Upper limit of normal serum thyroid-stimulating hormone: a moving and now an aging target?

Serum TSH measurement is the most sensitive screening test for thyroid dysfunction and for the diagnosis of subtle forms of hyperand hypothyroidism (1). Serum TSH has a loglinear relationship with circulating thyroid hormone levels; thus a small change in peripheral thyroid hormone concentrations, even within the normal laboratory reference range, may result in an increase or decrease in serum TSH outside its normal range (2). The laboratory reference range is usually chosen by determining the 95% confidence limits of a population of individuals free of known thyroid dysfunction. By this method, 2.5% of normal individuals may have high serum TSH values. However, if unrecognized thyroid failure is present in more than 2.5% of the population, mild hypothyroidism will be underdiagnosed. Thus, it has been proposed that more strict selection criteria should be the basis for determination of normal values (3). In interpreting serum TSH levels, one should consider physiological variations as well as occult thyroid disease. Values are highest in the early morning and lowest in the afternoon. This diurnal variation may be affected by depression, bipolar disorder, and night shift work (4–6). Serum TSH may be slightly higher in morbid obesity and may be reduced after weight loss (7). There are other causes of high TSH not associated with thyroid failure, such as heterophile antibodies (i.e. assay interference), recovery from nonthyroidal illness, thyroid hormone resistance, TSH-producing pituitary tumors, TSH molecules with lower biological activity, and certain cases of hypothalamic-pituitary disorder, when secretion of TSH with altered biological activity may occur. However, these conditions are uncommon, and most patients with persistently elevated TSH have autoimmune thyroid disease (1). It has been suggested that when individuals with thyroid autoantibodies, goiter, or a strong family history of thyroid disease are excluded, the 95% TSH reference range shrinks to between 0.3 and 2.5 or 3.0 mIU/liter (3). It has been argued that such a “refined normal” range is a better reflection of “thyroid health” than a standard population-based reference range (3) and that values between 2.5 and 4.5 mIU/liter predict future hypothyroidism (8). Consequently, a number of professional organizations (9), including the American Association of Clinical Endocrinologists (AACE), have endorsed lowering the upper TSH reference range to 3.0 mIU/liter, with the therapeutic target for T4 therapy below that level (9). In contrast to the controversy about subclinical hypothyroidism and serum TSH values at the upper limit of normal, there is no controversy about TSH levels below the lower limit of normal. One reason is that the number of individuals with serum TSH between 0.3 and 0.4 mIU/liter is not very high, and there is not a skewed TSH distribution at the lower end of serum TSH distribution as there is at the high end. There is also no controversy about management of subclinical hyperthyroidism because of strong evidence for adverse effects on bone health and cardiovascular system (10). The importance of having a consensus on the normal TSH range is that some authors have recommended routine screening for thyroid disease after age 35 (11) and other societies for women above age 50 (12). Screening of pregnant women and women anticipating pregnancy has also been proposed (13). With TSH screening and physician awareness of thyroid disorders, a large number of individuals, up to 10% of the tested population, will have serum TSH values above the level of 4.5 mIU/liter (14). A lowering of the upper limit of normal for TSH may result in a tripling of the frequency of abnormal results for persons over 50 yr old (15). The issues of screening (12), the treatment of subclinical hypothyroidism, and normal range for serum TSH (16) are subjects of intense debate within the thyroid community (16–21). Treatment of minimally elevated TSH between 4.5 and 10 mIU/liter has been suggested by national societies (19), and yet a consensus panel has found no evidence in favor of routine therapy of patients at these levels of TSH (21). However, there is consensus for T4 treatment of patients with elevated serum TSH levels above 10 mIU/liter (22). A common understanding among thyroidologists has been that 15–20% of older individuals have serum TSH levels above normal due to a higher prevalence of mild autoimmune hyroid disease in this age group. In this issue of the Journal, Surks and Hollowell (22) offer an alternative explanation. These authors examined the age-specific distribution of serum TSH and antithyroid antibodies in both National Health and Nutrition Examination Survey (NHANES) III (1988–1994) and NHANES (1999–2002). The authors obtained similar results in both NHANES groups. In NHANES III, there were 16,533 disease-free individuals not reporting thyroid disease, goiter, or thyroid-related medications. After exclusion of medications such as lithium and estrogens and individuals with positive thyroid peroxidase or thyroglobulin antibodies and overt hyperthyroidism or hypothyroidism, 1344 remained and were designated reference population (23). NHANES (1999–2002) included 4392 individuals. Surks and Hollowell (22) analyzed the data for different age groups, categorized by TSH levels: 0.4–2.5 mIU/liter (proposed normal by some national societies); 2.4 to 4.5 JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community. 0021-972X/07/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 92(12):4560–4562 Printed in U.S.A. Copyright © 2007 by The Endocrine Society doi: 10.1210/jc.2007-2285