| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Medicine, Mount Sinai Hospital (S.B., S.E.) and St. Michael’s Hospital (G.L.B.); Department of Pathology, University Health Network (S.L.A.) and St. Michael’s Hospital (K.K.); Department of Laboratory Medicine and Pathobiology (S.L.A., K.K.) and Department of Medicine, Ontario Cancer Institute and the Freeman Centre for Endocrine Oncology (S.B., S.L.A., S.E.), Toronto, Ontario, Canada M5G-1X5
Address all correspondence and requests for reprints to: Dr. Shereen Ezzat, Mount Sinai Hospital, 600 University Avenue, Room 437, Toronto, Ontario, Canada M5G 1X5. E-mail: sezzat{at}mtsinai.on.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to examine the significance of somatotroph adenoma type on response to SSA.
Design: This study was a retrospective examination of postoperatively treated acromegalic patients with the SSA octreotide.
Setting: The study was performed at a university-affiliated tertiary care center.
Patients: Forty patients with acromegaly were studied.
Main Outcome Measures: Normalization of IGF-I levels and GH responses were the main outcome measures.
Results: Univariate analysis revealed that responders were more likely to have densely granulated somatotroph adenomas (80% vs. 43.8%; P = 0.024), to be older (51.3 vs. 38.2 yr; P < 0.003), to have smaller tumors (stage 
3; 78.6% vs. 35.7%; P = 0.022), to have lower baseline IGF-I (453 vs. 716 µg/liter; P < 0.001) and GH levels (2.7 vs. 7.8 µg/liter; P < 0.05), and to require a lower maximum dose of SSA (24 vs. 31 mg every 4 wk; P = 0.013). Multivariate analysis confirmed that a densely granulated adenoma was the strongest predictor of complete response [adjusted odds ratio (OR), 58.41; 95% confidence interval (CI), 1.24–1000.00; P = 0.04] compared with other covariates, including older age at time of diagnosis (OR, 1.15/yr; 95% CI, 1.01–1.31; P = 0.03), and tumor stage of 3 or less (OR, 29.77; 95% CI, 1.01–885.45; P < 0.05).
Conclusions: Somatotroph tumor type represents a strong clinical predictor of response to SSA treatment and will help to identify patients who warrant more vigilant management of their disease.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recent studies have convincingly demonstrated that normalization of GH or IGF-I concentrations results in both a significant improvement in comorbidities (6, 7, 8) and a reduction in mortality to that of the general population (9, 10, 11). The development of selective pharmacological agents over the last decade and of somatostatin analogs (SSA) in particular has greatly advanced the treatment of acromegaly. Given the availability of these agents and the benefits of treatment, an aggressive therapeutic approach is necessary and warranted in these patients.
Therapeutic options for patients with acromegaly include surgical resection, primary or adjunctive medical therapy, and radiation therapy. Surgical treatment alone has been found to be curative in less than 60% of patients (5, 12). Similarly, primary therapy with octreotide has been shown to normalize IGF-I in a similar proportion and renders GH levels below 2.5 µg/liter in approximately 80% of patients. However, pituitary tumor size reduction with SSA is limited to a smaller subset of about only 40% of subjects (13, 14). A recent review of SSA treatment in acromegaly suggests that the effect appears to be related to the subject’s prior treatment, with approximately 57% and 28% of patients demonstrating tumor size shrinkage when this medical treatment was administered as primary and adjunctive therapies, respectively (15). Furthermore, when it does occur, tumor size reduction is usually of a modest degree. Therefore, most patients require surgical resection of their pituitary tumor, followed by adjunctive medical therapy with SSA for optimal biochemical control.
Despite aggressive therapy, however, response rates to SSA postoperatively are variable and are thought to depend mainly on somatostatin receptor subtype expression by the adenoma (16). Other possible predictive factors of responsiveness to SSA have been examined, including patient age, tumor size, and baseline IGF-I or GH levels (17). However, the various pituitary pathologies underlying acromegaly have not been extensively examined for their impact on SSA responsiveness (18). One previous study demonstrated that the GH inhibitory preoperative effects of octreotide were significantly greater in patients harboring densely granulated somatotroph adenomas than in those harboring sparsely granulated adenomas (19). We examined the significance of the various clinical predictive factors in the context of pathological classification of the somatotroph adenoma type and compared it with other putative predictors of response to octreotide treatment in acromegalic patients who have undergone previous transsphenoidal pituitary surgery.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We conducted a 10-yr retrospective analysis of patients with acromegaly at a single center (University of Toronto, Toronto, Canada) between February 1994 and September 2004. Patients treated with transsphenoidal pituitary surgery and adjunctively with the SSA octreotide for a minimum of 4 months were included in our analysis. Among 67 consecutive patients with acromegaly who were seen at our center during the study period, 40 were eligible for inclusion in our cohort. Patients not requiring SSA after surgery (n = 12), those treated with SSA alone (n = 6), or those requiring the addition of a dopamine agonist to SSA (n = 9) were not included in this study. All included patients were treated postoperatively with octreotide (Sandostatin-LAR, Novartis Pharmaceuticals, Montreal, Canada; doses ranging from 20 mg, im, every 6 wk to 30 mg, im, every 4 wk). Adjustments to the treatment dose were based on the aim of targeting age-adjusted normal IGF-I and GH levels as described below. Among those included in the analysis, nine patients had received radiation therapy (ranging from 6 months to 12 yr after surgery) before study inclusion.
During follow-up visits, patients were assessed clinically every 4 months and had serum fasting GH (Quest Diagnostics, San Juan Capistrano, CA) and IGF-I (Diagnostic Systems Laboratories, Inc., Webster, TX) levels evaluated at each visit. Patients with persistently elevated GH (>1.0 µg/liter) and/or age-adjusted IGF-I levels also underwent a 75-g oral glucose tolerance test (OGTT).
Response assignment
Patients in our cohort (n = 40) were followed for a mean of 28 months (range, 4–96 months). No patients were lost to follow-up or died during the observation period.
Patients achieving normalization of age-adjusted IGF-I levels after treatment with SSA were classified as responders; nonresponders maintained levels above this range for up to 7 yr. Nonresponders were additionally categorized into those who reached an IGF-I level within 25% of the upper limits of their normal values (partial responders); those who had an IGF-I level exceeding this range were defined as resistant. All patients who had a discrepancy between GH and IGF-I (normal IGF-I with a random GH >1.0 µg/liter) underwent a 75-g OGTT and were only classified as responders if the GH suppressed to below 1.0 µg/liter.
Patient variables
Tumor pathology served as our primary variable of interest. The pathology of the pituitary adenomas resected was examined using immunohistochemistry for GH and keratin for all cases, and tumors were classified according to the accepted Armed Forces Institute of Pathology and World Health Organization criteria (18, 20). This classification was also corroborated by electron microscopy. All patients were classified as having either a sparsely or a densely granulated somatotroph adenoma, and one patient with an acidophil stem cell adenoma on pathology was classified as "other" (see Table 2
).
|
Statistical analyses
The statistical program SPSS (SPSS, Inc., Chicago, IL) was used. Continuous baseline data are presented as the mean ± SD, and binomial variables are presented as the absolute number and percent prevalence in the population. Results were analyzed for statistically significant differences using an independent t test or 
2 test where appropriate. In addition, ANOVA was conducted to compare baseline characteristics across the following three groups: responders, partial responders, and resistant cases. Logistic regression was used to examine the effect of pituitary pathology (densely vs. sparsely granulated/other) on the likelihood of achieving a complete response, after adjusting for potential confounders. The dependent binomial variable was defined as response or no response in terms of IGF-I normalization. Independent variables used in the multivariate analysis included all baseline variables (as indicated above) that were related to IGF-I normalization on univariate analysis with a P < 0.1 For logistic regression, no data were missing for all included patients (n = 40) for age at diagnosis, maximum dose, and baseline IGF-I. Pituitary pathology assignment was not possible for technical reasons in four cases (three in the nonresponder group and one in the responder group). We assigned those four cases to the densely granulated group to avoid skewing the data in favor of our hypothesis that densely granulated pathology predicts response to SSA, given that most of the missing data were in the nonresponder group. The likelihood of complete response associated with each covariant was expressed as an adjusted odds ratio (OR) ± 95% confidence interval (CI).
Finally, Kaplan-Meier survival curves were constructed to reflect the time to a complete response between patients with densely vs. sparsely granulated pituitary lesions. Time was recorded as months in the study on SSA treatment. Patients were censored if they achieved an IGF-I level normal for age and gender or if the patient’s IGF-I level did not normalize, but they required other adjunctive treatment (repeat pituitary surgery, radiation therapy, addition of dopamine agonist, or addition of GH receptor antagonist) subsequent to treatment with SSA. No patient died or was lost to follow-up during the study time. A two-tailed P < 0.05 was considered significant.
The study protocol was approved by the research ethics board at Mount Sinai and St. Michael’s Hospitals, University of Toronto.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The baseline characteristics of the patients are shown in Table 1. Among the 40 patients, there was a relatively equal distribution of males and females (45% vs. 55%). The mean age of the cohort at the time of the study was 53.1 yr (range, 19–76), an average of 8 yr from the time of diagnosis. A comparison of potential confounders influencing the response to SSA treatment in the two study groups (responders and nonresponders) is summarized in Table 2
. Twenty-one (53%) were classified as complete responders, and 19 (48%) as nonresponders. Among the nonresponders, six of 19 (31.6%) were classified as partial responders, and 13 of 19 (68.4%) were classified as resistant.
|
In comparing the two groups of responders vs. nonresponders, responders were older at the time of diagnosis (51.3 vs. 38.2 yr; P < 0.003), were significantly more likely to have a densely granulated somatotroph adenoma (80% vs. 43.8%; P = 0.024), or tumor stage of 3 or lower (78.6% vs. 35.7%; P = 0.022), had lower IGF-I levels (453 vs. 716 µg/liter; P < 0.001) and lower GH levels (2.7 vs. 7.8 µg/liter; P < 0.05), and required a lower maximum dose of SSA (24 vs. 31 mg every 4 wk; P = 0.013; Table 2). When comparing across all three groups, (responder, partial responder, and resistant), there was a significant difference in mean age (P < 0.001), the proportion with a tumor stage 3 or lower (P < 0.01), mean pretreatment IGF-I levels (P < 0.001), and the maximum dose of SSA required (P = 0.008).
In general, patients who responded to SSA demonstrated an improvement in glycemic status. For instance, 20% (one of five) of responders with baseline DM achieved normalization of fasting blood sugars, and 50% (one of two) of partial responders with DM at baseline converted from diabetes back to impaired fasting glucose. In contrast, none of the patients resistant to treatment with SSA demonstrated improvement in their diabetic status (Fig. 1). None of the patients in the responder or partial responder groups had progression of diabetic status, whereas two patients in the resistant group did: one progressed from impaired glucose tolerance to DM, and another converted from normoglycemia to impaired glucose tolerance during follow-up. There was no clinically identifiable change in HTN or OSA status in any of the groups (Fig. 1).
|
Multivariate analyses
On multivariate analysis, the presence of a densely granulated lesion was the strongest predictor of a complete response (adjusted OR, 58.4; 95% CI, 1.24–1000.00; P = 0.04) after adjusting for other significant baseline predictors (Table 3). Other significant covariates included a higher age at the time of diagnosis (OR, 1.15/yr of age; 95% CI, 1.01–1.31; P = 0.03), tumor stage of 3 or lower (OR, 29.8; 95% CI, 1.001–885.4; P < 0.05), and lower pretreatment IGF-I levels (OR, 0.93/10 µg/liter rise in IGF-I; 95% CI, 0.87–0.99; P = 0.04). Although higher baseline GH levels were associated with a lower odds of responding to SSA on univariate analysis (OR, 0.75/10 µg/liter; 95% CI, 0.58–0.98; P = 0.04), it was no longer a statistically significant predictor of response once tumor pathology, stage, pretreatment IGF-I levels, and age at diagnosis were included in the model. In addition, treatment dose was no longer significantly related to treatment response when other variables were included in the model (OR, 0.83; 95% CI, 0.62–1.10; for the 20- vs. 30-mg dose SSA).
|
. No overall difference in the two groups was noted using the log-rank statistic (P = 0.460). However, at 48 months there was a trend toward a significant difference between the two groups (P = 0.06), suggesting an increased success of SSA treatment in those patients with densely granulated adenomas compared with that in patients with sparsely granulated adenomas.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Over the years, the definition of a biochemical cure of acromegaly has continued to change due to the availability of improved GH and IGF-I serum assays and additional evidence of increased survival in patients with normalization of these values (10). Before 1980, a posttreatment GH level below 5 µg/liter was believed to represent disease cure (9); by 1990, this was lowered to less than 2.5 µg/liter. A recent study, however, demonstrates that normalization of GH to below 2.0 µg/liter is not sufficient, and that reduction of mortality rates to that of the general population is achieved by normalization of GH to less than 1.0 µg/liter and reduction of IGF-I to the normal range (10). Current guidelines and those used to describe cure in this study define biochemical cure as a suppressible GH of less than 1 µg/liter after an OGTT and normalization of IGF-I (1). As a result of the more stringent criteria for cure, increasing numbers of patients are requiring adjunctive therapy after surgical resection, including treatments with SSAs, dopamine agonists, and/or radiation.
Because SSAs are the primary form of adjunctive treatment for persistent disease in acromegaly, it is crucial to examine predictors of response to this therapy among patients with acromegaly using current definitions of cure as outlined above. In this series, pathological classification of the adenoma as a densely granulated somatotroph variant was determined to be the strongest predictor of disease control in patients requiring adjunctive treatment with a SSA after surgery. This finding is consistent with the results of a previous study demonstrating that densely granulated somatotroph adenomas have increased preoperative responsiveness to octreotide compared with sparsely granulated adenomas (19). It is also consistent with the identified pathophysiology of densely granulated adenomas that are known to harbor activating mutations of the Gs
protein resulting in constitutive activation of adenylate cyclase (22, 23, 24). This pathway is amenable to suppression through somatostatin receptor activation. In contrast, the pathogenesis of sparsely granulated somatotroph adenomas is not known (25). Moreover, the relationship between the sparsely granulated adenoma phenotype and somatostatin receptor subtype expression remains to be determined and should be the subject of future investigation. In a recent retrospective study of 57 patients with acromegaly treated primarily with transsphenoidal surgery, similar rates of remission, using modern remission criteria, as outlined above, were noted in those with densely or sparsely granulated adenomas on pathology (26). In that study, however, there was a high rate of remission postoperatively, with only nine patients requiring adjunctive medical therapy with SSA. Thus, our population probably represents a cohort with more advanced disease, because only those requiring adjunctive medical therapy were enrolled. Interestingly, in that same series, sparsely granulated lesions had a higher rate of dural invasion (
50%) and a lower preoperative GH level (26).
After tumor pathology, a preoperative tumor stage below 3 was determined to be the next strongest predictor of response to treatment with SSA after transsphenoidal surgery. This observation is in keeping with a previous study of a series of patients with acromegaly in whom higher tumor stage, tumor grade, and preoperative GH levels were all predictive of disease persistence in patients with acromegaly after transsphenoidal pituitary surgery (9). In that study, however, a GH level less than 5 µg/liter was used to define cure, and only 12% of patients received adjunctive postoperative therapy with SSAs for persistent disease. Thus, our series provides additional information about the importance of tumor stage in acromegaly, because it predicts not only the response to initial transsphenoidal pituitary surgery but also that to adjunctive treatment with SSAs using the current definition of disease control.
In our analysis, tumor pathology was the strongest predictor of IGF-I normalization in patients with acromegaly receiving postoperative SSA therapy. Because of the wide CIs associated with our estimates, the relative importance of tumor pathology vs. tumor stage is uncertain; however, we can confidently conclude that both are significantly related to the subsequent response to treatment.
Finally, older age at the time of diagnosis was determined to be a predictor of complete response to treatment with SSA after transsphenoidal surgery. Our data demonstrated a significant difference of 13 yr in the mean age of responders compared with nonresponders. Previous studies have also shown a correlation between younger age and higher preoperative GH levels with disease persistence after surgery (27). However, in a recent series examining predictors of mortality in a cohort of 208 acromegalic patients, older age at the time of diagnosis was determined to be a significant predictor of mortality (10). This study, however, predated the use of routine SSAs in New Zealand, and therefore, subjects in this analysis did not receive this adjunctive therapy.
It should be noted that although 23% of the subjects (n = 9) had received previous pituitary radiation, twice as many patients in the nonresponder group (31.6%) underwent radiation therapy compared with the responder group (14.3%). Because previous radiation therapy would increase the likelihood of achieving normalization of IGF-I and GH levels, this would not have affected the response rate seen in the two groups, because fewer responders received radiation therapy.
To date, this is the first systematic analysis of predictors of responsiveness to adjunctive postoperative SSA therapy using the updated guidelines (GH <1.0 µg/liter and normalization of IGF-I) to define biochemical cure with respect to the pathological features of the pituitary adenoma type. The significance of tumor classification as a clinical predictor of response to pharmacotherapy with SSAs emphasizes the importance of accurate morphologic classification using immunohistochemistry and/or electron microscopy.
|
Footnotes |
|---|
Abbreviations: CI, Confidence interval; DM, diabetes mellitus; FPG, fasting plasma glucose; HTN, hypertension; OGTT, oral glucose tolerance test; OR, odds ratio; OSA, obstructive sleep apnea; SSA, somatostatin analog.
Received May 5, 2005.
Accepted August 15, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-chain of Gs and stimulate adenylate cyclase in human pituitary tumors. Nature 340:692–696[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  C. A. Leontiou, M. Gueorguiev, J. van der Spuy, R. Quinton, F. Lolli, S. Hassan, H. S. Chahal, S. C. Igreja, S. Jordan, J. Rowe, et al. The Role of the Aryl Hydrocarbon Receptor-Interacting Protein Gene in Familial and Sporadic Pituitary Adenomas J. Clin. Endocrinol. Metab., June 1, 2008; 93(6): 2390 - 2401. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. L. Asa, R. DiGiovanni, J. Jiang, M. L. Ward, K. Loesch, S. Yamada, T. Sano, K. Yoshimoto, S. J. Frank, and S. Ezzat A Growth Hormone Receptor Mutation Impairs Growth Hormone Autofeedback Signaling in Pituitary Tumors Cancer Res., August 1, 2007; 67(15): 7505 - 7511. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N Y Y Al-Brahim and S L Asa My approach to pathology of the pituitary gland J. Clin. Pathol., December 1, 2006; 59(12): 1245 - 1253. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
