Scope and Purpose

An estimated 5 percent of women and 3 percent of men in the U.S. have subclinical thyroid dysfunction,¹ and about 0.5 percent of the population may have undiagnosed overt thyroid disease.^2,3 In some studies, subclinical hypothyroidism is associated with increased risk for coronary artery disease^4,5 and congestive heart failure,⁶ and subclinical hyperthyroidism with increased risk for all-cause and coronary heart disease mortality, atrial fibrillation,⁷ decreased bone density,⁸ and potentially fractures.⁹ Overt thyroid disease is associated with negative cardiovascular, musculoskeletal, dermatologic, gastrointestinal, and other effects, but clinical manifestations are highly variable and depend on the severity of the thyroid abnormality. Thyroid dysfunction represents a continuum from subclinical dysfunction to overt disease.

Approximately 2 to 5 percent of persons with subclinical hypothyroidism and 1 to 2 percent of those with subclinical hyperthyroidism develop overt thyroid disease.¹⁰ However, as much as 40 percent of patients with subclinical thyroid dysfunction may revert to normal when followed over time.^11-13

In 2004, the U.S. Preventive Services Task Force (USPSTF) found insufficient evidence to recommend for or against screening for subclinical thyroid dysfunction in asymptomatic nonpregnant adults:^1,14,15

• The USPSTF found fair evidence that the thyroid-stimulating hormone (TSH) test can detect subclinical thyroid dysfunction in persons without symptoms, but poor evidence that treatment improves clinically important outcomes in screen-detected adults.
• Although the yield of screening is greater in certain high-risk groups (e.g., postpartum women, persons with Down syndrome, and the elderly), the USPSTF found poor evidence that screening these groups leads to clinically important benefits.
• There is the potential for harm caused by false-positive screening tests; however, the magnitude of harm is not known.
• There is evidence that overtreatment with levothyroxine occurs in a substantial proportion of patients, but the long-term harmful effects of overtreatment are not known.
• As a result, the balance of benefits and harms of screening in asymptomatic adults could not be determined.

The 2004 recommendation did not address effects of screening in or treatment of patients with undiagnosed overt thyroid disease.

A contemporaneous systematic review conducted for the American Thyroid Association, the American Association of Clinical Endocrinologists, and the Endocrine Society reached similar conclusions.⁸ Nonetheless, thyroid medication prescription rates in the United States have increased dramatically, from an estimated 49.8 million in 2006 to 70.5 million in 2010.¹⁶ Among community-dwelling persons older than age 65 years with subclinical hypothyroidism, the proportion taking thyroid hormone has more than doubled, from 8.1 to 20 percent from 1989 to 2005.¹⁷

This report was commissioned by the USPSTF in order to update its 2004 recommendation on screening for thyroid disease. It builds upon a 2011 Comparative Effectiveness Review funded by the Agency for Health Care Research and Quality¹⁸ and prior USPSTF reviews on identification and treatment of subclinical thyroid dysfunction.^1,15 Prior to updating its 2004 recommendation, the USPSTF determined that in addition to subclinical thyroid dysfunction, screening could also identify undiagnosed overt thyroid disease;^2,3 therefore, the decision to screen should also consider the potential benefits and harms of identifying and treating undiagnosed overt disease. This update therefore differs from prior USPSTF reviews and the 2011 review in that it also addresses identification and treatment of undiagnosed overt thyroid disease.

Condition Definition

The thyroid gland is involved in metabolic homeostasis in adults through secretion of two hormones, thyroxine (T4) and triiodothyronine (T3), and is regulated by TSH, which is secreted by the anterior pituitary. Hypothyroidism is the undersecretion of T4 and T3, while hyperthyroidism is the oversecretion of these hormones. Current assays for TSH are extremely sensitive at detecting changes in thyroid homeostasis prior to changes in T4 and T3 levels. Subclinical thyroid dysfunction is defined as an elevated or low TSH test (normal reference range, 0.45 to 4.5 mIU/L) in the setting of normal thyroid hormone levels, and overt thyroid disease is defined by the presence of abnormal thyroid hormone (free T4, with or without T3) levels^8,19 (Table 1). Some recent data suggest that normal TSH ranges may require adjustment for age.^13,20-22

Table 1

Classification of Thyroid Dysfunction: Biochemical Definition.

Symptoms of overt hypothyroidism include fatigue, feeling cold, weight gain, hair loss, poor concentration, dry skin, and constipation (Table 2). Because a number of these symptoms are so common and nonspecific, they may be subtle and unrecognized. Myxedema coma is an uncommon but life-threatening complication of severe untreated or undertreated hypothyroidism, usually seen in the elderly. This condition may be precipitated by factors that impair respiration and is marked by hypothermia, hypoventilation, decreased level of consciousness, and sometimes seizures.¹⁹ Symptoms of overt hyperthyroidism include palpitations, heat intolerance and sweating, weight loss, hyperactivity, and fatigue. Thyroid storm is a potentially life-threatening condition that results from an acute illness superimposed on undiagnosed or undertreated hyperthyroidism. It is accompanied by fever, delirium, seizures, and coma.¹⁹

Table 2

Symptoms and Signs of Overt Thyroid Disease.

Prevalence and Burden of Disease

Subclinical thyroid dysfunction is more common than overt thyroid disease, and subclinical hypothyroidism is more common than subclinical hyperthyroidism. The National Health and Nutrition Examination Survey (NHANES) III found that 0.3 percent of the population had overt hypothyroidism, 4.6 percent had subclinical hypothyroidism, 0.5 percent had overt hyperthyroidism, and 0.7 percent had subclinical hyperthyroidism.²

Prevalence estimates of subclinical hypothyroidism vary based on population factors and according to differences in the defined upper normal limit for TSH. In general, prevalence increases with age, is higher in whites than blacks, and is higher in women than men.⁸ Cross-sectional studies have found that about 5 percent of women and 3 percent of men have subclinical hypothyroidism.¹ For women, estimates range from 1.2 percent in non-Hispanic black women to 5.8 percent in non-Hispanic white women in NHANES III (abnormal TSH defined as >4.5 mIU/L),² and from 4 percent in women ages 18 to 44 years to more than 17 percent in women older than age 75 years who participated in the Whickham Survey (abnormal TSH defined as >6.0 mIU/L).²³ For men, estimates from NHANES III range from 1.8 percent in non-Hispanic black men to 2.4 percent in Mexican American men; estimates from the Whickham Survey range from 1 percent in men ages 18 to 65 years to 6.2 percent in men age 65 years or older.¹

When defined as an undetectable TSH level in a person with a normal free T4 level, the prevalence of subclinical hyperthyroidism is about 1 percent (95% confidence interval [CI], 0.4 to 1.7) in men and 1.5 percent (95% CI, 0.8 to 2.5) in women older than age 60 years.¹ Subclinical hyperthyroidism has a higher prevalence in women, blacks, the elderly, and persons with low iodine intake.⁸

Untreated subclinical hypothyroidism may increase risk for developing coronary artery disease, lipid disorders, hypertension, obesity, and memory or cognitive disorders; subclinical hyperthyroidism may increase risk for atrial fibrillation, osteoporosis, and osteoporotic fractures.^8,24-28 Although untreated subclinical hypothyroidism does not appear to increase risk for all-cause mortality, evidence on the association between subclinical hyperthyroidism and all-cause mortality is more mixed.^28,29

Studies on subclinical thyroid dysfunction generally enrolled referral and other clinical populations rather than asymptomatic persons identified through screening. It is unclear whether otherwise healthy persons with subclinical thyroid disease identified by screening are at higher risk for adverse clinical outcomes compared with persons with normal thyroid function.¹

Etiology, Natural History, and Risk Factors

The most common cause of hypothyroidism in the United States is chronic autoimmune (Hashimoto's) thyroiditis. Other causes include previously treated thyroid dysfunction, poor adherence to or undertreatment with levothyroxine, external beam radiation in the head and neck area, and untreated adrenal insufficiency.^8,19 In addition to demographic risk factors described earlier, risk factors for hypothyroidism include type 1 diabetes mellitus, a family history of thyroid dysfunction, and Down syndrome.

Causes of hyperthyroidism include Graves' disease, autoimmune thyroiditis (“Hashitoxicosis”), functional thyroid nodules, and overtreatment with levothyroxine.¹⁹ In addition to demographic risk factors described earlier, risk factors include a personal or family history of hyperthyroidism and ingestion of iodine-containing drugs, such as amiodarone.

Some of the best data regarding the natural history of subclinical hypothyroidism comes from the Whickham Survey, which followed 2,779 British residents older than age 20 years. It found that, for a 50-year-old woman who had a serum TSH level of 6 mIU/L and positive antithyroid antibodies, the risk for developing overt hypothyroidism over 20 years was 57 percent; for a serum TSH level of 9 mIU/L, the risk was 71 percent. A 50-year-old woman who had a normal TSH level and negative antibody test had a risk of only 4 percent over 20 years.^1,23

Subclinical hyperthyroidism is less common than subclinical hypothyroidism and not as well studied. However, it has been estimated that 1 to 2 percent of persons with a TSH level less than 0.1 mIU/L develop overt hyperthyroidism each year, with a low likelihood of progression for TSH levels between 0.1 and 0.45 mIU/L.⁸

Rationale for Screening

The rationale for screening for subclinical thyroid dysfunction includes its relatively high prevalence and the potential for identification of affected persons to improve clinical outcomes through treatment prior to progression to overt disease or to mitigate adverse physiologic changes associated with adverse health outcomes later in life.¹⁰ Screening could also identify persons with unrecognized overt thyroid disease or those with mild symptoms who have not sought care for their symptoms. Other factors in favor of screening include the low cost, wide availability, and acceptability of the screening test (serum TSH), as well as low-cost and widely available treatment (levothyroxine).^10,30

Treatment

Overt hypothyroidism is treated with thyroid hormone replacement therapy, while subclinical hypothyroidism can be treated with thyroid hormone replacement therapy or a strategy of watchful waiting. Most experts recommend treating persons with a TSH level of greater than 10 mIU/L, while treatment is more controversial for those with a TSH level between 4.5 and 10 mIU/L.¹⁰ Replacement therapy is not thought to prevent progression to overt hypothyroidism, but may reduce risk for symptoms of overt disease in those who do progress.

Hyperthyroidism is treated with antithyroid medications, such as methimazole, or ablation therapy, such as radioactive iodine or surgery. Treatment is often recommended for persons who have an undetectable TSH level or a TSH level of less than 0.1 mIU/L, or who have Graves' disease or nodular thyroid disease, because of the risk for atrial fibrillation or bone loss, particularly in older adults. However, treatment is not recommended for subclinical hyperthyroidism due to thyroiditis, which typically resolves spontaneously.⁸ Routine treatment is also not recommended in patients with a TSH level between 0.1 and 0.45 mIU/L.⁸

Current Clinical Practice

Screening for both hypothyroidism and hyperthyroidism is accomplished through testing of serum TSH, with testing of serum free T4 (and in some cases T3) if the TSH level falls outside of the normal range.¹⁹ Additional testing is not routinely performed but may be done depending on the results of initial tests and the need to exclude other associated conditions. Common symptoms of mild thyroid dysfunction, such as mild fatigue or weight changes, are nonspecific, very common, do not predict the presence of thyroid dysfunction, and may be unrecognized or unreported. Therefore, screening may also result in identification of persons with overt thyroid disease who are not technically asymptomatic.

Recommendations of Other Groups

In 2002, a consensus panel sponsored by the American Association of Clinical Endocrinologists, the American Thyroid Association, and the Endocrine Society found insufficient evidence to support population-based screening for thyroid dysfunction, although it recommended aggressive case-finding in those considered to be high risk, including pregnant women and women older than age 60 years.³¹ Subsequently, a second panel appointed by the same three organizations reviewed the same evidence. While the panel acknowledged that the evidence does not support screening, it thought that a “lack of definitive evidence for a benefit does not equate to evidence for lack of benefit,” and issued a separate dissenting consensus statement that recommended routine screening for subclinical hypothyroidism and hyperthyroidism in adults.¹⁰ The American Thyroid Association also recommends screening in adults beginning at age 35 years and every 5 years thereafter.³⁰

A committee appointed by the Institute of Medicine in 2003 examined screening for hypothyroidism and hyperthyroidism in the Medicare population and concluded that “there is insufficient evidence to recommend periodic, routine screening for thyroid dysfunction among asymptomatic persons using serum TSH levels.”³² The American Academy of Family Physicians adopted the 2004 USPSTF recommendation.³³ The American College of Physicians does not have a current guideline (its 1998 guideline is inactive), but refers readers to the USPSTF recommendation.³⁴