Abstract & Commentary
Source: Surks MI, et al. JAMA. 2004;291:228-238.
Synopsis: Data supporting associations of subclinical thyroid disease with symptoms or adverse clinical outcomes or benefits of treatment are few.
A national group of 13 experts participated in a consensus conference sponsored by The Endocrine Society to evaluate the literature on subclinical thyroid disease in order to address the question of treatment. This effort was stimulated by the obvious differences in conclusions and recommendations regarding this clinical problem in the medical literature. The literature was evaluated according to the evidence-based standards established by the US Preventive Services Task Force. The results of this exercise can be presented as answers to a series of questions:
What is the definition of subclinical hypothyroidism?
- A TSH above the upper limit of the reference range, 0.45-4.5 mIU/L, when free T4 is normal;
- Other causes have been ruled out: thyroid medication adjustments, recent illness.
What is the definition of subclinical hyperthyroidism?
- A TSH below the lower limit of the reference range when free T4 and T3 are normal; usually nondetectable to 0.40 mIU/L;
- Other causes have been ruled out: overadministration of thyroid medication, pregnancy, recent medications (especially glucocorticoids or dopamine).
What is the prevalence of subclinical thyroid disease?
- Subclinical hypothyroidism is present in 4-8.5% of US adults, increases with age, is less common in blacks, and is present in up to 20% of women older than age 60. About 2-5% will progress each year to overt hypothyroidism;
- Subclinical hyperthyroidism is less common, occurring in about 2% of the US population without known thyroid disease. It is more common in women and blacks. Progression to overt hyperthyroidism is essentially limited to those individuals with TSH levels lower than 0.1 mIU/L.
What are the consequences of subclinical hypothyroidism?
Possible consequences include progression to overt hypothyroidism, atherosclerosis, cardiac dysfunction, elevated total cholesterol and LDL-cholesterol, and neuropsychiatric symptoms. None of these risks have been definitively established by appropriate clinical trials, although sufficient evidence exists to be concerned. Similar reservations exist regarding treatment. Nevertheless, the panel recommended treatment if the TSH is elevated and the free T4 is below the range of 0.8-2.0 ng/dL. Support for treatment increased with rising levels of TSH, and treatment for a TSH above 10 mIU/L carries a reasonable prospect for improvement in symptoms and the lipid profile. The panel recommended follow-up evaluation every 6-12 months when the TSH is 4.5-10 mIU/L, unless symptomatic complaints indicate a need for treatment. There is evidence to suggest that subclinical hypothyroidism carries risks for fetus and mother during pregnancy. TSH levels should be measured every 6-8 weeks during pregnancy and maintained within the normal range.
What are the consequences of subclinical hyperthyroidism?
Possible consequences include progression to overt hyperthyroidism, atrial fibrillation, cardiac problems, reduced bone density and fractures, and neuropsychiatric symptoms. Good evidence indicates an increase in atrial fibrillation, especially in older women, when the TSH is less than 0.1 mIU/L, and suggestive with levels of 0.1-4.5 mIU/L. Treatment lowers the risk of this problem. The panel recommended treatment for TSH levels less than 0.1 mIU/L, especially to avoid bone loss and atrial fibrillation in elderly women or in those at increased risk for osteoporosis and heart disease. With TSH levels 0.1-4.5 mIU/L, treatment is indicated only in older individuals to avoid cardiac complications.
Comment by Leon Speroff, MD
The recommendations of this expert consensus panel are handicapped by the ultra-conservative position these exercises ultimately embrace. The influence of the "evidence-based movement" proves to be an inhibitory force in the exercise of clinical judgment. I would argue that this process provides little assistance for the clinician who, when all is said and done, is left with making a clinical decision for the individual patient. But then that is as it should be. If these decisions were clear-cut, reached by simply following an algorithm, there would be no need for clinicians experienced in the art and science of medicine.
For example, the panel recommended against population screening for thyroid disease, even in women planning to become pregnant. The diagnosis of subclinical thyroid disease is a laboratory diagnosis, and it seems to me that the relatively inexpensive measurement of TSH is warranted to screen our patients as they age, to make sure thyroid medication is given in the right dose and to avoid fetal and maternal consequences in pregnancy.
A simplified conclusion is possible: Confirm abnormal levels of TSH, adding a measurement of free T4 if hypothyroid and both free T4 and free T3 if hyperthyroid. Treat patients with TSH levels higher than 10 mIU/L or lower than 0.1 mIU/L. Follow those patients every 6-12 months with levels in the midrange of abnormal because in some cases, the values return to normal. Remember that it takes 8 weeks for the laboratory values to stabilize after changes in thyroid medication. In patients with subclinical hypothyroidism, it is worthwhile to measure thyroid antibodies; the presence of thyroid antibodies increases the risk of overt hypothyroidism at the rate of approximately 20% per year and increases the risk of miscarriage. During pregnancy, treatment should be guided by TSH values, obtained every 2-3 months.
A good reason to treat subclinical hypothyroidism is to avoid the development of a goiter. Furthermore, some patients in retrospect (after treatment) recognize improved physical and mental well-being. Patients with subclinical hypothyroidism have alterations in energy metabolism in skeletal muscle.1 An improvement in impaired cognitive function and emotional behavior has been documented with thyroxine treatment of subclinical hypothyroidism.2 Patients with an abnormal cholesterol-lipoprotein profile can show a rapid improvement with thyroxine treatment.3-5 About 10% of elderly women have subclinical hypothyroidism, a strong risk factor for coronary heart disease.6
TSH values below 0.1 mIU/L are regarded as nondetectable, and patients with overt hyperthyroidism usually have undetectable TSH. Subclinical hyperthyroidism is half as common in older people as subclinical hypothyroidism (excluding the most common cause, treatment with excessive doses of thyroxine). Keep in mind that the dose of thyroxine required to treat hypothyroidism declines with age (because of the decrease in metabolic clearance with age); all patients being treated with thyroid hormone should have their TSH levels assessed every year. Atrial fibrillation is a common cardiovascular problem associated with subclinical hyperthyroidism.7 If subclinical hyperthyroidism persists, it should be treated, especially in postmenopausal women, because of the cardiac complications and the loss of bone associated with excess thyroid hormone.8 TSH levels that are low but not undetectable (0.1-0.5 mIU/L) need not be treated, but TSH measurement is warranted every 6 months. Progression to overt hyperthyroidism is uncommon.
Although an increased prevalence of antithyroid antibodies and antinuclear antibodies has been observed in women with recurrent pregnancy loss, their relevance is uncertain because neither predicts subsequent pregnancy outcome, and there is no logical and proven effective treatment to offer.9-11 Tests to detect antinuclear and antithyroid antibodies have no clinical use in euthyroid women with recurrent pregnancy loss.
Risk of pregnancy loss may be increased for women with uncorrected overt or even subclinical hypothyroidism.12 Mild or subclinical disease generally has not been considered as having important clinical consequences.13,14 However, the results of a study of pregnancy outcomes in women with hypothyroidism challenge that notion. The incidence of pregnancy loss was very low in treated hypothyroid women having normal thyroid indices but markedly increased in women with elevated thyroid-stimulating hormone (TSH) levels, including both women with untreated subclinical disease and those with overt disease who received inadequate exogenous thyroid hormone replacement.12 These observations suggest that subclinical hypothyroidism may not be entirely benign and further justify earlier recommendations to include TSH screening in the evaluation of women with recurrent pregnancy loss.
1. Monzani F, et al. J Clin Endocrinol Metab. 1997;82:3315-3318.
2. Monzani F, et al. Clin Investigator. 1993;71:367.
3. Kinlaw WB. Endocrinologist. 1995;5:147.
4. Danese MD, et al. J Clin Endocrinol Metab. 2000;85:2993-3001.
5. Caraccio N, et al. J Clin Endocrinol Metab. 2002;87:1533-1538.
6. Hak AE, et al. Ann Intern Med. 2000;132:270-278.
7. Sawin CT, et al. N Engl J Med. 1994;331:1249.
8. Sgarbi JA, et al. J Clin Endocrinol Metab. 2003;88:1672-1677.
9. Fausett MB, et al. Endocrinol. 2000;18:379-392.
10. Esplin MS, et al. Am J Obstet Gynecol. 1998;179:1583-1586.
11. Rushworth FH, et al. Hum Reprod. 2000;15:1637-1639.
12. Abalovich M, et al. Thyroid. 2002;12:63-68.
13. Montoro M, et al. Ann Intern Med. 1981;94:31.
14. Roti E, et al. J Clin Endocrinol Metab. 1996;81:1679-1682.
Dr. Speroff is Professor of Obstetrics and Gynecology, Oregon Health Sciences University, Portland Ore.