|
A correction has been published: N Engl J Med 1993;328(16):1208.
|
| |||||||||||||||||||||||||||||||||
Background Patients with brain tumors who are treated with radiation frequently have growth hormone deficiency, but other neuroendocrine abnormalities are presumed to be uncommon.
Methods We studied endocrine function in 32 patients (age, 6 to 65 years) 2 to 13 years after they had received cranial radiotherapy for brain tumors. The doses of radiation to the hypothalamic-pituitary region ranged from 3960 to 7020 rad (39.6 to 70.2 Gy). Nine patients also received 1800 to 3960 rad (18.0 to 39.6 Gy) to the craniospinal axis. Serum concentrations of thyroid, gonadal, and pituitary hormones were measured at base line and after stimulation.
Results Nine patients (28 percent) had symptoms of thyroid deficiency, and 20 patients (62 percent) had low serum total or free thyroxine or total triiodothyronine concentrations. Of the 23 patients treated only with cranial radiation, 15 (65 percent) had hypothalamic or pituitary hypothyroidism. Of the nine patients who also received spinal (and thus direct thyroid) radiation, three (33 percent) had evidence of primary thyroid injury.
Seven of the 10 postpubertal, premenopausal women (70 percent) had oligomenorrhea, and 5 (50 percent) had low serum estradiol concentrations. Three of the 10 men (30 percent) had low serum testosterone concentrations. Overall, 14 of the 23 postpubertal patients (61 percent) had evidence of hypogonadism. Mild hyperprolactinemia was present in 50 percent of the patients. Responses to stimulation with corticotropin-releasing hormone and corticotropin were normal in all patients except one, who had panhypothalamic dysfunction. However, serum 11-deoxycortisol responses to the administration of metyrapone were low in 11 of the 31 patients (35 percent) tested.
Three of the 32 patients (9 percent) had no endocrine abnormalities, 9 (28 percent) had an abnormal result on tests of thyroid, gonadal, prolactin, or adrenal function, 8 (25 percent) had abnormalities in two axes, 8 (25 percent) in three axes, and 4 (12 percent) in all four axes.
Conclusions Cranial radiotherapy in children and adults with brain tumors frequently causes abnormal hypothalamic-pituitary function. The most frequent changes are hypothyroidism and gonadal dysfunction, although subtle abnormalities in adrenal function may also be present.
With respect to the secretion of growth hormone, 65 percent of children with acute lymphoblastic leukemia treated with 2000 to 3000 rad (20 to 30 Gy) of cranial radiation have impaired serum growth hormone responses to provocative stimulation,12 and children with brain tumors who receive higher doses are even more likely to have abnormal secretion of growth hormone and abnormal growth1,3,7,11,13. The site of the injury is apparently the hypothalamus, since growth hormone-releasing hormone appropriately stimulates the secretion of growth hormone in these patients19.
Most studies of other neuroendocrine sequelae of cranial radiation have been done in children who received relatively low doses of radiation. For example, less than 5 percent of children treated for medulloblastoma with craniospinal radiation have hypothalamic or pituitary hypothyroidism,3,11 but the dose of radiation to the hypothalamus and pituitary is generally 3600 rad (36 Gy) or less. Conversely, about 30 percent of patients treated for other head or neck (generally nasopharyngeal) tumors with high doses of radiation that include the hypothalamic-pituitary axis have hypothyroidism, and about 80 percent have some type of hypothalamic-pituitary dysfunction20,21,22,23.
We began to appreciate that neuroendocrine abnormalities were more frequent than generally presumed when we evaluated a group of children and adults with brain tumors who had received radiation therapy24. Therefore, we examined these patients more thoroughly to determine the frequency and severity of clinical and chemical abnormalities of thyroid, gonadal, and adrenal function; the site of the abnormalities; and their relation to demographic or therapeutic factors.
Methods
Patients
The study group consisted of patients treated at the University of Rochester Cancer Center for a brain tumor not involving any portion of the hypothalamic-pituitary axis. Patients were considered eligible if they had no evidence of tumor recurrence, had not received radiation or chemotherapy for at least two years, and had been seizure-free for at least one year. All patients receiving care in the radiation-oncology follow-up clinic who met these criteria were asked to participate in the study, and informed consent was obtained from the patients or their parents. Patients who were taking anticonvulsant agents or hormones discontinued their medications at least one month (two months in the case of estrogen and progesterone preparations) before testing began.
The study group contained 32 patients (16 female and 16 male) who had been treated since 1973. Twenty-one patients had gliomas, five medulloblastomas, two ependymomas, and four other types of tumors. Their mean age at the time of radiation therapy was 19 years (range, 2 to 57), and at the time of the study, it was 26 years (range, 6 to 65). The mean interval between therapy and testing was 7 years (range, 2 to 13). Twenty patients (10 girls and 10 boys) were children ( 
18 years of age) at the time of radiation therapy, and 16 were children (6 to 18 years) when tested.
Radiation Therapy
All patients had received radiation therapy, with a cobalt-60 unit or linear accelerators of 10 or 18 MV, to all or part of the hypothalamus and pituitary gland. Simulation or beam films were available for all patients and used to estimate the dose of radiation received by the two areas. Some patients had received whole-brain radiation, whereas in others treatment was localized. The calculated doses delivered to the hypothalamus and pituitary ranged from 3960 to 7020 rad (39.6 to 70.2 Gy), with a mean of 5360 rad (53.6 Gy). Nine patients had also received radiation to the spinal axis (1800 to 3960 rad [18.0 to 39.6 Gy]; mean, 3100 rad [31.0 Gy]). The dose of radiation to the thyroid gland, measured by thermoluminescent dosimetry, was 1.4 percent of the midline cranial dose. For the patients who received spinal-axis radiation, the dose to the thyroid averaged 92 percent of that prescribed.
Chemotherapy
Six adult patients received carmustine chemotherapy (80 mg per square meter of body-surface area intravenously for three days every two months for one to two years) concurrently with radiation therapy. One adult received six intravenous doses of methotrexate before undergoing any radiation therapy, and one child received vincristine, prednisone, and lomustine during and after radiation therapy.
Endocrinologic Evaluation
The patients were admitted to the University of Rochester Clinical Research Center early on the morning of study, at which time a history was obtained and a complete physical examination performed. Patients were considered postpubertal if they had mature secondary sexual characteristics (Tanner stage 5). Height was measured in the 20 patients who were irradiated during childhood, 19 of whom were 2 to 13 years of age when treated, and 6 of whom were 18 years or older when their height was measured.
All patients received thyrotropin-releasing hormone (TRH, 7 µg per kilogram of body weight; maximal dose, 200 µg), gonadotropin-releasing hormone (GnRH, 100 µg), and corticotropin (0.25 mg) intravenously in a single injection between 8 and 10 a.m. Fifteen adults also received corticotropin-releasing hormone (1 µg per kilogram) intravenously between 5 and 8 p.m. Blood samples were obtained for measurements of serum pituitary hormones and cortisol 20 minutes before the injection, at the time of the injection, and 15, 30, 45, 60, 90, and 150 minutes after the injection. Serum concentrations of free and total thyroxine (T4), total triiodothyronine (T3), estradiol, and testosterone were measured 20 minutes before the injection and at the time of the injection. Metyrapone was given orally (adults, 1.5 g; children, 30 mg per kilogram) between 11:30 p.m. and midnight; blood samples were obtained for measurements of serum 11-deoxycortisol, cortisol, and corticotropin at 8 a.m. the next morning. Serum concentrations of growth hormone or insulin-like growth factor I were not measured. Patients were not tested for posterior pituitary function, but no patient had symptoms of diabetes insipidus, excessive urinary output, or increased serum electrolyte concentrations.
Hormonal Assays
All hormones were measured by radioimmunoassay with commercial kits. The sensitivity and intraassay and interassay coefficients of variation were as follows: free T4: 0.085 ng per deciliter, 5.3 percent and 4.9 percent (Dade-Baxter Travenol Diagnostics, Cambridge, Mass.); total T4: 0.35 µg per deciliter, 5.0 percent and 5.3 percent (Dade-Baxter Travenol); total T3: 9.0 ng per deciliter, 3.6 percent and 4.8 percent (Clinical Assays, Stillwater, Minn.); thyrotropin: 0.3 mU per liter, 3.7 percent and 5.1 percent (Hybritech, San Diego, Calif.); cortisol: 0.20 µg per deciliter, 6.5 percent and 6.0 percent (Diagnostic Products, Los Angeles); testosterone: 4.03 ng per deciliter, 6.6 percent and 10.4 percent (Diagnostic Products); estradiol: 7.9 pg per milliliter, 4.3 percent and 6.8 percent (Diagnostic Products); prolactin: 3.7 ng per milliliter, 4.8 percent and 12.9 percent (Diagnostic Products); corticotropin: 15 ng per liter, 6.3 percent and 6.0 percent (Incstar, Stillwater, Minn.); luteinizing hormone (LH): 0.4 IU per liter, 3.4 percent and 4.0 percent (Clinetics, Tustin, Calif.); and follicle-stimulating hormone (FSH): 0.7 IU per liter, 3.3 percent and 3.9 percent (Clinetics). The basal values reported are the mean of the values in the samples obtained 20 minutes before the injection and at the time of injection.
The reference ranges (mean ±2 SD) for serum total and free T4 and total T3 concentrations were obtained by determining the levels in 80 normal subjects (31 men and 49 women), ranging from 19 to 67 years of age (median, 29). The reference ranges for the remaining hormones were obtained from the Laboratory Medicine Division of Strong Memorial Hospital, with the exception of the peak serum cortisol concentration in response to corticotropin-releasing hormone, which was from the literature25.
Statistical Analysis
The results are expressed as the mean ±SD. The serum total T4, free T4, total T3, and thyrotropin concentrations were compared between groups with the Mann-Whitney test26. Associations between endocrine measures and the dose of radiation to the hypothalamic-pituitary region, the age at treatment, and the interval between treatment and the study were assessed with Kendall's tau test27. The frequency of abnormalities in different groups was compared by Fisher's exact test28. The P values for all tests are two-sided. A patient's hormonal secretion was considered abnormal if he or she had a value outside the reference range, either basally or after provocative testing. In the summary analysis of data, patients were considered to have abnormal hormonal secretion if their basal serum hormone concentrations were low (high for prolactin only) or if they had a blunted response to provocative testing.
Results
Thyroid Function
Nine of the 32 patients had symptoms or signs of hypothyroidism, including lethargy (7 patients); intolerance to the cold (5 patients); dry hair, dry skin, and pronounced weight gain (2 patients each); and constipation (1 patient). These patients and nine others who were asymptomatic had low serum free T4 concentrations. All clinical abnormalities were ameliorated by thyroxine therapy.
The mean serum free T4, total T4, and total T3 concentrations in the 23 patients in the cranial-radiation group were significantly lower than those in the normal subjects (P<0.001 for each comparison) (Table 1). Fifteen (65 percent), seven (30 percent), and five (22 percent) patients had serum free T4, total T4, and total T3 concentrations below normal, respectively, and all patients with low serum total T4 concentrations had low serum free T4 concentrations. The mean basal and peak serum thyrotropin concentrations in response to stimulation with TRH were within the respective normal ranges, and no patient had a supranormal response. Twelve of the 15 patients with below-normal concentrations of serum free T4 (80 percent) were thought to have hypothalamic hypothyroidism, as defined by a low serum free T4 concentration and normal basal and peak serum thyrotropin concentrations in response to stimulation with TRH. Three patients with a low serum free T4 concentration and a blunted response of serum thyrotropin (<5.0 mU per liter) to TRH were considered to have either hypothalamic or pituitary hypothyroidism.
|
There was a significant negative relation between the serum free T4 concentration and the interval between radiation and testing (P = 0.02) (Figure 1A), but not between serum total T4 or T3 concentrations and this interval. Five of the 12 patients tested less than five years after receiving radiation had low serum free T4 concentrations, as compared with 13 of the 20 patients tested later.
|
Both the basal and peak serum thyrotropin concentrations after stimulation with TRH were negatively correlated with age in the patients treated with cranial radiation (P = 0.007 and P = 0.032, respectively). Serum total T4 and T3 but not free T4 concentrations also were negatively correlated with age in the combined study group (P = 0.007 and P = 0.003, respectively) and in the 23 patients who received cranial radiation (P = 0.02 and P = 0.021, respectively).
Prolactin Secretion
The basal serum prolactin concentration was elevated in 16 of the 32 patients (50 percent) (Table 2 and Figure 2). Two of the six women and one of the seven men with hyperprolactinemia reported decreased libido. On the basis of the patients' ages at the time of radiation therapy, hyperprolactinemia was more common in adults (75 percent) than in children (30 percent). One of the three patients with primary hypothyroidism had hyperprolactinemia.
|
|
Gonadal Function
Of the 10 female patients who were postpubertal and premenopausal when studied, 3 had primary oligomenorrhea and 4 had secondary oligomenorrhea. Serum estradiol concentrations were undetectable in four, low in one, normal in four (including two with normal menstruation), and elevated in a sexually mature 13-year-old girl (bone age, 16 years) who was menstruating (Table 2). Four of the seven patients with oligomenorrhea as well as one normally menstruating woman had hyperprolactinemia. The basal serum FSH and LH concentrations were low in 0 and 4 of the 10 postpubertal, premenopausal women, respectively. The 13-year-old sexually mature girl had an elevated basal serum FSH concentration (28.2 IU per liter) and elevated peak serum FSH and LH responses to stimulation with GnRH. Both postmenopausal women had low basal serum FSH and LH concentrations. The peak serum FSH and LH responses to GnRH were low in only 3 and 1 of 11 postpubertal women, respectively (Figure 3), despite the high frequency of oligomenorrhea.
|
Seven children (four boys and three girls) were prepubertal, and two children (one boy and one girl) were pubertal when studied. The pubertal girl, who was 15 years of age (and who was irradiated at the age of 11), had no menstrual cycles, partial sexual development (Tanner stage 2 to 3), an undetectable serum estradiol concentration, and basal serum FSH and LH concentrations of 7.3 and 1.1 IU per liter, respectively. Of the prepubertal girls, one who was 11 years old (and who was irradiated at the age of 5) had an accelerated bone age (16 years) but no sexual development. The remaining two girls, who were 6 and 10 years of age when studied (2 and 6 years of age, respectively, when treated) were developmentally normal for their age. Among the six female patients who were treated before puberty (age, 5 to 11 years) and studied after the expected age of puberty, all but one had Tanner stage 4 or 5 sexual characteristics. The pubertal boy was 15 years of age (irradiated at the age of 12) and had Tanner stage 3 development and a normal serum testosterone concentration. Of the prepubertal boys, one who was 14 years old (irradiated at the age of 7) had no sexual development (Tanner stage 1), a low serum testosterone concentration, and a normal serum prolactin concentration. The remaining three boys, who were 8 to 11 years old when studied (4 to 7 years of age when treated), were developmentally normal for their age. Four other male patients who were treated before puberty (age, 3 to 13 years) attained full sexual maturation. The mean (±SD) basal serum LH and FSH concentrations for all seven prepubertal children were 3.5 ±1.7 and 4.9 ±1.8 IU per liter, respectively. The two pubertal children had normal serum FSH and LH responses to stimulation with GnRH29,30. The gonadotropin responses to GnRH in the prepubertal children are not included in Table 2 and Figure 3 because values for normal children were not available and the values reported in the literature vary widely29,30,31,32.
Adrenal Function
No patient had symptoms or signs of adrenal insufficiency, but three patients had low basal serum cortisol concentrations, and one patient had a low serum cortisol response to stimulation with corticotropin (Table 3). This patient had panhypothalamic dysfunction, including hypothyroidism and hyperprolactinemia.
|
Eleven of the 31 patients tested (35 percent) had a subnormal serum 11-deoxycortisol response to the administration of metyrapone. There was an inverse correlation between the intensity of the serum 11-deoxycortisol response and the dose of radiation (P<0.001), and all patients with subnormal responses had received more than 5000 rad (50 Gy) (Figure 4).
|
Only 4 of the 11 children who received cranial radiation and 2 of the 8 children who received craniospinal radiation attained a height above the third percentile. One patient was fully grown when treated at the age of 18 years. However, formal testing of growth hormone secretion was not performed.
Chemotherapy
Eight of the 32 patients received chemotherapy. There was no significant difference in the frequency of abnormal hormonal test results between these 8 patients and the 24 patients who did not receive chemotherapy, with the exception of elevated basal serum prolactin concentrations (P = 0.037). Of the eight patients who received chemotherapy, seven had hyperprolactinemia, two had primary and three had hypothalamic hypothyroidism, three had low gonadal test results, and three had low adrenal test results.
Summary of Findings
The pattern of hormonal (i.e., thyroid, gonadal, prolactin, and adrenal) abnormalities varied (Table 4). There was no discernible predominance of any combination of hormonal deficits.
|
The reported frequency of pituitary or hypothalamic hypothyroidism and hypoadrenalism in patients with brain tumors not involving the pituitary gland who are treated with radiation is generally less than 5 percent, whereas that for hypogonadism is less than 20 percent1,2,3,4,5,6,7,8,9,10,11,12,13,33,34. In patients treated for other head or neck tumors with high doses of radiation to the hypothalamus and pituitary, the frequency of hormonal abnormalities ranges from 8 to 50 percent20,21,22,23. Possible explanations for this apparent inconsistency include the previous focus on children with medulloblastoma, who generally receive relatively low doses of radiation; the short interval between treatment and testing in some reports; the unreliability of clinical evaluations of endocrine deficiency; and the less sensitive and specific laboratory assessments used in the past. Because the results of our study are similar to those reported in patients with head or neck tumors, we suggest that endocrine dysfunction will occur frequently when patients with brain tumors are aggressively treated with large doses of radiation to a region that encompasses the hypothalamus and pituitary gland.
The most striking abnormality was the substantial frequency of pituitary or hypothalamic hypothyroidism in the patients who received cranial radiation alone. The greater number of patients found to have low serum free T4 concentrations, as compared with low serum total T4 or total T3 concentrations, presumably reflects the sensitivity of the free T4 measurement in the assessment of thyroid function. The most common pattern of abnormality was low concentrations of serum free or total T4 or T3 and normal basal and TRH-stimulated serum thyrotropin concentrations, findings suggestive of TRH deficiency35,36. Although the mechanism of radiation damage to the hypothalamus or pituitary is not known, it must involve a direct injury to the cells responsible for hormonal secretion, an injury to the stroma or its microvasculature, or an injury to the vascular channels that transfer the hypothalamic hormones to the pituitary. The correlation between the frequency of hypothyroidism and the interval between treatment and evaluation is compatible with both vascular injury and damage to parenchymal cells, which have a slow rate of turnover37. Since tissue injury caused by radiation is generally related to the dose of radiation, an increase in the frequency or severity of dysfunction with an increasing dose might be expected. Children treated with radiation therapy for leukemia or brain tumors are known to have growth hormone deficiency, with the onset dependent on the dose of radiation,38 and adults with pituitary tumors treated with radiation therapy have endocrine deficiencies more commonly after higher doses of radiation39. We found a clear inverse relation between the dose of radiation and serum total T4 and T3 concentrations, as well as basal and stimulated serum thyrotropin concentrations. The absence of an association between the dose of radiation and serum free T4 concentration is unexplained. The influence of age on the propensity for hypothalamic-pituitary injury also remains unclear. Although older age did appear to be associated with hypothyroidism, the interpretation of this association is complicated, because most of the children received lower doses of radiation.
The occurrence of primary hypothyroidism in patients who received craniospinal radiation was expected, because their thyroid gland was irradiated directly. Primary hypothyroidism resulting from direct radiation to the thyroid gland is well recognized; up to 80 percent of patients with Hodgkin's disease treated with mantle radiation have primary hypothyroidism, and the effect is dose-dependent40,41,42,43. The hypothyroidism that occurs in these patients could be due to radiation damage to the thyroid follicular cells, the thyroid vasculature, or the supporting stroma37.
The 50 percent frequency of hyperprolactinemia in our patients is best explained by an injury to the hypothalamus that disrupts its normal inhibition of prolactin secretion by the pituitary lactotrophs. The extent to which hyperprolactinemia contributed to gonadal dysfunction is uncertain, since the hyperprolactinemia was slight.
Seventy percent of the women and one of the pubertal girls had oligomenorrhea. The presence of low serum estradiol concentrations and low or normal basal and GnRH-stimulated serum LH and FSH concentrations is compatible with GnRH deficiency. Similarly, the decreased serum testosterone concentrations in some of the men may have resulted from GnRH deficiency, since their serum LH and FSH concentrations were normal or slightly low. The documented capability of the pituitary to respond to GnRH argues against pituitary, as opposed to hypothalamic, hypogonadism.
Basal pituitary-adrenal function was normal, or nearly so, in most patients. Thirty-five percent of the patients, however, had a low serum 11-deoxycortisol response to the administration of metyrapone, and the degree of the abnormality correlated inversely with the dose of radiation. Yet the results of corticotropin-releasing hormone and corticotropin stimulation were normal or slightly increased. This apparent inconsistency may be due to the fact that metyrapone stimulates the release of corticotropin by lowering cortisol secretion, whereas corticotropin and to a lesser extent corticotropin-releasing hormone were administered in pharmacologic doses.
In conclusion, both children and adults with brain tumors are at risk for hypothalamic and pituitary injury after radiation therapy. The spectrum of endocrinologic abnormalities that occurs suggests that hypothalamic dysfunction is more common than pituitary dysfunction. The clinical consequences of these abnormalities are most apparent in the thyroid and gonadal axes. Evaluating patients after radiation therapy for brain tumors for neuroendocrine sequelae, and treating them when necessary, should enhance the quality of their lives.
Supported in part by a general Clinical Research Center grant (RR00044) from the Division of Research Resources, National Institutes of Health.
We are indebted to Mrs. Ann G. Muhs for assistance in the preparation of the manuscript.
Source Information
From the Department of Radiation Oncology (L.S.C., P.R.), Pediatric Oncology Division, Department of Pediatrics (L.S.C.), Endocrinology Division, Department of Medicine (P.D.W.), and the Department of Biostatistics (R.F.R.), University of Rochester Medical Center, Rochester, N.Y.; the Department of Radiation Oncology, Dover General Hospital, Dover, N.J. (D.C.); and the Endocrinology Division, Department of Pediatrics, State University of New York Health Science Center, Syracuse (G.M., K.M.). Presented in part at the 33rd Annual Meeting of the American Society for Therapeutic Radiology and Oncology, Washington, D.C., November 4-8, 1991.
Address reprint requests to Dr. Constine at the Department of Radiation Oncology, University of Rochester Medical Center, Rochester, NY 14642-8647.
References
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
