On the basis of a 3.7% prevalence of hypothyroidism in the United States (NHANES 1999–2002) (2), as many as 10 million people in the United States, and perhaps as many as 200 million people worldwide, are taking thyroid hormone and are therefore at risk for subclinical thyrotoxicosis because of either overzealous replacement therapy or intentional suppressive therapy. When sensitive methods for measuring serum TSH first became available, an estimated 40% to 50% of patients
taking T₄ replacement therapy had subnormal serum TSH concentrations (3), but the proportion has diminished as a result of the use of these assays to monitor therapy and recognition of the potential dangers of over-replacement. Nonetheless, in a 2010 population-based study, 19.8% of patients taking thyroid hormone replacement had a TSH less than 0.4 mU/L (4). The prevalence of low serum TSH was even higher (41%) in individuals over age 65 years (5). Patients with thyroid cancer, thyroid nodules, or multinodular or diffuse goiter who are treated with T₄ to suppress serum TSH levels to below normal have subclinical (if not overt) thyrotoxicosis. Therapy with thyroid hormone to suppress serum TSH should be considered only when the benefits exceed the potential risks of subclinical thyrotoxicosis.

The common causes of endogenous subclinical thyrotoxicosis include multinodular goiter with autonomy, autonomously functioning thyroid adenomas, Graves’ disease, and painless (lymphocytic) or subacute (granulomatous) thyroiditis (6).

In a study of patients over age 55 years, hyperthyroidism from autonomy in a nodular goiter was subclinical in 57% of patients, while hyperthyroidism due to Graves’ disease was subclinical in only 6% of patients (7). The prevalence of autonomously functioning adenomas and nodular goiter varies considerably from one region of the world to another. This variation is the result of differences in dietary iodine intake and perhaps also genetic differences. For example, in a US study, the ratio of patients with thyrotoxicosis caused by an autonomously functioning adenoma to those with thyrotoxicosis caused by Graves’ disease was 1:50 (8), whereas in Switzerland it was 1:2 (9). Thus in a European study, 22% of patients with nodular goiter had subclinical thyrotoxicosis, and 28% had autonomous areas on thyroid radionuclide imaging (10). Subclinical thyrotoxicosis also occurs in patients with mild Graves’ disease (11), or Graves’ eye disease (12), and can persist following treatment. In one study, about 20% of patients had a chronically suppressed serum TSH after a course of antithyroid drugs or surgery, and were more likely to relapse than a comparable group with normal serum TSH levels (13).

The thyrotropic activity of high serum chorionic gonadotropin concentrations during early pregnancy may cause a transient increase in serum free T₄ and T₃ concentrations, usually within the normal range, and low serum TSH concentrations in approximately 10% of pregnant women (14). This transient subclinical hyperthyroidism may be considered normal physiology; it is not associated with adverse pregnancy outcomes (15).

The evaluation of patients with subclinical thyrotoxicosis is similar to that of overt thyrotoxicosis (see Chapter 31), except, as noted above, autonomy in a nodular goiter is more common than Graves’ disease when hyperthyroidism is minimal (7). Free T₃ should be measured if a reliable assay is available, since 3 of 148 patients (2%) with apparent subclinical thyrotoxicosis had free T3 toxicosis (16). Many defer definitive evaluation unless treatment is being considered. The management guidelines of the American Thyroid Association and the American Association of Clinical Endocrinologists recommend a radioactive iodine uptake, and scan when nodularity is present, to assess the etiology of thyrotoxicosis (1). One or more focal areas of increased uptake indicate an autonomous nodule or autonomy within an adenomatous goiter; diffuse uptake suggests Graves’ disease; and in the absence of an iodine load, nearly absent uptake suggests painless thyroiditis, subacute thyroiditis, factitious ingestion of thyroid hormone, or struma ovarii. Alternatively, Graves’ disease may be diagnosed by measuring TSH receptor antibodies.

There are other causes of low serum TSH concentrations, but many of the patients have abnormal serum T₄ and T₃ concentrations (Table 33.1). Patients with pituitary or hypothalamic disease who have hypothyroidism may have low serum TSH concentrations, and the concentrations are often low in patients with severe nonthyroidal illness, especially those receiving glucocorticoid or dopamine therapy (17). Following treatment and during resolution of overt thyrotoxicosis, low serum TSH concentrations may be associated with low or normal serum T₄ and T₃ concentrations due to delayed recovery of the pituitary–thyroid axis (18). Finally, epidemiologic data show that the distribution of serum TSH levels is shifted to the left in black persons in comparison to Caucasians (19) and in smokers (20). It is therefore possible that a mildly subnormal serum TSH could be normal for certain black individuals and cigarette smokers, and there is a risk of potential misdiagnosis of mild subclinical hyperthyroidism in these persons.

The negative feedback relationship between serum TSH and free T₄ (and T₃) concentrations is log-linear. Therefore, very small increases in endogenous thyroid hormone production, or only slightly excessive doses of the exogenous thyroid hormone, decrease serum TSH concentrations to below the normal range (31) (see Regulation of Thyrotropin Secretion, Chapter 10D). Groups of patients with low serum TSH concentrations who receive slightly excessive doses of T₄ have serum T₄ and T₃ concentrations greater than the mean values for normal subjects but within the respective normal ranges. Presently, most clinicians accept serum TSH measurements as the most sensitive indicator of thyroid hormone status (in the absence of pituitary or hypothalamic disease). Several findings, however, raise the question of whether this is always the case. In laboratory animals, intrapituitary deiodination of T₄ to T₃, a reaction that is catalyzed primarily by type 2 T_4–5′-deiodinase (see Chapter 7), contributes more of the T₃ bound to the T₃ nuclear receptors, and serum T₃ contributes less, compared with most other organs (32). Hence, the extent of T₃-induced inhibition, or lack thereof, on TSH secretion, may vary somewhat independently of the serum T₃ concentration. Polymorphisms in the type 2 T_4–5′-deiodinase with reduced activity have been described (33). In addition, the concentrations of the different isoforms of the T₃ nuclear receptor are different in the pituitary and peripheral organs (see Chapter 8), and these concentrations are altered discordantly in states of thyroid hormone excess or deficiency (34). Some euthyroid elderly patients with low serum T₄ and normal serum TSH concentrations appear to have a decreased set point for T₄– and T₃-mediated feedback inhibition of TSH secretion (35). Thus, it is possible that serum TSH levels do not always accurately reflect thyroid function in peripheral tissues.

Thyroid hormone has a direct resorptive effect on bone, and overt thyrotoxicosis is associated with increased bone resorption and, to a lesser extent, increased bone formation and an increase in fracture rate (36). While some data suggest that TSH has a direct role inhibiting both osteoclasts and osteoblasts (37), other studies suggest that bone loss in thyrotoxicosis is primarily mediated by thyroid hormone interaction with its alpha receptor (38). Cortical bone is affected more than trabecular bone. Patients with a history of thyrotoxicosis have a 1.5-fold increased risk of osteoporosis (39) and a 1.8-fold increased risk of hip fracture (40). Even in patients with normal thyroid function, lower bone density (41,42) and vertebral fracture (43) correlate with TSH values in the lower portion of the normal range.