Blood samples were drawn in the morning, after fasting since midnight. Serum IGF-I was assayed by immunoradiometric assay (Nichols Institute Diagnostics, Inc., San Juan Capistrano, CA); the normal range was 122–400 µg/liter in subjects aged 19–40 yr. The intraassay coefficient of variation (CV) was 16% at 60 µg/liter and 11% at 300 µg/liter. Serum GH was analyzed by an immunofluorometric method (DELFIA hGH, Wallac Oy, Turku, Finland). The detection level for serum GH was 0.01 µg/liter, and the intra- and interassay CVs were 5% and 3%, respectively, at 1.5 µg/liter, and 3% and 5%, respectively, at 7.7 µg/liter. The kit standards were calibrated against the First International Standard 80/505, and the assay is specific for 22-kDa GH. Serum prolactin (PRL) was analyzed with a prolactin assay on the Modular Analytics E170 (Roche, Indianapolis, IN). The reference interval for women was 4–27 µg/liter, and that for men was 4–24 µg/liter. At 10 µg/liter, the CV was 3.9%. Venous blood glucose was analyzed with the Hemocue blood glucose analyzer (Hemocue AB, Angelholm, Sweden). According to the manufacturer, the SD between the cuvettes is less than 0.3 mmol/liter.

Statistics

In the calculations, all controls were assumed to have normal GH levels or to be GH insufficient at most according to the ITT. The area under the receiver operating characteristics (ROC) curve was used as an overall measure of the ability of the GHRH-arginine test to discriminate between GHD and, at most, GH insufficiency (28). The area under the ROC curve (percentage) can be interpreted as the probability that a randomly chosen individual with GHD has a lower peak GH in the GHRH-arginine test than a randomly chosen individual without GHD. For the classification of GHD using 5, 7.5, 9, or 16 µg/liter as the cutoff level for the GHRH-arginine test, 95% confidence intervals for the sensitivity and specificity were calculated based on the binomial distribution. In addition, a positive predictive value, defined as the probability that a positive test result reflects a true GHD individual, was calculated for the patient group. The corresponding negative predictive value, defined as the probability that a negative result reflects a true non-GHD individual, was also calculated. These predictive values reflect the expected performance of the GHRH-arginine test in predicting GHD among childhood ALL patients in clinical settings.

Throughout the text, we regarded P < 0.05 as statistically significant. Differences in peak GH during the GHRH-arginine test and in body mass index BMI among patients and controls were tested with the Wilcoxon signed-rank test. Bivariate associations among the patients (and, if feasible, among the controls) between peak GH during the GHRH-arginine test and gender, age at CRT, age at test, years since CRT, target dose of CRT, treatment with MTX above 1 g/m2, PRL, IGF-I, and BMI were assessed by Spearman rank correlation coefficient (continuous variables) and Mann-Whitney U test (dichotomous and dichotomized variables). In the multivariate analyses, multiple linear regression with an intercept term was employed. To obtain a more meaningful intercept term, a BMI of minus 20 with negative values assigned to zero was used as the BMI variable in the regression modeling. The natural logarithm of the outcome variable, peak GH during the GHRH-arginine test, was used to obtain regression residuals that conformed more closely to the normal distribution. The logarithmic transformation imposes a multiplicative relation between the explanatory variables of the regression model. Explanatory variables were excluded one at a time if they had P values above 0.30, starting with the regression model, with all significant variables from the bivariate analyses included. Explanatory variables with P values between 0.05 and 0.30 were also excluded one at a time if exclusion changed the regression coefficients of the significant explanatory variables by less than 10%. The results of the multivariate analyses are presented for the patients only.


   Results
Top
 Abstract
 Introduction
 Patients and Methods
 Results
Discussion
 References
 
Sensitivity, specificity, and predictive values of the GHRH-arginine test

The GHRH-arginine test was performed in 43 patients (Table 1Go), and the ITT results for 37 of these subjects could also be assessed (Fig. 1Go). The GH response to the GHRH-arginine test in the controls was markedly higher (median, 25; range, 3.2–110; Table 1Go) than that in the patients (median, 6.2; range, 0.85–43; P < 0.001). BMI was significantly higher in the patients than the controls (median, 25 vs. 22; Table 1Go; P = 0.002). Using the ITT as the gold standard for detecting GHD with less than 3 µg/liter as the cutoff level and assuming that the controls had normal GH levels or were, at most, GH insufficient, implied a high overall discrimination ability of the GHRH-arginine test [area under the ROC curve, 87%; 95% confidence interval (CI), 79–95%; Fig. 2Go]. The sensitivity of the GHRH-arginine test using 9 µg/liter as the cutoff level was 66% (95% CI, 47–81%; 21 of 32 patients with GHD according to the ITT were classified correctly; Table 2Go). The specificity of the GHRH-arginine test at 9 µg/liter was 88% (95% CI, 75–95%; 42 of 48 subjects without GHD were classified correctly). Before conducting the GHRH-arginine test, the probability that a randomly selected patient was GHD was 86% (32 patients with GHD of 37).



View larger version (12K):
[in this window]
[in a new window]
  		FIG. 1. Relation between peak GH values in the GHRH-arginine test and the ITT among the 37 patients, with results from both tests. The vertical and horizontal lines show cutoff levels of 7.5 and 3 µg/liter, respectively.

 


View larger version (12K):
[in this window]
[in a new window]
  		FIG. 2. ROC curve of the GHRH-arginine test for classifying GHD, defined as ITT below 3 µg/liter. The combination of sensitivity and specificity for the peak GH cutoff levels of 5, 7.5, 9, and 16.5 µg/liter are indicated (•). Note that the specificity axis is labeled in descending order.

 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Sensitivity and specificity of the GHRH-arginine test for 37 patients and 43 controls using a peak GH of 9 µg/liter as the cutoff level and using the ITT as the gold standard for GHD

 
Using 9 µg/liter as the cutoff level for the GHRH-arginine test, the probability of GHD increased to 91% if the GHRH-arginine test was positive (21 of 23 patients with GHRH-arginine test results <9 µg/liter were truly GHD according to the ITT; Table 3Go). If the GHRH-arginine test was negative, i.e. above 9 µg/liter, the probability of GHD decreased to 79%, i.e. the negative predictive value was 100–79 = 21% (three of 14 patients with GHRH-arginine test results of ≥9 µg/liter were not GHD according to the ITT). The sensitivity and specificity of GHD classification for other choices of cutoff levels for the GHRH-arginine test can be obtained from Table 4Go or from the ROC curve (Fig. 2Go). Aiming for a higher specificity of the classification by regarding the result of the GHRH-arginine test as GHD only if it was less than 7.5 µg/liter would increase the specificity of the test to 94% (95% CI, 83–99%; 45 of 48 individuals without GHD according to the ITT were classified correctly) while leaving the sensitivity unchanged at 66% (Table 4Go). Fifteen patients had a GH response to the GHRH-arginine test of at least 7.5 µg/liter, but only four of them were had normal GH levels on the basis of a GH response to the ITT of 3 µg/liter or higher, giving a negative predictive value of only 27% if this response was used to define normality in the GHRH-arginine test (Fig. 1Go). In contrast, 21 of 22 patients with a GH response to the GHRH-arginine test less than 7.5 µg/liter were GHD on the basis of a GH response to the ITT below 3 µg/liter, yielding a positive predictive value of 95% (Fig. 1Go). Using an even lower cutoff (5 µg/liter) would decrease the sensitivity substantially (Table 4Go). In contrast, aiming at a substantially higher sensitivity by increasing the cutoff level would reduce the specificity markedly. As an example, regarding GHRH-arginine test as GHD if the result was less than 16.5 µg/liter would imply a sensitivity of 88% (28 of 32 individuals with GHD according to the ITT were classified correctly) and a specificity of 69% (33 of 48 individuals without GHD according to the ITT were classified correctly). Such a broad classification of GHD based on the GHRH-arginine test implied positive and negative predictive values of 90% and 33%, respectively (Table 4Go).


View this table:
[in this window]
[in a new window]
  		TABLE 3. Predictive values of the GHRH-arginine test among 37 patients, using a peak GH of 9 µg/liter as the cutoff level and using the ITT as the gold standard for GHD

 

View this table:
[in this window]
[in a new window]
  		TABLE 4. Sensitivity, specificity, positive and negative predictive values of the GHRH-arginine test for various cutoff levels for the peak GH

 
Factors related to the GHRH-arginine test among the controls

A significant negative correlation between the GH peak during the GHRH-arginine test and BMI was observed among the controls (r = –0.58; P < 0.001; Fig. 3Go). Accordingly, controls with BMI less than 25 kg/m2 had significantly higher GH peak during the GHRH-arginine test (median, 32 µg/liter; range, 9.6–110 µg/liter; n = 32) than the controls with BMI of 25 kg/m2 or higher (median, 13 µg/liter; range, 3.2–35 µg/liter; n = 11; P < 0.001). Four of the 11 controls with BMI of 25 kg/m2 or higher had a GH peak during the GHRH-arginine test less than 9 µg/liter compared with none of the 32 controls with BMI below 25 kg/m2. A higher GH peak during the GHRH-arginine test was indicated, but was nonsignificant, in females (median, 34 µg/liter; range, 9.6–110 µg/liter; n = 21) than in males (median, 21 µg/liter; range, 3.2–65 µg/liter; n = 22; P = 0.06). A similar gender difference was indicated after adjustment for BMI in a multivariate analysis (results not shown). No apparent association between GH peak and age at test (P = 0.16), PRL (P > 0.30), or IGF-I (P = 0.14) was discerned.



View larger version (13K):
[in this window]
[in a new window]
  		FIG. 3. Relation between BMI (kg/m2) and peak GH (µg/liter) of the GHRH-arginine test in patients ({circ}) and controls (•) and of the ITT in patients ({blacktriangleup}).

 
Factors related to the GHRH-arginine test among the patients

The GH peak during the GHRH-arginine test correlated negatively with BMI (r = –0.54; P < 0.001; Fig. 3Go) and positively with age at CRT (r = 0.35; P = 0.02) among the patients. Accordingly, for the GHRH-arginine test, patients with BMI less than 25 kg/m2 had higher GH peak than patients with BMI of 25 kg/m2 or more (P = 0.003), and patients who were at least 5 yr old at CRT had higher GH peak than those who were younger (P = 0.02; Table 5Go). Moreover, there was a significant positive correlation between GH peak during the GHRH-arginine test and PRL level (r = 0.37; P = 0.01). The PRL level correlated significantly with age at CRT (r = 0.37; P = 0.01), but not with years since CRT (P = 0.18). Patients treated with MTX above 1 g/m2 had significantly higher responses to the GHRH-arginine test than the others [median, 10.4 (n = 9) vs. 4.8 (n = 34); P = 0.002]. However, patients treated with more than 1 g/m2 MTX were also treated with CRT at an older age than the others (median age, 13 vs. 4.3 yr). No associations between GH peak and gender (P = 0.19), age at test (P = 0.21), target dose of CRT (P > 0.30), years since CRT (P > 0.30), or IGF-I (P > 0.30) were discerned. In the multivariate regression analysis, only age at CRT and BMI remained significant predictors of the GH peak (Table 6Go). Each unit of BMI above 20 can be expected to decrease the GH peak by 10% (95% CI, 5–16%). Each extra year of age before exposure to CRT can be expected to increase the GH peak by 10% (95% CI, 4–17%).


View this table:
[in this window]
[in a new window]
  		TABLE 5. The outcome of the GHRH-arginine test among the patients (n = 43), measured as peak GH, stratified by age at CRT and BMI

 

View this table:
[in this window]
[in a new window]
  		TABLE 6. Multiple regression analysis exhibiting a significant association among the patients (n = 43) between the peak GH during the GHRH-arginine test, age at CRT, and BMI (adjusted R2 = 39%)

 
Factors related to the ITT in the patients

The peak GH during the ITT in the patients (n = 37) showed a weak correlation with age at treatment that was close to significance (r = 0.32; P = 0.05). BMI did not correlate with the ITT among the patients (r = 0.02; P > 0.30; Fig. 3Go), and the peak GH during the ITT in patients with BMI of 25 kg/m2 or more was not noticeably different (median, 1.0 µg/liter; range, 0.06–4.1 µg/liter) from that in patients with BMI below 25 kg/m2 (median, 0.75 µg/liter; range, 0.06–6.9 µg/liter; P > 0.30). Four patients of 20 had normal GH levels in the patients with BMI below 25 kg/m2 on the basis of a GH response to the ITT of 3 µg/liter or more compared with one patient of 17 with BMI of 25 kg/m2 or more. No obvious correlation between the peak GH during the ITT and target dose of CRT (P = 0.16), years since treatment (P > 0.30), PRL (P > 0.30), IGF-I (r = –0.30; P = 0.07), or peak GH during the GHRH-arginine test (P = 0.23) was observed.


   Discussion
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
References
 
The present study showed that the GHRH-arginine test cannot be used solely for diagnosing GHD in former childhood ALL subjects treated with CRT. Using 7.5 µg/liter as the cutoff yielded high specificity (94%), but, at the same time, the sensitivity was as low as 66%. In contrast, a failed GH response to the GHRH-arginine test accurately reflects the presence of radiation-induced GHD, illustrated by a high positive predictive (95%) value of the test when 7.5 µg/liter was used as the cutoff.

In a previous study of patients irradiated in childhood and adulthood, the sensitivity (50%) and specificity (94%) observed when using a cut-off level of 9 µg/liter during the GHRH-arginine test was comparable with the results from the present study (14). Biller et al. (23) suggested a much lower cutoff level of GH peak, e.g. less than 5 µg/liter during the GHRH-arginine test, which resulted in a specificity of 95% and a sensitivity of 91%, respectively. However, their results were based on patients and controls with a median BMI of 30 kg/m2, and the majority of the patients had a diagnosis of pituitary adenoma and were not treated with CRT. Using less than 5 µg/liter as the cutoff in the present study would yield similar specificity (96%), but much lower sensitivity (41%).

The most alarming effect of the sole use of the GHRH-arginine test in the clinical setting was the low negative predictive value, which means that the test cannot be used to rule out GHD. This is probably also true for other patient groups subjected to CRT, e.g. in patients radiated for brain tumors, as shown in the previous study by Darzy et al. (14). In contrast, a failed GH response to the GHRH-arginine test would in many situations make a complementary ITT unnecessary, because the GHRH-arginine test accurately reflects the presence of radiation-induced GHD (14), which is also illustrated by the high positive predictive value of the test observed in the present study. Among patients with a high suspicion of GHD after CRT, we therefore generally recommend a two-step procedure, starting with the GHRH-arginine test and followed by ITT, if a GH response, e.g. at least 7.5 µg/liter, is observed during the GHRH-arginine test.

In the present study we generally noticed a blunt GH response to the ITT when there was a rather clear peak in GH during the GHRH-arginine test. Damage of the lateral hypothalamus by CRT could explain this phenomenon, because this area is thought to contain neurons responsible for promoting GH secretion after stimulation by hypoglycemia (29, 30, 31). Also, in children with GHD and involvement of the hypothalamic area by, for instance, idiopathic inflammatory pituitary stalk thickness or Langerhans cell histiocytosis, a greater GH response to the GHRH-arginine test was recorded than in patients with GHD due to lesions of both hypothalamus and pituitary as in a craniopharyngioma (32).

In the multivariate regression analysis, only age at CRT and BMI remained significant predictors of the GHRH-arginine test among the patients. Each unit of BMI above 20 kg/m2 can be expected to decrease the peak GH level by 10%. Previously it was shown that BMI had no significant impact on the peak GH during GHRH-arginine stimulation among patients (23), but did have an effect in healthy males with BMI greater than 25 kg/m2 (24). In our study, the GH peak during the GHRH-arginine test also correlated negatively with BMI in the control group, comprising both males and females with a wide range in BMI. Almost half of the controls with BMI of 25 kg/m2 or higher would have been classified as GHD in the GHRH-arginine test with 9 µg/liter as the cutoff, which additionally stresses the inaccuracy of the GHRH-arginine test even for mildly elevated BMI values (24). Unfortunately, the number of patients in the present study was too small to allow an evaluation of different GHD cutoff levels depending on BMI for the GHRH-arginine test. Establishing appropriate GH cutoffs for subjects with elevated BMI would require large population studies (24). Permanent GHD due to organic hypothalamic-pituitary disease may indeed be difficult to distinguish from reversible blunting of GH secretion in obesity not only in the GHRH-arginine test, but also in the ITT (23). Recently, a correlation between peak GH during ITT and BMI was shown among healthy controls (33). BMI did not correlate with peak GH during ITT in our patients. To date, no study has been able to show a correlation between peak GH during ITT and BMI among patients with a narrow range of peak GH and a high proportion of truly GHD (23). Thus, among patients, the ITT seems less sensitive than the GHRH-arginine test to the effect of obesity on the GH peak, and we would therefore recommend using ITT to confirm GHD in such cases.

Age at CRT was also significant in the multivariate analyses in the present study, and each extra year of age above zero at CRT can be expected to increase the peak GH of the GHRH-arginine test by 10%. Thus, a low peak GH during a GHRH-arginine test together with a low age at CRT indicates a high likelihood of true GHD. This conclusion is also based on the parallel correlation between peak GH during ITT and age at CRT (P = 0.05) in the present study. This is in contrast with the report by Darzy et al. (14) that showed no correlation between age at CRT and peak GH during the GHRH-arginine test, which might be explained by the fact that they included a mixed group of childhood- and adult-onset GHD, and among the childhood-onset patients, the median age at CRT was 9 yr compared with 4 yr in the present study. In addition, our findings are in accordance with those presented by Kirk et al. (5), who investigated 77 ALL children 3–9 yr after diagnosis and showed that children with a younger age at diagnosis had more severe growth failure. It is well known that the brain is more vulnerable to CRT at a younger age, and it has been shown that the development of cortical gray matter peaks at approximately 4 yr of age, and cortical white matter volume continuously increases until age 20 yr (34). It is also of interest that of all the tested explanatory variables in the present study, the only correlation with peak GH during the ITT was a weak positive correlation to age at treatment (P = 0.05).

We recorded no correlation between the peak GH during the GHRH-arginine test and time since CRT, which is in contrast to the report by Darzy et al. (14), which showed a significant negative correlation. This is a new finding in adults subjected to CRT in childhood, and it probably means that when CRT is given at such an early age as in the present study population (median, 4 yr), this has much stronger impact on hormone secretion than time since CRT. We recorded no significant effect of gender on the peak GH during the GHRH-arginine test, which is in accordance with the findings of Darzy et al. (14).

As shown previously (14), no correlation was recorded between serum IGF-I and the peak GH response to the GHRH-arginine test in the present study population. This is in agreement with Brennan et al. (18), who showed that measurements of serum IGF-I or IGFBP-3 levels were poor markers for GHD in childhood ALL patients treated with CRT.

In a previous study there was an association between PRL and GHD (35), which is in contrast to our multivariate analysis that did not show any remaining effect of PRL on peak GH when controlling for age since treatment and BMI. In the previous study (34), the definition of PRL deficiency was PRL levels below the limit of detection, and no such low levels were recorded in the present study. There was no correlation between a median target dose of 24 Gy and peak GH during the GHRH-arginine test in the present study, which is in contrast to the findings of a previous study (14) where the range of the CRT dose was much broader, which may explain the differing results. Brennan et al. (18) recorded that ALL patients treated with lower doses of CRT of 18 Gy, were considered GH insufficient or even normal after ITT or arginine tests, but these patients were also older at CRT and thus had a shorter time since CRT.

In conclusion, the present study showed that the GHRH-arginine test cannot be used solely for diagnosing GHD in former childhood ALL subjects treated with CRT. Using 7.5 µg/liter as the cutoff yielded high specificity (94%), but at the same time, the sensitivity was as low as 66%. In contrast, a failed GH response to the GHRH-arginine test would in many situations make a complementary ITT unnecessary, because the GHRH-arginine test accurately reflects the presence of radiation-induced GHD, as illustrated by the high positive predictive value (95% at 7.5 µg/liter) of the test reported in the present study. Among patients with a high suspicion of GHD after CRT, we generally recommend a two-step procedure, starting with the GHRH-arginine test and followed by ITT if a GH response is observed during the GHRH-arginine test. Complementary ITT is also warranted to confirm GHD in obese patients.


   Footnotes
 
This work was supported by the Swedish Research Council (Grant K1999-27X013074-01A), the Medical Faculty of Lund University, the Swedish Children’s Cancer Foundation, and Eli Lilly & Co.

First Published Online August 30, 2005

Abbreviations: ALL, Acute lymphoblastic leukemia; BMI, body mass index; CI, confidence interval; CRT, cranial irradiation; CV, coefficient of variation; GHD, GH deficiency; ITT, insulin tolerance test; MTX, methotrexate; PRL, prolactin; ROC, receiver operating characteristics.

Received February 14, 2005.

Accepted August 15, 2005.


   References
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
 References
 

  1. Littley MD, Shalet SM, Beardwell CG, Ahmed SR, Applegate G, Sutton ML 1988 Hypopituitarism following external radiotherapy for pituitary tumors in adults. Q J Med 262:145–160
  2. Lam KS, Tse VK, Wang C, Yeung RT, Ma JT, Ho JH 1987 Early effects of cranial irradiation on hypothalamic-pituitary function. J Clin Endocrinol Metab 64:418–424[Abstract]
  3. Chrousos GP, Poplack D, Brown T, O’Neill D, Schwade J, Bercu BB 1982 Effects of cranial radiation on hypothalamic-adenohypophyseal function: abnormal growth hormone secretory dynamics. J Clin Endocrinol Metab 54:1135–1139[Abstract]
  4. Constine LS, Woolf PD, Cann D, Mick G, McCormick K, Raubertas RF, Rubin P 1993 Hypothalamic-pituitary dysfunction after radiation for brain tumors. N Engl J Med 328:87–94[Abstract/Free Full Text]
  5. Kirk JA, Raghupathy P, Stevens MM, Cowell CT, Menser MA, Bergin M, Tink A, Vines RH, Silink M 1987 Growth failure and growth-hormone deficiency after treatment for acute lymphoblastic leukaemia. Lancet 1:190–193[CrossRef][Medline]
  6. Growth Hormone Research Society 1998 Consensus guidelines for the diagnosis and treatment of adults with growth hormone deficiency: summary statement of the growth hormone research society workshop on adult growth hormone deficiency. J Clin Endocrinol Metab 83:379–381[Abstract/Free Full Text]
  7. Ghigo E, Aimaretti G, Arvatr E, Camanni F 2001 Growth hormone-releasing hormone combined with Arginine or Growth hormone secretagogues for the diagnosis of growth hormone deficiency in adults. Endocrine 15:29–38[CrossRef][Medline]
  8. Ghigo E, Gofi S, Nicolosi M, Arvat E, Valente V, Mazza E, Ghigo MC, Manni F 1990 Growth hormone responsiveness to combined administration of arginine and GH-releasing hormone does not vary with age in man. J Clin Endocrinol Metab 71:1481–1485[Abstract]
  9. Valetto MR, Bellone J, Baffoni C, Savio P, Aimaretti G, Gianotti L, Arvat E, Camanni F, Ghigo E 1996 Reproducibility of the growth hormone response to stimulation with growth hormone-releasing hormone plus arginine during lifespan. Eur J Endocrinol 135:568–572[Abstract]
  10. Ghigo E, Aimaretti G, Gianotti L, Bellone J, Arvat E, Camanni F 1996 New approach to the diagnosis of growth hormone deficiency in adults. Eur J Endocrinol 134:352–356[Abstract]
  11. Aimaretti G, Corneli G, Razzore P, Bellone S, Baffoni C, Arvat, E, Camanni F, Ghigo E 1998 Comparison between insulin-induced hypoglycemia and growth hormone (GH)-releasing hormone and arginine as provocative tests for the diagnosis of GH deficiency in adults. J Clin Endocrinol Metab 83:1615–1618[Abstract/Free Full Text]
  12. Ahmed SR, Shalet SM 1984 Hypothalamic growth hormone releasing factor deficiency following cranial irradiation. Clin Endocrinol (Oxf) 21:483–488[Medline]
  13. Grossman A, Lytras N, Savage MO, Wass JA, Coy DH, Rees LH, Jones AE, Besser GM 1984 Growth hormone releasing factor: comparison of two analogues and demonstration of hypothalamic defect in growth hormone release after radiotherapy. Br Med J 288:1785–1787
  14. Darzy KH, Aimaretti G, Wieringa G, Rao Gattamaneni H, Ghigo E, Shalet SM 2003 The usefulness of the combined growth hormone (GH)-releasing hormone and arginine stimulation test in the diagnosis of radiation-induced GH deficiency is dependent on the post-irradiation time interval. J Clin Endocrinol Metab 88:95–102[Abstract/Free Full Text]
  15. 2004 Stat bite: incidence of selected childhood cancers by age group. J Natl Cancer Inst 96:1809
  16. Chessells JM, Bailey C, Richards SM 1995 Intensification of treatment and survival in all children with lymphoblastic leukaemia: results of UK Medical Research Council trial UKALL X. Lancet 345:143–148[CrossRef][Medline]
  17. Moëll C, Garwicz S, Marky I, Mellander L, Karlberg J 1988 Growth in children treated for acute lymphoblastic leukemia with and without prophylactic cranial irradiation. Acta Paediatr Scand 77:688–692[Medline]
  18. Brennan BMD, Rahim A, Mackie EM, Eden OB, Shalet SM 1998 Growth hormone status in adults treated for acute lymphoblastic leukaemia in childhood. Clin Endocrinol (Oxf) 48:777–783[CrossRef][Medline]
  19. Hoffman DM, O’Sullivan AJ, Baxter RC, Ho KK 1994 Diagnosis of growth-hormone deficiency in adults. Lancet 343:1064–1068[CrossRef][Medline]
  20. Drake WM, Howell SJ, Monson JP, Shalet SM 2001 Optimizing GH therapy in adults and children. Endocr Rev 22:425–450[Abstract/Free Full Text]
  21. Tillmann V, Buckler JM, Kibirige MS, Price DA, Shalet SM, Wales JK, Addison MG, Gill MS, Whatmore AJ, Clayton PE 1997 Biochemical tests in the diagnosis of childhood growth hormone deficiency. J Clin Endocrinol Metab 82:531–535[Abstract/Free Full Text]
  22. Link K, Moell C, Garwicz S, Cavallin-Stahl E, Bjork J, Thilen U, Ahren B, Erfurth EM 2004 Growth hormone deficiency predicts cardiovascular risk in young adults treated for acute lymphoblastic leukemia in childhood. J Clin Endocrinol Metab 89:5003–5012[Abstract/Free Full Text]
  23. Biller BMK, Samuels MH, Zagar A, Cook DM, Arafah BM, Bonert V, Stavrou S, Kleinberg DL, Chipman JJ, Hartman ML 2002 Sensitivity and specificity of six tests for the diagnosis of adult GH deficiency. J Clin Endocrinol Metab 87:2067–2079[Abstract/Free Full Text]
  24. Bonert VS, Elashoff JD, Barnett P, Melmed S 2004 Body mass index determines evoked growth hormone (GH) responsiveness in normal healthy male subjects: diagnostic caveat for adult GH deficiency. J Clin Endocrinol Metab 89:3397–3401[Abstract/Free Full Text]
  25. Gustafsson G, Kreuger A, Dohlwitz A 1981 Acute lymphoblastic leukemia in Swedish children 1973–1978. Acta Paediatr Scand 70:609–614[Medline]
  26. Gustafsson G, Garwicz S, Hertz H, Johanesson G, Jonmundsson G, Moe PJ, Salmi T, Seip M, Siimes MA, Yssing M, Åhström L 1987 A population-based study of childhood acute lymphoblastic leukemia diagnosed from July 1981 through June 1985 in the five Nordic countries. Acta Paediatr Scand 76:781–788[Medline]
  27. Gustafsson G, Kreuger A, Clausen N, Garwicz S, Kristinson J, Lie SO, Moe PJ, Perkkiö M, Yssing M, Saarinen-Pihkala UM 1998 Intensified treatment of acute childhood lymphoblastic leukaemia has improved prognosis, especially in non-high risk patients: the Nordic experience of 2.648 patients diagnosed between 1981 and 1996. Acta Paediatr 87:1151–1161[CrossRef][Medline]
  28. Hanley JA, McNeil BJ 1982 The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143:29–36[Abstract]
  29. Abrams RL, Parker ML, Blanco S, Reichlin S, Daughaday WH 1966 Hypothalamic regulation of growth hormone secretion. Endocrinology 78:605
  30. Himsworth RL, Carmel PW, Frantz AG 1972 The location of the chemoreceptor controlling growth hormone secretion during hypoglycemia in primates. Endocrinology 91:217
  31. Krey LC, Lu KH, Butler WR, Hotchkiss J, Piva F, Knobil E 1975 Surgical disconnection of the medial basal hypothalamus and pituitary function in the rhesus monkey. II. GH and cortisol secretion. Endocrinology 96:1088
  32. Maghnie M, Cavigioli F, Tinelli C, Autelli M, Arico M, Aimaretti G, Ghigo E 2002 GHRH plus arginine in the diagnosis of acquired GH deficiency of childhood-onset. J Clin Endocrinol Metab 87:2740–2744[Abstract/Free Full Text]
  33. Qu XD, Gaw Gonzalo IT, Al Sayed MY, Cohan P, Christenson PD, Swerdloff RS, Kelly DF, Wang C 2005 Influence of body mass index and gender on growth hormone (GH) responses to GH-releasing hormone plus arginine and insulin tolerance tests. J Clin Endocrinol Metab 90:1563–1569[Abstract/Free Full Text]
  34. Pfefferbaum A, Mathalon DH, Sullivan EV, Rawles JM, Zipursky RB, Lim KO 1994 A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch Neurol 51:874–887[Abstract]
  35. Mukherjee A, Murray RD, Columb B, Gleeson HK, Shalet SM 2003 Acquired prolactin deficiency indicates hypopituitarism in patients with disease of the hypothalamic-pituitary axis. Clin Endocrinol (Oxf) 59:743–748[CrossRef][Medline]



This article has been cited by other articles:


Home page
 Eur J EndocrinolHome page
K. Chihara, A. Shimatsu, N. Hizuka, T. Tanaka, Y. Seino, Y. Katofor, and the KP-102 Study Group
A simple diagnostic test using GH-releasing peptide-2 in adult GH deficiency
Eur. J. Endocrinol., July 1, 2007; 157(1): 19 - 27.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
90/11/6048    most recent
Author Manuscript (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Björk, J.
Right arrow Articles by Erfurth, E. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Björk, J.
Right arrow Articles by Erfurth, E. M.
Related Collections
Right arrow Endocrine Oncology
Right arrow Neuroendocrinology and Pituitary
Right arrow Pediatric Endocrinology

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology Endocrine Reviews J. Clin. End. & Metab.
Molecular Endocrinology Recent Prog. Horm. Res. All Endocrine Journals