| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Molecular and Cellular Endocrinology Branch, National Institute of Diabetes and Digestive and Kidney Diseases (F.B.-D., M.C.S., B.D.W.); National Institute of Neurological Disorders and Stroke (E.H.O.); and the Department of Radiology (J.L.D.), National Institutes of Health, Bethesda, Maryland 20892
Address all correspondence to: Françoise Brucker-Davis, M.D., National Institutes of Health, Building 10, Room 8S235-B, 10 Center Drive, MSC 1770, Bethesda, Maryland 20892-1770.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-subunit, and high
-subunit/TSH ratio) were present, respectively, in 10, 8, and 12
cases (sensitivity, 71%, 75%, and 83%; specificity, 96%, 90%, and
65%). In treated patients, the respective sensitivities of the TRH
test, -subunit, and -subunit/TSH ratio were 64%, 90%, and 90%,
and their specificities were 100%, 82%, and 73%. Studies of thyroid
hormone action revealed no evidence of acquired resistance to thyroid
hormone in TSH-secreting tumors. Apparent cure was achieved in 35% of
cases by surgery alone and in 22% more by combined therapies. Three
deaths occurred, including 1 from metastatic thyrotroph carcinoma. Six
patients had residual tumor, with symptoms of hyperthyroidism
controlled with octreotide in 5. The size and invasiveness of the
tumor, duration of symptoms, and intensity of hyperthyroidism were the
main prognostic factors. Thus, early diagnosis and treatment are the
keys to a good outcome. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The present study of 25 patients assessed and followed at the NIH from 1982–1996 describes their clinical and biochemical features and analyzes the diagnostic criteria and therapeutic outcome of these rare tumors. Moreover, as some patients with TSH-secreting tumors display only mild or no hyperthyroid symptoms, we also assess tissue responsiveness to thyroid hormone for comparison with patients with resistance to thyroid hormone and Graves’ disease.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
After giving informed consent, as approved by the NIDDK institutional review board, 25 patients with TSH-secreting tumor were studied at the NIH between 1982–1996. Partial data for 8 patients have been previously reported (9, 10, 11). The results of diagnostic testing and thyroid hormone action parameters in patients with TSH-secreting tumors were compared to the results obtained in 51 adult patients with resistance to thyroid hormone (including 34 with intact thyroid), 12 hyperthyroid adult patients with Graves’ disease, and 60 euthyroid adults, unaffected relatives of patients with resistance to thyroid hormone. Patients underwent a complete physical exam, including palpation of the thyroid and cardiovascular assessment. Information on hyperthyroid symptoms, headaches, visual problems, menstrual history, and galactorrhea were also recorded. Visual field testing (n = 24) was performed using computerized Goldmann’s perimetry.
Basal biochemical testing of serum
TSH was measured with a first generation RIA until 1985 (Solid Phase hTSH System, Beckman Coulter, Inc., Hialeah, FL), a second generation assay from 1985–1994 (MAIAclone, Serono Diagnostics, Walpole, MA), and a third generation assay since then (Access TSH Assay, Sanofi Diagnostics Pasteur, Inc., Chaska, MN).
Measurements of T4 (fluorescein polarization immunoassay, Abbott TDx, Abbott Laboratories, North Chicago, IL), T3 (Quanticoat, Kallesad Diagnostic, Chasco, MN), free T4 (Gammacoat two-step RIA, Incstar Corp., Stillwater, MN), free T3 (RIA, Becton Dickinson and Co. kit, SmithKline Beecham Laboratories, Van Nuys, CA), and T4-binding globulin (Corning’s Immunophase, TBG 125I, Corning Medical, Norwood, MA) were performed at the NIH Clinical Center.
-Subunit (RIA) and PRL (TOSOH AIA-1200) levels were measured by
Hazelton-Washington Laboratories (Vienna, VA).
Several parameters of thyroid hormone action classically assessed in resistance to thyroid hormone (6, 12) were studied. They included cholesterol, ferritin (Abbott Laboratories), carotene (SmithKline Beecham Clinical Laboratories), testosterone-binding globulin (Hazelton-Washington), and recently osteocalcin and angiotensin-converting enzyme (Mayo Medical Laboratories, Rochester, MN). Two nonbiochemical parameters were also studied. The basal metabolic rate (n = 14) was measured at Georgetown University Hospital (Washington DC) up to 1994 by using a Sensor Medics 2900 metabolic cart (Sensor Medics Corp., Yorba Linda, CA) and at the NIH since then (Deltatrac II, Sensor Medics). The right ankle reflex was measured with an achillometer (Polymed GmbH, Polymed Medical Center, Medizintechnik, Glattbugg, Switzerland) connected to a 1511B electrocardiograph (Hewlett-Packard Co., Waltram, MA) in nine patients.
Dynamic testing
TRH-stimulated TSH was measured in a standard 3-h test (after iv
injection of 500 µg Protirelin, Ferring Laboratory, Suffern,
NY) in 24 patients; PRL and -subunit levels were measured at 0 and
30 min.
An octreotide test, a modification of the Beck-Peccoz protocol (13),
was performed in 10 patients before surgery. TSH and -subunit were
measured before and after sc injection of octreotide (100 µg) given
three times daily over 72 h.
The T3 suppression test, proposed by Nicoloff and
Spencer (14), was performed in two patients before surgery and in seven
after surgery. Patients received T3 (Cytomel; 300
µg) at noon; TSH and -subunit levels were measured before and
48 h after treatment.
Imaging
In 19 patients, magnetic resonance imaging (MRI) of the pituitary gland was performed at 1.5 Tesla. T1-weighted coronal and sagittal images were performed before and after gadolinium injection (1.0 mmol/kg). Contrast-enhanced computerized tomography (CT) of the pituitary was performed in the 6 patients seen before 1986. In addition, an ultrasound of the thyroid was performed in 16 patients.
Treatment
Twenty-two patients underwent surgery at the NIH. Six had
pituitary surgery before referral to the NIH: three had transsphenoidal
surgery, two had a transfrontal approach, and one had both (Table 2a
).
In the absence of surgical cure, external radiation of the pituitary
was performed (n = 12); high dose focused irradiation (
-knife
therapy) was used in one patient. Palliative medical therapy, including
bromocriptine and octreotide, was given to patients with
persistent, clinically significant tumoral TSH secretion. Patients were
followed at the NIH twice the first year and then yearly. For two
patients who had not recently returned to the NIH, follow-up
information was obtained from their private physicians.
|
Apparent cure was defined as euthyroidism with a normalized TRH test and the absence of residual tumor on MRI. However, because these criteria are not suitable for patients who have postoperative hypopituitarism or who had a normal preoperative TRH test, we have more recently been testing the predictive value of the newly developed T3 suppression test, defining a normal response as a decline of TSH to less than 10% of the baseline (14).
Statistical analysis
Data were analyzed using StatView 4.2 software (Abacus Concepts,
Inc., Berkeley, CA). The t test was used for continuous
variables, and 
2 test was used for qualitative
variables. The level of significance was set at P =
0.01 to take into account the multiple comparisons.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The patients’ characteristics are shown in Tables 1, 2a
, and 2b. Patients
were, on the average, 44 yr old (range, 15–80 yr), and were
predominantly women; 9 were menopausal (however, only 5 had elevated
gonadotropin levels). The symptoms leading to diagnosis were usually
linked to thyrotoxicosis and had been present in most patients for many
years (range, 0–26 yr); thyrotoxic symptoms and signs were considered
severe in 14 patients, mild in 8, and absent in 3. The diagnosis was
made incidentally in 3 asymptomatic men (2 had systematic blood tests
for occupational purposes showing central hyperthyroidism and 1 had a
CT scan for mastoiditis revealing a pituitary macroadenoma) as well as
in 2 women presenting with amenorrhea-galactorrhea and in 1 with
acromegaly. Headaches were present in 39% of the patients, and visual
fields were impaired in 36%. Signs of other pituitary hormone
secretion were present in 4 patients (GH, 1 case; PRL, 3 cases); in 1
patient, cosecretion of LH and FSH was suggested, but not proven
(patient 1).
|
|
-subunit and the -subunit to
TSH molar ratio were elevated. There was no difference between the 2
groups in age, sex distribution, length of symptoms, size of tumor,
secretion of other pituitary hormones, goiter, or thyroid nodule
(P = 0.119). There was a trend, however, for
hyperthyroid symptoms to be milder in the untreated group and for the
visual fields to be less impaired (P = 0.05). The
previously treated group was characterized by lower thyroid hormone
levels associated with much higher TSH levels compared to the group
with intact thyroid. Radiological assessment
MRI and/or CT imaging revealed a macroadenoma (
1 cm maximum
diameter) in 23 patients and a microadenoma in 2 patients (Table 1).
The average maximal diameter of the adenoma was 23 mm (range, 7–60
mm); there was no difference in adenoma size whether patients had had
thyroid treatment or not. Invasion of cavernous sinus was common
(obvious in 12 cases and suspected in 6 other cases) as was expansion
into the sphenoid sinus or chiasmatic compression (respectively, in 15
cases and 10 cases). Only 5 patients had no radiographic evidence of
extrasellar extension. Figures 1 and 2 show the MRI imaging in 2 patients,
illustrating the differences in radiological features at
presentation.
|
|
Patients with untreated thyroid. The diagnostic value of
classical tests was compared in 14 untreated patients with
TSH-secreting tumor, 34 patients with resistance to thyroid hormone
(with similar elevation of thyroid hormones), and 60 normal subjects
(Table 3).
|
The average TSH stimulation after TRH challenge was much lower in patients with TSH-secreting tumor (P < 0.0001). Seven of 14 tumor patients had an absent response to TRH, defined as a TSH increment of less than 2 mU/L, vs. none in the group of patients with resistance to thyroid hormone. Three TSH-secreting tumor patients had a decreased response (<200% of baseline or <5 mU/L), and four had a normal response (>200% of baseline or >5 mU/L). One patient had a delayed (peak at 120 min) but quantitatively normal response.
Baseline -subunit tended to be higher in patients with TSH-secreting
tumor (range, 0.9–135 µg/L) than in controls (range, 0.5–4.9
µg/L) and patients with resistance to thyroid hormone (range,
0.5–6.9 µg/L). However, -subunit was normal in 6 of these 14
patients. The 10 TSH-secreting tumor patients with normal or low
gonadotropin levels had, as expected, lower -subunit than the 4 with
elevated gonadotropins (4.7 ± 1.4 vs. 45.3 ± 31
µg/L); importantly, their -subunit levels were higher than those
in premenopausal controls (1.1 ± 0.1 µg/L; n = 28;
P < 0.0001) and patients with resistance to thyroid
hormone (1.02 ± 0.07 µg/L; n = 18; P <
0.002).
The -subunit to TSH molar ratio was calculated using the following
formula (-subunit in micrograms per L divided by TSH in milliunits
per L) x 10 (15). There was still an overlap between patients
with TSH-secreting tumor (range, 3.5–241), resistance to thyroid
hormone (range, 1.2–36.3), and controls (range, 1.5–24.5). When
studying patients with normal or low gonadotropin levels, the ratio
tended to be higher in patients with TSH-secreting tumor (9.9 ±
2.8; n = 10) than in patients with resistance to thyroid hormone
(3.9 ± 0.5; n = 18; P = 0.01) and controls
(6.3 ± 0.6; n = 28; P = 0.07). Using the
criteria proposed by Beck-Peccoz et al. (16) (if TSH is
normal, ratio <5.7 in normogonadotropic patients and <29.1 in
hypergonadotropic patients; if TSH is elevated, ratio <0.7 in
normogonadotropic patients and <1.0 in hypergonadotropic patients), 2
of 14 TSH-secreting tumor patients had a normal ratio.
Patients with prior thyroid treatment. Prior thyroidectomy
resulted in major changes in the hormonal profile (Table 1). Baseline
TSH was elevated in all 11 cases including 8 receiving
T4 therapy (range, 6.7–127 mU/L), with free
T4 usually elevated but normal in 4 cases (range,
0.8–4.2 ng/dL).
The response to TRH challenge was considered normal in four cases
(stimulated TSH >200% of baseline), decreased in four cases
(120–200% of baseline), and flat in three cases (<120% of
baseline). In patients with normal or low gonadotropins (n = 9),
-subunit (13.6 ± 3 µg/L) was higher than in controls
(P < 0.0001), patients with resistance to thyroid
hormone (P < 0.0001), or TSH-secreting tumor with
intact thyroid (P = 0.02). Conversely, the
-subunit/TSH ratio (3.1 ± 0.7) tended to be lower than that in
controls or untreated TSH-secreting tumor patients, but was similar to
that in patients with resistance to thyroid hormone. However, in the
absence of normalized data for patients with primary hypothyroidism,
-subunit values (range, 2.2–26.4 µg/L) were difficult to
interpret, given the high TSH in these patients. The
-subunit/TSH ratio (range, 0.145–6.7) was normal in two
patients, using the criteria described earlier. Combining the
-subunit/TSH ratio and TRH test results permitted appropriate
diagnosis in all but one patient.
Diagnostic value of classical tests
Table 4 compares the respective
value of each test for the diagnosis of TSH-secreting tumor in the
presence of inappropriate secretion of TSH (14 tumors and 34 resistance
to thyroid hormone with intact thyroid and 11 tumors and 17 resistance
to thyroid hormone with prior thyroid treatment). In patients with
intact thyroid, the most sensitive test to identify a TSH-secreting
adenoma was an elevated -subunit/TSH ratio (83%), followed by an
elevated -subunit (75%), a flat or decreased response to TRH
(71%), and an elevated baseline TSH (43%). A flat or decreased
response to TRH (96%) was the most specific test for a TSH-secreting
adenoma, followed by elevated -subunit (90%), elevated baseline TSH
(88%), and elevated -subunit/TSH ratio (65%). The TSH response to
TRH had the best positive and negative predictive value. In patients
with prior thyroid treatment, the TRH test was less sensitive (64%),
but was highly specific (100%).
|
An octreotide test was performed in 10 patients, showing a suppressed TSH on day 4 to below 12% of baseline in 8 patients; 1 patient partially suppressed to 53% of baseline, and another had no response. In contrast, patients with resistance to thyroid hormone showed only slight to moderate decreases (12) (our personal results, not shown).
The short T3 suppression test, performed preoperatively in two patients, showed a small decrease in TSH levels (to 69% of the baseline value in each case), whereas normal controls suppressed to less than 10% of the baseline.
Assessment of thyroid hormone action
Table 5 summarizes the results of
parameters of thyroid hormone action in tumor patients with intact
thyroid compared with those in patients with Graves’ disease, those
with resistance to thyroid hormone, and euthyroid controls. These
results show the absence of acquired resistance to thyroid hormone in
patients with TSH-secreting tumors.
|
Treatment (Table 2b)
Description. All 22 patients who underwent surgery at the NIH had transsphenoidal surgery. Two patients also received transcranial surgery for lateral and extrasellar extension. The cavernous sinuses and/or sphenoid sinus were invaded in 14 patients, and isolated macroscopic dura invasion was observed in 2 cases. In 6 cases, no invasiveness was visible at surgery. An unusually hard and fibrous consistency of the tumor was typical of these tumors, which, combined with frequent dural invasion, resulted in difficult resection and dissection. Tumor immunostaining confirmed the diagnosis of TSH-secreting tumors in all cases.
External radiation of the pituitary, performed in 11 patients in whom
surgical cure was not achieved, was typically given at a dose of
4500–5500 rads (range, 3000–9340); 1 patient underwent a second
course of therapy because of malignant transformation of her tumor, for
a total of 9340 rads. In addition, 1 patient received focused high dose
irradiation with a -knife unit.
Adjuvant medical therapy, including bromocriptine in six patients seen before 1983 and subsequently octreotide at doses of 100-1500 µg daily, was used to maintain euthyroidism or control TSH secretion, usually after a combination of surgery and external radiation.
Treatment outcome. Outcome was assessed in the 23 patients who
were followed for at least 1 yr (Table 2b). Three patients died, 2 in
the immediate postoperative period and 1 of metastatic TSH carcinoma,
previously reported (10, 16). Eight patients (35%) were considered
cured after surgery alone (4 males and 4 females); 2 of these patients
had a short T3 suppression test that was
considered normal.
In patients 4, 11, 13, 16, and 18, although baseline thyroid function
and -subunit levels were normal, the outcome was unclear, as dynamic
tests did not return to normal, including the TRH and
T3 suppression tests. Three of these subjects
have had external radiation as adjunct therapy. In addition, patient 4
has had a 12-yr follow-up without evidence of recurrence, but her TSH
cannot be completely suppressed, suggesting an abnormal pituitary
feedback, such as observed in resistance to thyroid hormone. Two years
after surgery, patient 18 is euthyroid with a normal TRH test, but her
TRH test was also normal before surgery; the only concern is the
T3 suppression test with a suppression of TSH to
only 20% of baseline. Only a longer follow-up, a repetition of the
T3 suppression test, and its validation will
allow correct outcome assessment. Six patients (no. 1, 3, 12, 14, 19,
and 23) had residual macroscopic tumor, with -subunit remaining
elevated in three cases (patients 1, 3, and 12), normalizing in two
cases (patients 19 and 23), and staying normal in one case (patient
14). Five of the six patients were treated and controlled with
octreotide for both endocrine and tumoral parameters. Patient 5
relapsed and was followed elsewhere.
Table 6 summarizes the features that were
recognized to influence the prognosis. Patients meeting the criteria
for cure differed from patients clearly not cured by milder symptoms of
shorter duration, a smaller tumor with absence of visual field
involvement, and a trend for lower levels of TSH for similar levels of
free T4 and for lower levels of -subunits; in
addition, male sex and absence of prior thyroidectomy tended to be
beneficial (cure in, respectively, 66% of men vs. 33% of
women, and in 55% of patients with intact thyroid vs. 28%
of thyroidectomized patients). In addition, the presence of invasion of
the cavernous sinus at surgery was a poor prognostic sign, with none of
the eight cases cured, whereas six of nine cases without invasion were
cured (P = 0.01). The invasiveness at surgery tended to
correlate with imaging size (P = 0.03), but not with
prior thyroidectomy, TSH levels, or any hormonal parameter.
|
In addition to the two deaths previously reported [one at NIH (10) and one elsewhere (17)], early postoperative complications included cerebrospinal leak, panhypopituitarism, and transient or permanent diabetes insipidus. In addition, transient symptomatic inappropriate secretion of ADH occurred in three patients.
Late complications included isolated persistent central hypothyroidism in two patients cured by surgery alone, suggesting damage to normal thyrotroph cells (no recovery after 2.5 and 1.5 yr), secondary hypogonadism after external radiation (one case), and the side-effects of octreotide, such as abdominal cramping, diarrhea, gallbladder dysfunction, or leg pain. Late complications linked to progression of the tumor included death secondary to metastatic thyrotroph carcinoma and sudden irreversible monocular blindness occurring before the availability of octreotide.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Clinical features
The female predominance in our series probably comes from a recruitment bias, as this was not observed in a recent review of the literature (1).
Most patients were diagnosed at the stage of macroadenoma and often
after having been mistaken for having Graves’ disease, resulting in
unwarranted thyroidectomy which altered the hormonal profile. The mean
duration of symptoms before accurate diagnosis was long (9 yr). Tumoral
symptoms and signs, such as headaches and/or visual field defects, were
frequent at diagnosis (
40% of the cases). Patients with visual
field defects tended to have a longer delay to diagnosis (12
vs. 7 yr).
About two thirds of patients had thyroid nodules, probably due to sustained TSH stimulation over many years. The natural history of such nodules is being assessed, but similar nodules have been reported coincident with Graves’ disease after long-standing stimulation by thyroid-stimulating antibodies.
Imaging
The presence of a pituitary tumor in a patient with inappropriate secretion of TSH, although strongly suggestive, is not diagnostic of TSH-secreting tumor, as pituitary incidentalomas have been found on MRI in up to 10% of normal subjects (21). Furthermore, the development of TSH-secreting tumor in patients with resistance to thyroid hormone is theoretically possible, but is a rare occurrence (1, 22). MRI is an excellent technique for assessing these tumors with the caveat that it is not always accurate in the determination of cavernous sinus invasion.
Diagnostic tests
In a standardized manner, the diagnostic value of each classical
test was assessed, drawing on our large study population of patients
with inappropriate secretion of TSH. Baseline TSH values, often in the
normal range in patients with intact thyroid, cannot be used to
diagnose TSH-secreting tumors. Elevation of thyroid hormones in the
face of normal TSH levels is explained by an increased bioactivity of
neoplasm-derived TSH (23). Cosecretion of -subunit was the rule, and
GH or PRL was cosecreted in four cases (16%).
We found that a decreased or lack of response of TSH during the TRH test had a good sensitivity and excellent specificity in patients with intact thyroid, with a slight decrease in sensitivity after thyroidectomy. The lack of response of TSH to TRH in 50% of the patients with intact thyroid suggested that the TRH receptor could be absent (24) or nonfunctional (25) in the tumor; however, TRH receptors are found on the majority of tumors studied (1). A partial or even normal response was retained in other patients, pointing toward possible molecular or etiological differences among TSH-secreting tumors. Conversely, patients with resistance to thyroid hormone typically have a robust response of TSH to TRH (6, 12).
Absolute values of -subunit have excellent sensitivity and
specificity; however, if used alone it could be misleading in
menopausal women, for whom determination of LH and FSH levels is
mandatory. In addition, in patients with prior thyroid treatment, the
elevation of TSH is associated with an elevation of -subunit values,
and we did not have normative data for -subunit at different levels
of TSH.
The -subunit/TSH molar ratio had excellent sensitivity, but less
specificity, in patients with intact thyroid. New norms have been
established for this ratio (16), which take into account gonadotropins
and TSH levels, because the prior cut-off value of 1 (15) was
inadequate in many patients. These norms clearly represent an
improvement; however, we found that they are of limited value when TSH
is only mildly elevated.
In summary, the combination of TRH test, -subunit, and
-subunit/TSH ratio was diagnostic in all untreated patients and in
all but one patient with prior thyroid treatment. For patients with
intact thyroid, the best combined sensitivity and specificity were seen
with the TRH test (71% and 96%) and -subunit (75% and 90%). For
patients with previous thyroid treatment, the best combined sensitivity
and specificity was seen with the -subunit (90% and 82%) and the
-subunit/TSH ratio (90% and 73%).
The short T3 suppression test is easy to perform and well tolerated (14); TSH-secreting tumors are characterized by an absent or inappropriate negative feedback of thyroid hormone on TSH secretion. Our experience with this test as a diagnostic tool is still preliminary, as normative data, and data in other thyroid conditions are not fully established. It could be especially useful in thyroidectomized patients to distinguish a TSH-secreting tumor from mere noncompliance to thyroid hormone replacement resulting in secondary pituitary hypertrophy (1). In addition, it seems promising as a criterion for cure (see below).
The octreotide test was useful before surgery to predict whether the drug could be used as therapy if surgery alone was not curative. This study confirms its diagnostic value, as a marked decrease in TSH was usually observed in TSH-secreting tumors, whereas there was no or little decrease in patients with resistance to thyroid hormone (13). The different response is due to the high concentration of somatostatin receptors (26, 27) on TSH-secreting tumors, which usually respond immediately to somastostatin with a marked decrease in TSH secretion. The test also had the advantage of preparing the patients for surgery, resulting in decreased TSH and thyroid hormone levels. However, for diagnostic purposes, these tests should be reserved for patients in whom the classical tests are atypical, and they may be particularly helpful in patients with prior thyroidectomy or in menopausal hypergonadotropic patients.
Thyroid hormone action parameters
It is striking that a subgroup of patients with TSH-secreting tumors (n = 3 in our series) did not display symptoms of thyrotoxicosis despite frankly elevated thyroid hormone levels; this led to the hypothesis that an acquired state of resistance to thyroid hormone was possible in TSH-secreting tumor patients. However, in agreement with Beck-Peccoz et al. (28), our results do not support this concept. Our group of TSH-secreting tumors resembles Graves’ patients more than patients with resistance to thyroid hormone despite the fact that free T4 levels were, on the average, lower in tumor patients than in Graves’ patients. The apparent tolerance of elevated thyroid hormone in some patients may be linked to the gradual development of hyperthyroidism, leading to compensatory mechanisms, such as desensitization or down-regulation of thyroid hormone receptors.
Therapy
Transsphenoidal surgery was the first line therapy, and related complications increased with the size of the tumor. The 2 patients who died in the immediate postoperative period early in this series had unusually large tumors. Attempts to completely remove the tumor resulted in irreversible brain damage in both patients. Because of this early experience and the advent of octreotide therapy, a more conservative approach is now proposed. Indeed, the fibrous and hard consistency of the tumor may be linked to the secretion by the tumor of a basic fibroblast growth factor (29), making resection more difficult and traumatic. Other complications were relatively minor and transient; 1 worth noting is the occurrence of inappropriate secretion of antidiuretic hormone, which was as frequent as diabetes insipidus (3 of 22 patients). It occurred 7–10 days after surgery, as in other types of pituitary tumors (30), with frequently an abrupt drop in serum sodium associated with variable clinical manifestations, from mild headaches to acute neurological manifestations. This justifies systematic postoperative monitoring of electrolytes before discharge. In the case of apparent surgical cures, the time necessary for the recovery of normal thyrotrophs was variable, 3–12 months. Transient central hypothyroidism may occur, but thyroid hormone replacement is rarely necessary; however, apparent permanent central hypothyroidism in the absence of other pituitary dysfunction occurred in 2 patients.
Twelve patients received external irradiation, including one treated
with a -knife unit. Irradiation should be proposed routinely when
surgery is known to be noncurative even if the patient is still
euthyroid, as relapse is inevitable and the full effect of irradiation
requires months or years. However, in patients who refuse irradiation
because of its potential long term complications (hypopituitarism,
infertility, and impaired cognitive function), somatostatin analogs can
be proposed.
For the patients who were still hyperthyroid despite surgery and external radiation, medical treatment was necessary. As dopamine receptors were detected on TSH-secreting tumors (31), initially bromocriptine was used, but with limited success, before the era of somatostatin. The mechanisms of action of somatostatin (27) may include a decrease in TSH secretion and TSH bioactivity, prompted by a change in glycoisomer distribution (32). The absence of effect of octreotide reported in some patients could be explained by the presence of different subtypes of somatostatin receptors on the tumor (33). Our first patients were started on octreotide in the mid-1980s (34). It proved to be effective for long term therapy in five patients, with a follow-up of up to 13 yr in one case. Tumor size was usually stabilized, but we did not observe the shrinkage observed by others (35). Moreover, this treatment was only transiently effective in the patient with malignant transformation of her TSH-secreting tumor who eventually died of metastatic disease. This treatment is tedious, requiring two or three injections a day. However, new, slow release formulation of somatostatin analogs, not yet available in the U.S., have proven their ability to control TSH-secreting tumors (36) and appear promising to secure compliance and improve quality of life.
Finally, thyroid hormone therapy was necessary in patients with prior thyroidectomy or ablation. The proper dose should maintain circulating thyroid hormones at the upper limit of normal to keep the benefit of even partial negative feedback (25, 37) without causing clinical hyperthyroidism.
Criteria for cure
There has been a recent interest in defining criteria for cure in TSH-secreting tumors. In our opinion, cure should not be defined by mere euthyroidism and absence of visible tumor on imaging. Indeed, some patients relapse despite no apparent macroscopic residual tumor at surgery, probably because of the frequency of microscopic dura invasion (38). Postoperative changes usually preclude an early diagnosis of relapse by MRI, and biochemical clues are the best early indicators. Different tests and new approaches are now used for their capacity to predict cure early. TSH measured 1–7 days after surgery has been proposed recently (39). At that time, the normal thyrotrophs are still suppressed by high thyroid hormone levels, and the measured TSH should reflect tumoral TSH. Later, normalization of dynamic tests should be obtained, including TRH and T3 suppression tests. However, the absence of normalization of TRH test can be seen even in cured patients in the case of central postsurgical hypothyroidism, with low or normal thyroid hormone levels. In any case, long term follow-up is necessary to detect a relapse.
Outcome
The rate of apparent surgical cure here (35%) was similar to that in other series, even though criteria for cure and length of follow-up differed from those in other series (17, 39, 40, 41). Our criteria for cure were more stringent than those in all other series, except for that of Losa (39). We observed an increase in the rate of surgical cure in recent years (47% since 1990 vs. 13% before), perhaps reflecting the earlier diagnosis of these tumors.
Importantly, all patients except 1, who died of malignant
transformation (10), were controlled after adjunctive medical
treatment. Thirteen patients, of the 19 alive and followed-up for at
least 1 yr, were euthyroid and under control without need for
octreotide; 5 needed octreotide to control tumoral TSH production, and
1 has relapsed with borderline hyperthyroidism and has just undergone
focused radiation with a -knife.
Prognostic factors
The main prognostic factor was the size of the tumor and its invasiveness; diagnosis delay was also important, usually due to misdiagnosis, resulting in unwarranted thyroidectomy and larger tumors.
That males had a slightly better prognosis may be due to the fact that Graves’ disease is rare in males, so the diagnosis of secondary hyperthyroidism may be entertained earlier. Also, estrogen receptors have been found on TSH-secreting tumors and, at least in theory, could foster the growth of the tumor (42). However, this in vitro observation has not proven to be important clinically. Therefore, estrogen replacement may not be contraindicated in menopausal women, if the tumor is under control. In fact, it should help maintain bone mass in menopausal women who have been exposed to hyperthyroidism for several years.
In conclusion, this study of a large series of TSH-secreting tumors
followed prospectively at one institution emphasizes the importance of
early recognition. This is now possible with the availability of third
generation TSH assays, which should be requested for patients without
the specific features of Graves’ disease, such as ophtalmopathy and
dermopathy. As a rule, the TRH test, and -subunit and
-subunit/TSH ratio measurements together with MRI of the pituitary
are diagnostic. However, in difficult cases, more recently proposed
dynamic tests, including octreotide and T3
suppression tests may be useful. These tests are also helpful after
surgery to predict apparent surgical cure, thus sparing the need for
systematic sellar irradiation. However, long term follow-up is needed
for early detection and treatment of recurrence.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Present address: Department of Medicine, Division of Endocrinology,
University of Maryland School of Medicine, and the Institute of Human
Virology, 725 West Lombard Street, Baltimore, Maryland 21201. 
Received May 13, 1998.
Revised November 9, 1998.
Accepted November 12, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and ß subunit levels to identify patients
with primary tumors. J Clin Endocrinol Metab. 45:534–543.[Medline]
-subunit in pituitary adenomas. Trends
Endocrinol Metab. 3:41–45.
This article has been cited by other articles:
 |
|
 |
|
  E. S. Mittra, R. D. Niederkohr, C. Rodriguez, T. El-Maghraby, and I. R. McDougall Uncommon Causes of Thyrotoxicosis J. Nucl. Med., February 1, 2008; 49(2): 265 - 278. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kienitz, M. Quinkler, C. J Strasburger, and M. Ventz Long-term management in five cases of TSH-secreting pituitary adenomas: a single center study and review of the literature Eur. J. Endocrinol., July 1, 2007; 157(1): 39 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Daousi, P M Foy, and I A MacFarlane Ablative thyroid treatment for thyrotoxicosis due to thyrotropin-producing pituitary tumours J. Neurol. Neurosurg. Psychiatry, January 1, 2007; 78(1): 93 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. C. Nemergut, A. S. Dumont, U. T. Barry, and E. R. Laws Perioperative Management of Patients Undergoing Transsphenoidal Pituitary Surgery Anesth. Analg., October 1, 2005; 101(4): 1170 - 1181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Furumoto, H. Ying, G. V. R. Chandramouli, L. Zhao, R. L. Walker, P. S. Meltzer, M. C. Willingham, and S.-Y. Cheng An Unliganded Thyroid Hormone {beta} Receptor Activates the Cyclin D1/Cyclin-Dependent Kinase/Retinoblastoma/E2F Pathway and Induces Pituitary Tumorigenesis Mol. Cell. Biol., January 1, 2005; 25(1): 124 - 135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. J. Sarlis, L. Gourgiotis, C. A. Koch, M. C. Skarulis, F. Brucker-Davis, J. L. Doppman, E. H. Oldfield, and N. J. Patronas MR Imaging Features of Thyrotropin-Secreting Pituitary Adenomas at Initial Presentation Am. J. Roentgenol., August 1, 2003; 181(2): 577 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ando, N. J. Sarlis, E. H. Oldfield, and P. M. Yen Somatic Mutation of TR{beta} Can Cause a Defect in Negative Regulation of TSH in a TSH-Secreting Pituitary Tumor J. Clin. Endocrinol. Metab., November 1, 2001; 86(11): 5572 - 5576. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ando, N. J. Sarlis, J. Krishnan, X. Feng, S. Refetoff, M. Q. Zhang, E. H. Oldfield, and P. M. Yen Aberrant Alternative Splicing of Thyroid Hormone Receptor in a TSH-Secreting Pituitary Tumor Is A Mechanism for Hormone Resistance Mol. Endocrinol., September 1, 2001; 15(9): 1529 - 1538. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Confusion in an elderly patient: an uncommon diagnosis for such a common event Postgrad. Med. J., August 1, 2001; 77(910): 537d - 537. [Full Text] [PDF]
|
|
|
|
|
|
P. Caron, S. Arlot, C. Bauters, P. Chanson, J.-M. Kuhn, M. Pugeat, R. Marechaud, C. Teutsch, E. Vidal, and P. Sassano Efficacy of the Long-Acting Octreotide Formulation (Octreotide-Lar) in Patients with Thyrotropin-Secreting Pituitary Adenomas J. Clin. Endocrinol. Metab., June 1, 2001; 86(6): 2849 - 2853. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Pituitary Tumors in Children and Adolescents J. Clin. Endocrinol. Metab., December 1, 1999; 84(12): 4317 - 4323. [Full Text]
|
|
|
|
|
|
P. U. Freda and S. L. Wardlaw Diagnosis and Treatment of Pituitary Tumors J. Clin. Endocrinol. Metab., November 1, 1999; 84(11): 3859 - 3866. [Full Text]
|
|
| ||||||||||||||
