Current issues of ACP Journal Club are published in Annals of Internal Medicine


Thyroxine decreased TSH-receptor antibodies and the risk for recurrence of hyperthyroidism in Graves disease

ACP J Club. 1991 July-Aug;115:17. doi:10.7326/ACPJC-1991-115-1-017

Source Citation

Hashizume K, Ichikawa K, Sakurai A, et al. Administration of thyroxine in treated Graves' disease. Effects on the level of antibodies to thyroid-stimulating hormone receptors and on the risk of recurrence of hyperthyroidism. N Engl J Med. 1991 Apr 4;324:947-53.



To evaluate the effectiveness of thyroxine in decreasing both the level of antibodies to thyroid-stimulating hormone (TSH) receptors and the rate of recurrence of hyperthyroidism in patients with hyperthyroidism caused by Graves disease after thyroid hormone secretion had been normalized by methimazole.


Randomized, single-blind, placebo-controlled trial of up to 4.5 years duration.


Not described.


109 patients (85 women and 24 men, ranging in age from 17 to 58 years) with untreated hyperthyroidism caused by Graves disease were enrolled in the study. All patients had diffuse goiter, an increased basal metabolic rate, increased thyroidal uptake of 123I, and increased serum levels of thyroxine, tri-iodothyronine, and antibodies to TSH receptors. All patients were treated with methimazole, 10 mg every 8 hours, and were euthyroid by 6 months.


At 6 months, patients were stratified into those with high and low TSH-receptor antibody levels. Each group was randomly assigned to 100 mg of thyroxine and 10 mg of methimazole once daily (n = 60) or to placebo and 10 mg of methimazole once daily (n = 49). One year later, methimazole therapy was discontinued; patients continued taking thyroxine or placebo for an additional 3 years.

Main outcome measures

The TSH-receptor antibody level and the recurrence of hyperthyroidism.

Main results

The TSH-receptor antibody levels (±SD) decreased from 28% ± 10% to 10% ± 3% after 1 month of treatment with thyroxine and methimazole but did not change in the 2 groups of patients receiving placebo and methimazole (P < 0.01). After withdrawal of methimazole, the TSH-receptor antibody levels decreased further in patients receiving thyroxine (from 6.6% ± 3.2% to 2.1% ± 1.2%) but increased in patients receiving placebo (from 9.1 ± 4.8 to 17.3 ± 5.8) (P < 0.01). Within 3 years of discontinuing methimazole therapy, hyperthyroidism recurred in 1 patient (2%) receiving thyroxine and in 17 patients (35%) receiving placebo (P < 0.001). {This absolute risk reduction of 33% means that 3 patients would need to be treated with thyroxine (compared with placebo) to prevent 1 additional recurrence of hyperthyroidism within 3 years of discontinuing methimazole, 95% CI 2 to 5; the relative risk reduction was 95%, CI 73% to 99%.}*


After patients with Graves disease are rendered euthyroid, administration of thyroxine decreased both the level of TSH-receptor antibodies and the risk for recurrence of hyperthyroidism.

Source of funding: Not stated.

Address for article reprint: Dr. K. Hashizume, Department of Geriatrics, Endocrinology, and Metabolism, Shinshu University School of Medicine, Matsumoto 390, Japan.

*Numbers calculated from data in article.


The low recurrence rate for hyperthyroidism achieved in this study is startling. The combination of antithyroid drug with thyroid hormone was used in the 1960s to prevent hypothyroidism and obviate the need for measuring thyroid hormone levels to assess the control of the hyperthyroidism; sufficient antithyroid drug was given to block thyroid hormone production and enough hormone was given to maintain a euthyroid state. The combination was stopped after about 1 year of therapy. Recurrence rates were no better than with an antithyroid drug given alone in the lowest dose to control the hyperthyroidism.

The new wrinkle in this Japanese study is the continuation of thyroxine for 3 years after stopping methimazole. No one has reported previously that addition of thyroid hormone nearly eliminates recurrence of hyperthyroidism: Only 1 of 60 patients had a recurrence in this study. It is almost too good to be true. What is the explanation? The investigators have shown that the TSH-receptor antibody levels were much lower in those taking thyroxine than in those receiving placebo. They speculate that suppression of TSH secretion by administration of thyroxine prevents the thyroid from releasing cell-surface components, such as the TSH receptor, and the reduction of this antigen lowers TSH-receptor antibody levels. Although their assay for receptor did not measure thyroid-stimulating antibody, there is a close correlation between TSH- displacing activity (which they measured) and TSH-stimulating antibody.

Will this article change my treatment strategy? Yes, I shall try to emulate the excellent results of these investigators by prescribing thyroxine with antithyroid drug for 1 year after controlling the hyperthyroidism followed by 3 years of thyroxine alone. However, to confirm that this treatment results in a very low recurrence rate, we must embark on a similar controlled study on this side of the Pacific.

Jerome M. Hershman, MD
West Los Angeles Veterans Affairs Medical CenterLos Angeles, California, USA