The impact of age and gender on papillary thyroid cancer survival.

CONTEXT: Thyroid cancer predominately affects women, carries a worse prognosis in older age, and may have higher mortality in men. Superimposed on these observations is the fact that most women have attained menopause by age 55 yr. OBJECTIVE: The objective of the study was to determine whether men contribute disproportionately to papillary thyroid cancer (PTC) mortality or whether menopause affects PTC prognosis. DESIGN: Gender-specific mortality was normalized using age-matched subjects from the U.S. population. Multivariate Cox proportional hazard regression models incorporating gender, age, and National Thyroid Cancer Treatment Cooperative Study Group stage were used to model disease-specific survival (DSS). PARTICIPANTS AND SETTING: Patients were followed in a prospective registry. MAIN OUTCOME MEASURE: The relationships between gender, age, and PTC outcomes were analyzed. RESULTS: The unadjusted hazard ratio (HR) for DSS for women was 0.40 [confidence interval (CI) 0.24-0.65]. This female advantage diminished when DSS was adjusted for age at diagnosis and stage with a HR encompassing unity (HR 0.72, CI 0.44-1.19). Additional multivariate models of DSS considering gender, disease stage, and various age groupings showed that the DSS for women diagnosed at under 55 yr was improved over men (HR 0.33, CI 0.13-0.81). However, the HR for DSS increased to become similar to men for women diagnosed at 55-69 yr (HR 1.01, CI 0.42-2.37) and at 70 yr or greater (HR 1.17, CI 0.48-2.85). CONCLUSIONS: Although the overall outcome of women with PTC is similar to men, subgroup analysis showed that this composite outcome is composed of two periods with different outcomes. The first period is a period with better outcomes for women than men when the diagnosis occurs at younger than 55 yr; the second is a period with similar outcomes for both women and men diagnosed at ages greater than 55 yr. These data raise the question of whether an older age cutoff would improve current staging systems. We hypothesize that older age modifies the effect of gender on outcomes due to menopause-associated hormonal alterations.

American Thyroid Association guidelines for detection of thyroid dysfunction.

OBJECTIVE: To define the optimal approach to identify patients with thyroid dysfunction. PARTICIPANTS: The 8-member Standards of Care Committee of the American Thyroid Association prepared a draft, which was reviewed by the association's 780 members, 50 of whom responded with suggested revisions. EVIDENCE: Relevant published studies were identified through MEDLINE and the association membership's personal resources. CONSENSUS PROCESS: Consensus was reached at group meetings. The first draft was prepared by a single author (P.W.L.) after group discussion. Suggested revisions were incorporated after consideration by the committee. CONCLUSIONS: The American Thyroid Association recommends that adults be screened for thyroid dysfunction by measurement of the serum thyrotropin concentration, beginning at age 35 years and every 5 years thereafter. The indication for screening is particularly compelling in women, but it can also be justified in men as a relatively cost-effective measure in the context of the periodic health examination. Individuals with symptoms and signs potentially attributable to thyroid dysfunction and those with risk factors for its development may require more frequent serum thyrotropin testing.

Outcome after treatment of high-risk papillary and non-Hürthle-cell follicular thyroid carcinoma.

BACKGROUND: Treatment of differentiated thyroid cancer has been studied for many years, but the benefits of extensive initial thyroid surgery and the addition of radioiodine therapy or external radiation therapy remain controversial. OBJECTIVE: To determine the relations among extent of surgery, radioiodine therapy, and external radiation therapy in the treatment of high-risk papillary and non-Hürthle-cell follicular thyroid carcinoma. DESIGN: Analysis of data from a multicenter study. SETTING: 14 institutions in the United States and Canada participating in the National Thyroid Cancer Treatment Cooperative Study Registry. PATIENT: 385 patients with high-risk thyroid cancer (303 with papillary carcinoma and 82 with follicular carcinoma). MEASUREMENTS: Death, disease progression, and disease-free survival. RESULTS: Total or near-total thyroidectomy was done in 85.3% of patients with papillary carcinoma and 71.3% of patients with follicular cancer. Overall surgical complication rate was 14.3%. Total or near-total thyroidectomy improved overall survival (risk ratio [RR], 0.37 [95% CI, 0.18 to 0.75]) but not cancer-specific mortality, progression, or disease-free survival in patients with papillary cancer. No effect of extent of surgery was seen in patients with follicular thyroid cancer. Postoperative iodine-131 was given to 85.4% of patients with papillary cancer and 79.3% of patients with follicular cancer. In patients with papillary cancer, radioiodine therapy was associated with improvement in cancer-specific mortality (RR, 0.30 [CI, 0.09 to 0.93 by multivariate analysis only]) and progression (RR, 0.30 [CI, 0.13 to 0.72]). When tall-cell variants were excluded, the effect on outcome was not significant. After radioiodine therapy, patients with follicular thyroid cancer had improvement in overall mortality (RR, 0.17 [CI, 0.06 to 0.47]), cancer-specific mortality (RR, 0.12 [CI, 0.04 to 0.42]), progression (RR, 0.21 [CI, 0.08 to 0.56]), and disease-free survival (RR, 0.29 [CI, 0.08 to 1.01]). External radiation therapy to the neck was given to 18.5% of patients and was not associated with improved survival, lack of progression, or disease-free survival. CONCLUSIONS: This study supports improvement in overall and cancer-specific mortality among patients with papillary and follicular thyroid cancer after postoperative iodine-131 therapy. Radioiodine therapy was also associated with improvement in progression in patients with papillary cancer and improvement in progression and disease-free survival in patients with follicular carcinoma.

Thyrotropin suppression and disease progression in patients with differentiated thyroid cancer: results from the National Thyroid Cancer Treatment Cooperative Registry.

The ideal therapy for differentiated thyroid cancer is uncertain. Although thyroid hormone treatment is pivotal, the degree of thyrotropin (TSH) suppression that is required to prevent recurrences has not been studied in detail. We have examined the relation of TSH suppression to baseline disease characteristics and to the likelihood of disease progression in a cohort of thyroid cancer patients who have been followed in a multicenter thyroid cancer registry that was established in 1986. The present study describes 617 patients with papillary and 66 patients with follicular thyroid cancer followed annually for a median of 4.5 years (range 1-8.6 years). Cancer staging was assessed using a staging scheme developed and validated by the registry. Cancer status was defined as no residual disease; progressive disease at any follow-up time; or death from thyroid cancer. A mean TSH score was calculated for each patient by averaging all available TSH determinations, where 1 = undetectable TSH; 2 = subnormal TSH; 3 = normal TSH; and 4 = elevated TSH. Patients were also grouped by their TSH scores: group 1: mean TSH score 1.0-1.99; group 2: mean TSH score 2.0-2.99; group 3: mean TSH score 3.0-4.0. The degree of TSH suppression did not differ between papillary and follicular thyroid cancer patients. However, TSH suppression was greater in papillary cancer patients who were initially classified as being at higher risk for recurrence. This was not the case for follicular cancer patients, where TSH suppression was similar for all patients. For all stages of papillary cancer, a Cox proportional hazards model showed that disease stage, patient age, and radioiodine therapy all predicted disease progression, but TSH score category did not. However, TSH score category was an independent predictor of disease progression in high risk patients (p = 0.03), but was no longer significant when radioiodine therapy was included in the model (p = 0.09). There were too few patients with follicular cancer for multivariate analysis. These data suggest that physicians use greater degrees of TSH suppression in higher risk papillary cancer patients. Our data do not support the concept that greater degrees of TSH suppression are required to prevent disease progression in low-risk patients, but this possibility remains in high-risk patients. Additional studies with more patients and longer follow-up may provide the answer to this important question.

Prospective multicenter study of thyroiscarcinoma treatment: initial analysis of staging and outcome. National Thyroid Cancer Treatment Cooperative Study Registry Group.

BACKGROUND: A novel prognostic staging classification encompassing all forms of thyroid carcinoma was created for the National Thyroid Cancer Treatment Cooperative Study (NTCTCS) Registry, with the goal of prospective validation and comparison with other available staging classifications. METHODS: Patient information was recorded prospectively from 14 institutions. Clinicopathologic staging was based on patient age at diagnosis, tumor histology, tumor size, intrathyroidal multifocality, extraglandular invasion, metastases, and tumor differentiation. RESULTS: Between 1987 and 1995, 1607 patients were registered. Approximately 43% of patients were classified as NTCTCS Stage I, 24% Stage II, 24% Stage III, and 9% Stage IV. Patients with follicular carcinoma were more likely to have "high risk" Stage III or IV disease than those with papillary carcinoma. Of 1562 patients for whom censored follow-up was available (median follow-up, 40 months), 78 died of thyroid carcinoma or complications of its treatment. Five-year product-limit patient disease specific survival was 99.8% for Stage I, 100% for Stage II, 91.9% for Stage III, and 48.9% for Stage IV (P < 0.0001). The frequency of remaining disease free also declined significantly with increasing stage (94.3% for Stage I, 93.1%for Stage II, 77.8% for Stage III, and 24.6% for Stage IV). The same patients also were staged applying six previously published classifications as appropriate for their tumor type. The predictive value of the NTCTCS Registry staging classification consistently was among the highest for disease specific mortality and for remaining disease free, regardless of the tumor type. CONCLUSIONS: The NTCTCS Registry staging classification provides a prospectively validated scheme for predicting short term prognosis for patients with thyroid carcinoma.

Serum growth hormone response to growth hormone-releasing hormone in non-obese and obese adults with hypopituitarism.

The clinical consequences of growth hormone (GH) deficiency (GHD) in adults have not been defined. Standard methods of measuring GH reserve in children may not be reliable in adults. In addition, obesity in normal adults diminishes GH responsiveness to provocative stimuli; this inhibition of GH release is reversed with pyridostigmine (PD). We investigated the use of GH-releasing hormone (GHRH) as a method to assess pituitary GH secretory reserve (as defined by peak GH response to GHRH) in both non-obese and obese (ie, > 115% ideal body mass index [BMI]) adults with hypothalamic-pituitary tumors. Nine non-obese patients (NOP) and 10 obese patients (OP) were studied with 11 non-obese controls (NOC) and 10 obese controls (OC). All study groups received GHRH (1 microgram/kg intravenous bolus) with blood sampling at -15, 0, 15, 30, 45, 60, 75, and 90 minutes. OC and OP received 120 mg PD orally 1 hour before GHRH injection. Mean serum GH responses in NOC and OC were significantly higher (P < .05) than those in NOP and OP, respectively, 15 minutes after GHRH injection, and remained so throughout the time-course of the test. Mean +/- standard error of the mean (SEM) peak GH level (microgram/L) was lower in NOP than in NOC (5.1 +/- 1.6 v 21.2 +/- 4.4, P < .01) and lower in OP than in OC (4.6 +/- 1.8 v 15.5 +/- 2.2, P < .01). Mean +/- SEM peak GH level was also lower in NOP than in OC (5.1 +/- 1.6 v 15.5 +/- 2.2 micrograms/L, P < .01).(ABSTRACT TRUNCATED AT 250 WORDS)

Differential changes in serum concentrations of androgens and estrogens (in relation with cortisol) in postmenopausal women with acute illness.

Previous studies of adrenal androgens and estrogens in critical illness were limited by measuring only selected sex steroids and by including men (who have confounding simultaneous changes in gonadal steroids). We evaluated relationships between changes in serum levels of cortisol (F), androgens, estrogens, and gonadotropins in 20 postmenopausal women with acute critical illness to determine if changes in adrenal androgens and estrogens paralleled gonadal axis suppression or adrenal stimulation. Two patterns of changes in sex steroids were observed. Admission serum levels of androstenedione (delta 4-A), estradiol, and estrone, like F, were increased compared to healthy controls (P < 0.0001). delta 4-A and estrone then decreased toward normal by day 5 in parallel with cortisol (r = 0.56 and 0.60). In contrast, admission serum dehydroepiandrosterone (DHEA) and DHEA-sulfate (DHEA-S) were not elevated and testosterone (T) was decreased in our patients compared to controls (P < 0.0005) in parallel with serum gonadotropin levels. Serum levels of DHEA and T continued to decrease by day 5 in parallel with gonadotropins. We conclude that in agonadal patients with acute critical illness, serum levels of DHEA-S and T are selectively decreased in relation to F, delta 4-A, and estrogens. The decreased serum T levels suggest inhibition of 17 beta-OH-dehydrogenase and/or increased aromatization to estradiol. The marked increase in serum estrogen levels also suggests increased aromatization. The absence of increases in DHEA and DHEA-S suggest enhanced activity of 3 beta-hydroxysteroid dehydrogenase and/or inhibition of C17,20-lyase activity of P-450c17. The clinical significance of this marked increase in the ratio of estrogens to androgens in acute illness requires further investigation.

Both hyper- and hypogonadotropic hypogonadism occur transiently in acute illness: bio- and immunoactive gonadotropins.

Previous reports of hypogonadotropic hypogonadism in critically ill men may not reflect the complexity of changes in the hypothalamic-pituitary-gonadal (HPG) axis during acute illness. We sampled blood throughout hospitalization in 55 men admitted to acute care units to delineate the spectrum of changes in circulating gonadotropin and sex steroid levels at the onset and during recovery from acute illness. Bioactive LH and FSH were measured in a subset of patients. Percent free testosterone was measured to assess changes in binding to sex hormone binding globulin. Medications and serum estrogen and prolactin levels were monitored as potential causes of hypogonadotropism. Sustained suppression of serum testosterone levels below the normal range occurred in 62% of men with varying diagnoses and disease severity. Percent free testosterone remained constant. Hypogonadotropism was observed in most men (60%) and occurred independently from head injury, surgery, medications, or hyperprolactinemia. In a subset of men (n = 16), LH and/or FSH rose transiently above the normal range. Bioactivity of both LH and FSH remained constant while serum testosterone levels decreased. In contrast to serum testosterone levels, mean serum levels of E1, E2 and androstenedione were not less than control values. We conclude that both primary and secondary hypogonadism occur transiently in acutely ill men and cannot be explained solely by medications, hyperprolactinemia, or hyperestrogenemia. Neither biopotency of gonadotropins nor binding of testosterone to SHBG change across the course of acute illness. The hypogonadism, often severe and prolonged, may contribute to the persistent catabolic state observed in many critically ill patients.

Comparison of serum free thyroxine measurements by chemiluminescence and equilibrium dialysis.

A study of 239 patients compared free thyroxine (FT4) measurements made by equilibrium dialysis (ED) with measurements made using the Magic Lite FT4 chemiluminescence (CI) immunoassay (Ciba Corning Immunodiagnostics). Patient groups: 41 normals; 27 hyperthyroid; 29 hypothyroid; 37 sick euthyroid; 10 chronic renal failure (CRF) and 25 pregnant patients; 13 oestrogen; 10 heparin; 12 salicylate; and 9 dilantin-treated patients; 3 lipaemic; 5 haemolysed; 6 hyperbilirubinaemic patients; 6 low thyroid binding protein (TBP) and 6 high TBP level patients. The two assays gave comparable results in most groups. Both assays tended to give elevated values in heparinized patients but FT4-ED results were more obviously affected. Pregnant patients and women on oral oestrogen had higher mean values with FT4-ED. In both assays the sick euthyroid and CRF patients had mean FT4 values similar to healthy euthyroid patients; the range of values in sick euthyroid and CRF patients was similar in both assays but wider than in healthy euthyroid patients. A supplemental study of 81 unselected acutely ill patients using FT4-Cl alone confirmed the wider range of values to be anticipated in sick euthyroid patients.

Autoimmune thyroid disease: immunological, pathological, and clinical aspects.

Autoimmune thyroiditis, most notably Hashimoto's thyroiditis, appears to be increasing in prevalence and is now more easily detected by sensitive laboratory tests and more invasive procedures such as fine needle aspiration. During the last decade, marked progress has been made in the understanding of these diseases. There is a greater awareness of the interaction between the humoral and cell-mediated arms of the immune system in autoimmune thyroiditis. Recent studies implicate a subpopulation of suppressor T lymphocytes which have an antigen-specific defect, resulting in their suboptimal interaction with the helper T lymphocytes and subsequent autoimmune manifestations. There is some evidence that thyroid epithelial cells which inappropriately express HLA-DR may enhance presentation of thyroid antigens to the immune system, possibly significant in the initiation or enhancement of the autoimmune response. The presence of various antithyroid autoantibodies allows the use of laboratory assays to confirm the clinical diagnosis and predict the results of treatment. There appears to be predisposing genetic factors in the development of autoimmune thyroiditis, with some geographical and racial differences. Environmental factors, most notably dietary intake of iodine, have also been implicated in the pathogenesis of Hashimoto's thyroiditis. Several animal models have been developed addressing such issues. Ongoing studies in the areas of postpartum thyroiditis and childhood thyroiditis are helpful in clarifying their relationship with Hashimoto's thyroiditis. Graves' disease and postpartum thyroiditis are being investigated as possible causes of postpartum depression. The association of Hashimoto's thyroiditis and carcinoma of the thyroid gland is still controversial, but its relationship with malignant lymphoma is now well accepted. Thus, although the pathogenesis of autoimmune thyroiditis remains elusive, there has been significant refinement of the clinical diagnosis, and immunological abnormalities of specific intrathyroidal lymphocytes have been identified. Hopefully, these new areas of knowledge will assist in the treatment of these diseases and in the prevention of the development of malignant lymphomas of the thyroid gland.

Percutaneous tetracycline instillation for sclerosis of recurrent thyroid cysts.

Of 770 hypofunctioning thyroid nodules subjected to needle biopsy since 1977, the fluid of 172 cysts (22%) was aspirated. The fluid, cytologically examined, showed no evidence of carcinoma except in three instances: papillary carcinomas that were subsequently treated surgically. Of the cystic lesions, 19 (11%) recurred and were treated by reaspiration and the instillation of tetracycline hydrochloride into the cyst cavity. In all but one instance this resulted in obliteration of the cysts (95%). This is a considerably higher rate of success than that reported from aspiration alone (40% to 85%). The follow-up period ranged from 1 to 42 months, with no recurrence or subsequent development of a thyroid nodule in these patients. The one recurrent cyst was reaspirated and reinstilled a second time, which resulted in great reduction in size. Not considered for this procedure were patients in whom the cytologic condition of the fluid was abnormal or indicative of a malignancy. Also excluded were those in whom there was incomplete decompression of the cyst, manifested by a persistent nodule. No patients had a history of head or neck irradiation or cervical lymphadenopathy, both contraindications for this procedure. No patients had grossly bloody aspirates. Recurrence of thyroid cysts after aspiration was not thought to be, in itself, a criterion for surgical selection. The effective use of tetracycline hydrochloride as a sclerosing agent in the treatment of malignant pleural effusions is well documented. Its action in producing symphysis is thought to be related to its low pH (2.0). This procedure is safe, simple, cost effective, and well tolerated by patients. It obviates the need for excision in patients who fulfill the aforementioned criteria.

The efficacy of simultaneous fine-needle aspiration and large-needle biopsy of the thyroid gland.

Fine-needle aspirations and cutting needle biopsies were performed on 415 patients with solitary thyroid nodules. All nodules were considered hypofunctioning by scintiscans. Specimens were adequate in 399 patients. Ten percent of these patients had thyroid cancers documented by thyroidectomies. Results from either or both procedures were positive or suspicious in 58 patients (14.5%). Thyroid cancers were found in 40 of the 58 patients (69%). Aspirates alone detected 34 of 40 cancers (85%) and biopsy specimens alone detected 30 of 40 cancers (75%). All thyroid cancers were detected when both results were considered together, demonstrating that needle aspirates and cutting needle biopsies are complementary. The difficulty of making definitive diagnoses of follicular neoplasms by needle aspirates or cutting needle biopsies is reemphasized.

The aging thyroid. Relationship between elevated serum thyrotropin level and thyroid antibodies in elderly patients.

The relationship of thyroid antibodies and the serum level of thyrotropin in older adults (over age 60) was studied to determine whether thyroid antibodies were a good clue to thyroid failure in elderly persons. Of those with thyroid failure, evidenced by clearly elevated serum thyrotropin values (more than 10 microU/ml), 67 percent had positive antimicrosomal antibody levels, a prevalence much greater (p less than 0.001) than that among those of comparable age with normal thyroid function (18 percent). Nevertheless, one third (33 percent) had thyroid failure without positive antimicrosomal antibody levels; this was true whether or not a low serum thyroxine value was present. Furthermore, of those with positive antimicrosomal antibody levels, most (68 percent) did not have thyroid failure. Thus, although positive antimicrosomal antibody levels occurred more often in elderly patients with thyroid failure than in those with normal thyroid function, a sizable fraction of those with thyroid failure did not have positive antimicrosomal antibody levels. Hence, measurement of thyroid antimicrosomal antibodies is not a good test of early thyroid failure in older patients; direct demonstration of a clearly elevated serum thyrotropin value is a better approach.

A thyrotropin radioimmunoassay kit evaluated vs two reference assays.

We evaluated a double-antibody radioimmunoassay kit for thyrotropin that includes calibrators prepared in a matrix of human serum and involves overnight nonequilibrium. Results were compared with those from two reference assays for thyrotropin. The range of within-assay CVs for the kit for thyrotropin values between 0.9 and 2.4 milli-int. units/L was 2.2 to 5.3%, that for between-assay CVs was 8.3 to 30%. The estimated minimum detectable concentration of thyrotropin was 0.6 milli-int. unit/L. We saw no cross reactivity with human choriogonadotropin by any of 48 sera from pregnant women. The original lot of serum specified as thyrotropin-free contained small but measurable amounts of thyrotropin; a second lot did not. Clinical data generated with the kit and the reference assays correlated well and were consistent with the clinical status of various categories of patients.

Five cases of parathyroid lipohyperplasia.

Lipoadenoma is the accepted diagnosis of a single enlarged parathyroid gland that contains large quantities of mature fat cells and focal myxoid stroma, all widely separating small parenchymal cell nests in patients with hyperparathyroidism. Here we are reporting, for the first time, on five cases of hyperparathyroidism in which all four parathyroid glands are enlarged and each gland is noted to have an admixture of fat and parenchymal cells. We will introduce the descriptive diagnosis of lipohyperplasia to name this condition and keep it in perspective with other forms of parathyroid disease. All five patients were women between the ages of 36 and 62 years who underwent neck exploration, at which time four enlarged light-tan parathyroid glands were observed. Three and one half gland resections were performed, and all patients returned to a normocalcemic state except one who had borderline serum hypercalcemia after operation. Most of the resected parathyroid glands weighed in the range of 100 to 200 mg. The largest measured gland weighed 820 mg. Parathyroid histology showed an admixture of mature fat cells with parathyroid parenchymal cells often in a 1:1 ratio. One patient who had renal failure exhibited a lower ratio of fat cells. Two patients had chronic lymphocytic thyroiditis that was severe enough to require synthetic thyroid hormone therapy. Two patients had a history of urinary tract infections. Three patients had hypertensive cardiovascular disease, and several patients had arteriosclerotic cardiovascular disease. One patient had diabetes mellitus, one had a history of pituitary adenoma, and one had polydipsia. All of these patients were first seen with parathyroid glands measuring an average of five times normal size, yet they showed the usual 50% fat/50% parenchyma pattern of normal mature parathyroid glands. This means that the enlarged glands contain a 500% increase in parathyroid tissue, justifying the diagnostic term "lipohyperplasia." This easily represents enough parathyroid tissue to generate excessive parathyroid hormone production. At this time, there is no explanation of the pathogenesis of lipohyperplasia or how it varies from other previously described forms of parathyroid hyperplasia.

Cushing's disease: management by transsphenoidal pituitary microsurgery.

Over 15 yr, 24 patients underwent transsphenoidal pituitary surgery for Cushing's disease with a median follow-up of 12 months. Cures included 7 patients with normal sella turcicas (6 microadenomas), 6 patients with focal depressions (6 microadenomas) of the sella (grade I), and 3 patients (3 adenomas) with enlarged sellas (grade II). Three patients with sella destruction (grades III and IV), 2 with normal sellas, and 1 with focal sella depression (grade I) were not cured. Two apparent cures (microadenomas) recurred. Tumor histology revealed 19 basophilic adenomas; electron microscopy (14 tumors) and immunochemical studies (10 tumors) revealed only ACTH cells. Circadian rhythm returned in 6 cured patients. Impotence (in 2), amenorrhea (in 7), and galactorrhea (in 3) resolved in affected cured patients. The major surgical complication was hemorrhage at the operative site (3 patients). Transsphenoidal pituitary surgery is a valuable method for managing Cushing's disease in many patients.

Cyclophosphamide in the management of advanced Graves' ophthalmopathy. A preliminary report.

Three patients with advanced Graves' ophthalmopathy were treated with cyclophosphamide. All had proptosis and diplopia. One patient had disabling iatrogenic Cushing's syndrome from steroid treatment of the ophthalmopathy and had undergone bilateral orbital decompression before cyclophosphamide therapy; steroids could not subsequently be withdrawn. When cyclophosphamide was administered to the cushingoid patient, withdrawal of glucocorticoid therapy was then possible. Diplopia completely resolved in two patients and improved in the third coincident with administration of cyclophosphamide. Deteriorating visual acuity resolved in one patient. Chemosis improved in the two affected patients. Proptosis was unchanged in all three patients. Cyclophosphamide deserves further study as a therapeutic agent in Graves' disease.