Nonthyroidal Disease Research Articles

Iodine contamination in subjects admitted to a general hospital.

Urinary iodine excretion was measured in 300 consecutive patients admitted to a general hospital for nonthyroidal disease. Iodine contamination (values greater than 300 micrograms l/g creatinine) was found in 29.3% of cases, mostly due to amiodarone and iodinated contrastographic agents. The source of iodine contamination was not detected in 23% of cases.

Do thyroid function tests independently predict survival in the critically ill?

We studied the ability of thyroid function tests to predict hospital survival in 116 critically ill patients and compared the results with independent predictions of survival made by critical care physicians. Eleven patients (9.5%) had clinically unsuspected hypothyroidism and were less likely to survive (p = 0.03). In patients critically ill with nonthyroidal disease, low T3, low FT3I, low T4, low FT4I, high TSH, and high T3U levels each showed significant correlation with nonsurvival (all p less than 0.02). Of these, however, only low T3 (p less than 0.001) and high TSH (p = 0.016) showed significant independent prediction of nonsurvival, and only low T3 (p = 0.011) added any significant independent prediction of nonsurvival beyond that made clinically by the group of critical care physicians.

Heat stress and hydrocortisone are independent stimulators of triiodothyronine-induced growth hormone production in cultured rat somatotrophic tumour cells

We have reported that, in cultured GC cells, the stress of incubation at 41 degrees C enhances thyroid hormone stimulation of growth hormone (GH) in a manner similar to the effects observed in a model of nonthyroidal disease in rats. Since glucocorticoids are potentially involved in stress responses both in vivo and in cell culture, we studied the role of glucocorticoid in the enhancement of (which are rat somatotrophic tumor cells) triiodothyronine (T3)-induced GH synthesis due to heat stress. Hydrocortisone addition increased T3-induced GH synthesis and GH mRNA content in cultured GC cells at both 37 degrees C and 41 degrees C. Depletion of glucocorticoid endogenous to serum supplement of the tissue culture medium did not prevent the enhancement of T3-induced GH synthesis that occurred during incubation at 41 degrees C. The levels and affinity of glucocorticoid cytosolic receptors were not enhanced during incubation at 41 degrees C. Lastly, no change in the sedimentation coefficient of the cytosolic glucocorticoid receptor or in its translocation into the nucleus occurred during incubation at 41 degrees C. Thus, the enhancement of T3-induced GH production in GC cells by heat stress appeared independent of the effect of glucocorticoids and not mediated through glucocorticoid receptors.

Anti-Thyroid Peroxidase Antibodies in Thyroid Disorders and Non-Thyroid Autoimmune Diseases

A new commercial method for measurement of anti-thyroid peroxidase (anti-TPO DYNOtest, Henning, Berlin) was evaluated in normal subjects and in patients with autoimmune thyroid and non-thyroid diseases, and compared to an immune fluorescence method for measurement of anti-microsomal antibodies (MicAb), and a radioimmunological method for quantifying thyroglobulin antibodies (TgAb). The majority of normal subjects had anti-TPO levels below 52 U/ml and patients with Hashimoto's thyroiditis had levels above 200 U/ml, with a good correlation to MicAb. In other autoimmune thyroid diseases the correlation was less pronounced. In non-thyroid autoimmune diseases MicAb showed falsely positive reactions in the presence of other autoantibodies, e.g. mitochondrial antibodies. The present study indicates that the anti-TPO method should probably replace measurements of MicAb for routine clinical use, thus providing a sensitive, precise, antigen specific method with the ability to reveal quantitative fluctuations. The study also indicates that TgAb could be abolished in routine diagnosis of autoimmune thyroid diseases and be reserved for special clinical situations, research purposes as well as measurement in sera before evaluation of serum thyroglobulin levels.

The Effects of Nonthyroid Disease and Drugs on Thyroid Function Tests

Serious nonthyroid illness and caloric deprivation, which so often accompany systemic illness, have diverse and still incompletely understood effects on thyroid hormone economy. We have discussed the pathophysiologic basis for the most common pattern of alterations in routine thyroid function tests: a decreased serum T3 concentration; normal or, in critically ill patients, a low total serum T4 level; and a normal free T4 concentration. Another, less frequent pattern (high total and free T4 with a normal serum T3) can be encountered transiently in the acutely ill medical or psychiatric patient. With the recent advent of sensitive assays for TSH and better methods for serum free T4, it is now possible to define more quickly and accurately the thyroid-metabolic status of most of these sick patients; the vast majority are euthyroid. Certain drugs confound the picture. The most important of these include dopamine and high-dose glucocorticoids, both of which suppress TSH secretion from the pituitary and may actually cause a state of central hypothyroidism. Other drugs have multiple effects on thyroid hormone indices (e.g., amiodarone). Knowledge of all of the ways in which systemic illness, starvation, and certain drugs may influence thyroid function tests is crucial in assessing the thyroid status of patients with serious nonthyroid disease.

Thyrotoxicosis in the Critically Ill

The thyrotoxic patient offers a considerable challenge to the critical care physician because the "obvious" diagnosis often will be a cardiac (or other nonthyroidal) problem, but the "correct" diagnosis will be an endocrinologic one. The importance of considering the diagnosis of thyrotoxicosis in any patient with tachyarrhythmias, new-onset congestive heart failure, weight loss, or change in mental status cannot be overstated. Treatment for presumed thyroid disease sometimes will have to be initiated prior to the availability of the results of diagnostic tests. Timely and appropriate treatment of the thyroid problem is vital for a successful outcome in treating patients with thyrotoxicosis.

Effect of nonthyroidal illness on serum thyroxine concentrations in cats: 494 cases (1988)

Summary We reviewed the medical records of 494 cats with a variety of nonthyroidal diseases in which serum thyroxine (T4) concentration was determined as part of diagnostic evaluation. The cats were grouped by category of disease (ie, renal disease, congestive heart failure, diabetes mellitus, focal neoplasia, systemic neoplasia, hepatopathy, inflammatory bowel disease, inflammatory pulmonary disease, miscellaneous diseases, or undiagnosed disease), degree of illness (ie, mild, moderate, or severe), survival (ie, lived, died, or euthanatized), and presence or absence of a palpable thyroid gland. The mean (± sd) serum T4 concentrations in all 10 groups of cats, which ranged from 10.5 ± 11.1 nmol/L in cats with diabetes mellitus to 18.7 ± 7.8 nmol/L in cats with focal neoplasia, were significantly (P < 0.001) lower than those of normal cats (27.0 ± 10.4 nmol/L). The number of ill cats with low serum T4 concentrations (<10 nmol/L) was highest in the cats with diabetes mellitus (59%), hepatopathy (54%), renal failure (48%), and systemic neoplasia (41%). When the serum T4 concentrations in cats with mild, moderate, and severe illness were compared, mean concentrations were progressively lower (21.3 ± 6.8, 14.8 ± 8.1, and 6.5 ± 5.8 nmol/L, respectively) as degree of illness increased. Severity of illness had a more significant (P < 0.001) effect in lowering serum T4 concentrations than did disease category. Mean serum T4 concentrations in the cats that died (7.8 ± 9.8 nmol/L) or were euthanatized (10.0 ± 7.0 nmol/L) were also significantly (P < 0.001) lower than those of cats that survived (15.2 ± 8.8 nmol/L). Of the 182 cats with low serum T4 concentrations, 74 (41%) died or were euthanatized. The 63 cats with palpable thyroid nodules had significantly (P < 0.001) higher mean serum T4 concentrations (21.7 ± 10.4 nmol/L) than the cats in which a thyroid nodule was not palpated (12.7 ± 8.1 nmol/L). Adenomatous hyperplasia of the thyroid gland was confirmed at necropsy in 2 cats with a palpable thyroid gland, and 4 cats that survived have subsequently developed overt hyperthyroidism. This suggests that, in some cats with early or mild hyperthyroidism, concomitant nonthyroidal disease may suppress serum T4 concentrations into the normal range.

Triiodothyronine (T3) Suppression Test

The purpose of this study was to develop a T3 suppression test to help in the diagnosis of mild hyperthyroidism in cats. We evaluated the response in circulating T4 concentrations to exogenous T3 (liothyronine) administration in 44 clinically normal cats, 77 cats with hyperthyroidism, and 22 cats with nonthyroidal disease. The test was performed by first collecting blood samples for basal serum T4 and T3 determinations, administering liothyronine at an oral dosage of 25 micrograms three times daily for seven doses, and, on the morning of the third day, again collecting serum samples for T4 and T3 determinations 2 to 4 hours after the seventh dose of liothyronine. The mean basal serum concentrations of T4 (53.1 nmol/L) and T3 (1.8 nmol/L) were significantly higher in the cats with hyperthyroidism than in the normal cats (T4 = 25.3 nmol/L, T3 = 1.3 nmol/L) and the cats with nonthyroidal disease (T4 = 29.5 nmol/L, T3 = 1.4 nmol/L); however, there was a great deal of overlap of basal values between the three groups of cats. Of the 77 cats with mild hyperthyroidism, 41 (53%) had serum T4 values and 55 (71%) had T3 values that were within the established normal ranges. After administration of liothyronine, mean serum T4 concentrations fell much more markedly in the normal cats and the cats with nonthyroidal disease than in the hyperthyroid cats.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of active immunization against L-triiodothyronine on serum thyrotropin levels and liver mitochondrial α-glycerophosphate dehydrogenase activity in rats: evidence for reduced hormone disposal to cells

Thyroid hormone binding autoantibodies have been observed in serum from patients affected by either thyroid disorders or autoimmune nonthyroid diseases. In most instances, no major involvement of thyroid status of these patients has been reported. However, some authors have attributed the occurrence of hypothyroidism to high capacity thyroid hormone binding autoantibodies. In order to verify such a hypothesis, the effect of these autoantibodies on serum TSH and liver mitochondrial alpha-glycerophosphate dehydrogenase has been investigated in rats, wherein circulating T3-binding antibodies (max cap 10(-12) 10(-8) to mol/l) were induced by active immunization. Starting from the 3rd week after antigen injection, the binding of 125I-T3 to serum immunoglobulins was progressively increased, with a peak at the 5th week. In immunized animals considered as a whole group, serum TSH levels were significantly higher and liver mitochondrial alpha-glycerophosphate dehydrogenase activity was significantly lower than in normal rats (m +/- SE: 145 +/- 15 vs 34 +/- 2 ng/dl and 0.0450 +/- 008 vs 0.0980 +/- 005 delta A. mg/prot/.min, respectively; p less than 0.001). The pattern of both TSH increase and alpha-glycerophosphate dehydrogenase decrease were similar to the evolution of 125-I-T3 binding to serum immunoglobulins, the maximal TSH values and lowest enzyme levels being observed at the 5th week after the beginning of the experiments.(ABSTRACT TRUNCATED AT 250 WORDS)

Differential Response to L-Triiodothyronine of Anterior Pituitary Growth Hormone Messenger Ribonucleic Acid (mRNA) and β-Thyrotropin mRNA in a Hypothyroid Walker 256 Carcinoma-Bearing Rat Model of Nonthyroidal Disease*

To continue our studies on the influence of T3 on TSH regulation in the Walker 256 carcinoma-bearing rat model of nonthyroidal disease, we measured the effect of T3 on pituitary content of beta TSH mRNA and rat (r) TSH in hypothyroid control (C) and tumor-bearing (T) rats. The effect of T3 on TSH regulation was compared to effects on GH mRNA and rGH in the same animals. mRNA content was normalized to a pool of pituitaries from euthyroid rats (= 1.0). beta TSH mRNA increased 18-fold in both hypothyroid C and T rats and then decreased similarly with increasing T3 infusion to a value of 0.1. GH mRNA content decreased to 0.11 +/- 0.01 in hypothyroid C rats, but to only 0.38 +/- 0.02 in T rats (P less than 0.001). The pituitary contents of GH mRNA and rGH in hypothyroid T rats was significantly greater than those in C rats at all T3 infusion rates. These data together with our previous report of decreased nuclear T3 in T rats suggest that regulation of beta TSH mRNA by T3 is intact in T rats, but occurs at a lower concentration of nuclear T3. In contrast, the GH mRNA response is enhanced, displaying differential regulation of these two T3-responsive gene products in this model of nonthyroidal illness.

Decreased anterior pituitary T3 nuclear receptors in a Walker 256 carcinoma-bearing rat model of nonthyroidal disease.

Rats bearing the Walker 256 carcinoma have decreased pituitary nuclear T3 but normal pituitary TSH content and response to experimental hypothyroidism. To elucidate further the role of T3 receptor occupancy and biological response in the tumor-bearing rat model of nonthyroidal disease, we measured the concentration of T3 nuclear receptors, rTSH and rGH and beta-TSH mRNA and GH mRNA in the anterior pituitary of euthyroid rats bearing the Walker 256 carcinoma. The abundance of T3 nuclear receptors was decreased in tumor-bearing rats and was associated with a decrease in mRNA content for beta-TSH and GH. alpha-tubulin mRNA was decreased to a comparable degree. The pituitary content of rTSH and rGH was, however, the same as in control animals. Since tumor rats have normal regulation of TSH secretion by thyroid hormone, the present findings suggest that TSH secretion in T rats is maintained by a lower T3 nuclear receptor occupancy than in controls. The decrease in beta-TSH mRNA may precede a decrease in TSH synthesis and changes in pituitary TSH stores. Since the decrease in GH mRNA was comparable to the decrease in alpha-tubulin mRNA, it does not appear to be specifically related to decreased T3 nuclear receptor occupancy. We conclude that, in the tumor-bearing rat model of nonthyroidal disease, decreases in beta-TSH mRNA occur despite a decreased T3 receptor occupancy. Both thyroid-dependent and thyroid-independent factors may be involved in regulating beta-TSH mRNA.

Heat shock of cultured GC cells enhances the level of triiodothyronine induced growth hormone (GH) and GH messenger ribonucleic acid.

We have previously proposed that the effects of heat shock on thyroid hormone-responsive rat pituitary tumor (GC) cells may be a model relevant to the in vivo effects of nonthyroidal disease on thyroid hormone action. To determine the effects of heat shock on thyroid hormone responses, GC cells (normally cultured at 37 C) were studied after incubation at 41 C. After 18 h at 41 C there was enhanced synthesis of proteins (mol wt, 70,000 and 90,000) considered to be universal markers of the cellular response to heat shock. Incubation at 41 C also resulted in a significant decrease in GC cell viability and (after 24 h) arrest of GC cell growth. However, the induction of GH synthesis by T3 was significantly enhanced in GC cells stressed by incubation at 41 C. The addition of 5 nM T3 to thyroid hormone-depeleted GC cells resulted in a significantly greater (P less than 0.001) accumulation of GH (2642 +/- 280 ng/18 h) during 41 C incubation than during 37 C incubation (1223 +/- 175 ng/18 h). The enhanced T3-induced production of GH was coincident with a proportional increase (P less than 0.05) in cellular GH mRNA determined by dot hybridization analysis. Thus, the stress of 41 C incubation elicits a heat shock response in GC cells characterized by decreased viability and growth arrest, but enhanced accumulation of GH mRNA in response to T3. Our recent report on the identical effects due to the stress of implantation of the Walker 256 carcinoma on T3-induced rat pituitary GH mRNA in vivo suggests that heat shock of cultured GC cells is a valid in vitro model of nonthyroidal disease.

Thyroid function tests in ageing and their relation to associated nonthyroidal disease.

Thyroid function tests of 179 euthyroid geriatric inpatients (83 +/- 6 yr) unaffected by acute diseases or malnutrition were investigated and compared with those of 76 ambulatory healthy younger subjects (42 +/- 13 yr). Elderly population was divided in three groups, respectively: group G I (n = 37, 65-78 yr), group G II (n = 64, 79-85 yr) and group G III (n = 78, over 85 yr). Severity-of-illness index of the patients was evaluated at entry in the study protocol. While total thyroxine (TT4), free triiodothyronine (FT3) and TSH levels remained unchanged, circulating total triiodothyronine (TT3) was significantly lower (113 +/- 32 vs 150 +/- 31 ng/dl, p less than 0.05) and free thyroxine (FT4) was significantly higher (12.4 +/- 2.7 vs 10.3 +/- 2.3 pg/ml, p less than 0.05) in aged people. Furthermore, TT3 decreased significantly from 130 +/- 36 in G I to 110 +/- 33 in G II and to 108 +/- 25 in G III (p less than 0.01), and FT4 increased progressively although not significantly in the same groups. A close correlation was found between TT3 and severity index in male observations only (r = -0.43, p less than 0.01), as well as between FT4 and severity index in both sexes (r = 0.51, p less than 0.001 for men, r = 0.21, p less than 0.01 for women). These data suggest that thyroid function tests have to be cautiously interpreted in a geriatric population, particularly in relation to the severity of the clinical state, and reference values should be determined for TT3 and FT4 in the ageing process.

3,5,3'-Triiodothyronine determines the viability of GC cells after heat shock.

The heat shock (HS) response is a characteristic disruption of protein synthesis which occurs in cells exposed to a variety of noxious stimuli. The effects of HS on thyroid hormone-responsive GC cells were studied in an attempt to devise an in vitro model for the adaptive changes in thyroid hormone action caused by nonthyroidal disease. HS enhanced GC cell synthesis of 70 K and 90 K proteins in a manner previously described as characteristic of the HS response in many tissues. A step-wise decrease in GC cell viability occurred when cells were exposed to 45 C for 10 to 35 min. HS (45 C, 20 min) resulted in a rapid decrease in binding of T3 to nuclear receptors. Two hours after HS, analysis of T3 binding to isolated nuclei showed a 50% fall in binding capacity (240 fmol/100 micrograms DNA) compared to non-HS control cells (540 fmol/100 micrograms DNA); no difference in dissociation constant (Kd) was observed. The effect of thyroid hormone on cell viability after HS was then determined. Thyroid hormone depletion (less than or equal to 0.02 nM T3) resulted in significantly (P less than 0.05) enhanced cell viability compared to cells cultured with physiological T3 (0.2 nM) after incubation at 45 C for intervals of 10-35 min. This inverse relationship between medium T3 content and cell tolerance of HS occurred over a wide range of T3 concentrations. Mean cell viability after exposure to 45 C for 20 min was 44 +/- 3% in T3-depleted cultures (less than or equal to 0.02 nM), 27 +/- 1% to 32 +/- 5% in cultures containing 0.07-0.5 nM T3, and 13 +/- 3% in cultures containing 5 nM T3. Our results thus characterize the response to HS in GC cells and the relationship of this response to medium T3. Similar to the effect of various nonthyroidal diseases on rat hepatocytes in vivo, HS resulted in a decrease in T3 nuclear receptors. Similar to the adverse effect of thyroid hormone on morbidity in animals with experimental diseases or injury, GC cell viability after HS was inversely related to medium T3 content. Thus the HS response in GC cells may be a valuable in vitro model relevant to the effect on thyroid hormone action caused by nonthyroidal disease.

頚部腫りゅう症例の検討 疾患統計と受診者の経路

We analysed 7, 307 outpatients who had visited our clinic in the Dokkyo University School of Medicine Hospital between 1974 and 1987. The number of patients with swelling of the neck was 346 (4.8%). Benign disease accounted for 218 cases (62.1%) and malignant disease 133 cases (37.9%). Malignant disease increased in patients over 50 years old, and exceeded the number of benign disease in those over 60 years old.There were 171 cases of thyroid disease (35 malignant cases and 136 benign cases) and 180 cases of non-thyroid disease. The latter included 44 cases of inflammation (24.4%), 105 cases of tumors (58.3%) and 31 cases of congenital disease and miscellaneous (17.2%). Out of 105 tumor cases, primary malignant tumors were found in 13 cases, metastatic tumors below the clavicle were in 8 cases, metastatic tumors above the clavicle were in 73 cases, primary benign tumors were in 7 cases and primary unknown malignant tumors were in 4 cases.In the frontal neck region, almost 80% of neck masses were benign disease and almost 90% were thyroid disease. In the lateral neck region, almost 60% were malignant disease. In the sternocleidomastoideus area, over 90% were metastatic tumors. The metastatic neck masses from the head and neck region were found in the upper and median parts of the neck. Metastatic tumors below the clavicle were found only in the lower part of the neck.Almost 70% of the patients (230 out of 334 cases) with cervical masses were referred to our department by other departments or institutes. The patients were referred to our department slightly more than one year after noticing the first symptoms. The patients were first treated or diagnosed by referring doctors for a period of 44 days. We wish to enlighten the general public and doctors in other departments about these cervical masses.

Enzyme immunoassay for antibodies to thyroxine in human serum using synthesized antibody as a calibrator

We prepared human IgG labeled with rabbit antibodies to thyroxine, and used it for the assay of antibodies to thyroxine in human serum as a calibrator. The assay system consisted of thyroxine labeled with β- D-galactosidase and a microcolumn containing goat antibodies to human IgG immobilized on Sepharose 4B. Each serum sample or standard serum containing human IgG labeled with anti-thyroxine antibodies was incubated with the enzyme-labeled thyroxine, and the reaction mixture was passed through the microcolumn. The column was washed to remove the unbound label, and enzyme activity of the bound label was assayed. The minimum detectable amount of the anti-thyroxine antibodies was 0.3 ng/assay tube. The analytical recoveries of human IgG labeled with the antibodies added to human serum were from 103% to 108%. We measured anti-thyroxine antibodies in human sera from 187 patients with autoimmune thyroid diseases. The antibodies were detected in 32 serum samples, and the values in positive samples varied from 0.31 to 2.31 μg/ml. On the other hand, in 58 sera from patients with non-thyroid diseases, the antibodies were detected in only two samples.

Diminished hepatic L-triiodothyronine nuclear receptor concentration in rats bearing the Walker 256 carcinoma is not due to a generalized decrease in nonhistone protein synthesis.

To determine whether the decrease in hepatic T3 nuclear receptors in rats bearing the Walker 256 carcinoma (T rats) is a consequence of nonthyroidal disease or related to a generalized decrease in hepatic protein synthesis, we measured hepatic T3 nuclear receptors and [3H]leucine incorporation into protein in control and T rats. Mean protein contents of liver homogenate, purified nuclei, 0.4 M NaCl extracts of nuclei, and nonhistone protein fractions were similar in control and T rats. In pooled data from three experiments, [3H]leucine incorporation into protein in these liver fractions was significantly increased (55-73%) in T rats. Since the precursor nonradioactive leucine pool, measured by a novel HPLC method, was similar in control and T rats, the increased incorporation of [3H]leucine in all liver fractions suggests an increase in protein synthesis in T rats. Additional measurements in the same rats (two experiments) indicated that hepatic nuclear receptors were decreased to 50% of the control value in T rats (P less than 0.001). Our findings suggest that in the Walker tumor model of nonthyroidal disease, decreased hepatic nuclear T3 receptors appears relatively specific and is not associated with a generalized decrease in hepatic protein synthesis.

Uptake and deiodination of T4 by anterior pituitary cells in the presence of free fatty acids and sera from patients with nonthyroidal disease

Uptake and deiodination of T4 by anterior pituitary cells in the presence of free fatty acids and sera from patients with nonthyroidal disease

Performance of direct equilibrium dialysis and analogue-type free thyroid hormone assays, and an immunoradiometric TSH method in patients with thyroid dysfunction

Direct equilibrium dialysis and analogue-type radio-immunoassays for free triiodothyronine (FT3) and free thyroxine (FT4) in serum were compared in 168 subjects with various states of thyroid function. A good diagnostic efficacy for FT3 and FT4 by either type of assay was observed in hyperthyroidism. In hypothyroidism the free thyroid hormone assays, particularly the FT3 assays, performed diagnostically less well, partly because patients with mild disease were included in the study. No significant differences in the percentages of misclassifications of thyroid dysfunction patients by corresponding dialysis and analogue assays were found. We observed a good linear correlation between dialysis and analogue methods for FT3 (r = 0.98) and FT4 (r = 0.97) in this study comprising out-patients not suffering from severe non-thyroidal disease, known from earlier studies in this and other laboratories to interfere in these assays. It is concluded that analogue assays may be used on out-patients in whom severe systemic diseases are less frequent than in hospitalized patients. There are, however, other limitations to the use of analogue assays than systemic diseases. We observed two euthyroid patients with thyroxine auto-antibodies causing very high FT4 concentrations as determined by analogue assay; their dialysable FT4 concentrations were normal. We also tested a recently developed immunoradiometric serum TSH assay, which was found to perform well in primary hypo- and hyperthyroidism. Serum TSH was elevated in one patient hyperthyroid because of a TSH-producing pituitary adenoma, and within the reference limits in a patient with secondary hypothyroidism.

Application of an immunoradiometric assay for thyrotrophin in evaluation of thyroidal and nonthyroidal disease states

We evaluated an immunoradiometric assay for serum TSH (IRMA-TSH) and compared it with established indices of thyroid function in 208 patients with either frank thyroid disease, conditions of abnormal thyroxine protein binding, or conditions which are known to produce discordant thyroid results, viz. pregnancy or estrogen treatment and nonthyroidal illness (NTI). As expected, a wide scatter of TSH results was found in treated thyroid disease: 53 patients (53%) from both groups (initially hypo- or hyperthyroid) had TSH values less than 0.5 mU/l, but only one was considered to be clinically mildly toxic. All the pregnant, estrogen-treated and abnormal thyroxine protein-binding patients had TSH results within the reference range 0.5-6.0 mU/l, except one familial dysalbuminemic hyperthyroxinemia (FDH) patient (T4 level, 178 nmol/l) lost to follow-up with a TSH level of 8.5 mU/l, and one euthyroid, low-TBG patient being treated inappropriately with thyroxine, with a TSH level less than 0.5 mU/l. All the untreated thyrotoxic patients and 15 (26.3%) of the NTI patients had TSH results of 0.5 mU/l or lower. Because of this high incidence of low TSH levels in euthyroid NTI, we cannot recommend this IRMA-TSH as the initial test of thyroid function.