NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Rugge B, Balshem H, Sehgal R, et al. Screening and Treatment of Subclinical Hypothyroidism or Hyperthyroidism [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2011 Oct. (Comparative Effectiveness Reviews, No. 24.)

Cover of Screening and Treatment of Subclinical Hypothyroidism or Hyperthyroidism

Screening and Treatment of Subclinical Hypothyroidism or Hyperthyroidism [Internet].

Show details



A mildly elevated thyroid stimulating hormone (TSH – also called thyrotropin) concentration is the most common thyroid function test abnormality encountered in everyday practice. Most patients who have a mildly elevated TSH have a normal free thyroxine (T4) level. The treatment of such patients is controversial, particularly when they have few or no symptoms and no other clinical evidence of thyroid disease. Less frequently, clinicians encounter patients who have a low or undetectable serum TSH and normal triiodothyronine (T3) and free T4 levels. The management of these patients is also unclear.

The main purpose of this review is to compare the effectiveness of different strategies for managing individuals who have mildly elevated serum TSH concentrations and those who have mildly diminished TSH concentrations with normal free T3 and/or free T4. We also address the evidence for whether the primary care clinician should screen for thyroid function in patients who have no specific indication for thyroid testing and who come to the clinician for other reasons. For the purposes of this review, we considered overt thyroid disease to be a well-defined clinical entity that has clear signs and symptoms, and thus, outside the scope of our review.

This topic was also the focus of a 2004 U.S. Preventive Services Task Force review.1

Description of Condition

Disorders of the thyroid gland are among the most common endocrine conditions that U.S. clinicians evaluate and treat. Hyperthyroidism or hypothyroidism affects about five percent of adults in the United States.2

The thyroid gland is involved in metabolic homeostasis in adults. It accomplishes this through secretion of two hormones, thyroxine (T4) and triiodothyronine (T3), and is regulated by thyroid stimulating hormone (TSH), which is secreted by the anterior pituitary. Hypothyroidism is the under-secretion of thyroid hormones, while hyperthyroidism is the over-secretion of these hormones.

Symptoms of overt hypothyroidism are subtle and nonspecific and may include fatigue, feeling cold, weight gain, hair loss, poor concentration, dry skin, and constipation (Table 1). If overt hypothyroidism is allowed to progress due to lack of treatment or under-treatment, then myxedema coma, a life-threatening condition, can occur. Myxedema coma is generally seen in the elderly and may be precipitated by factors that impair respiration; it is marked by hypothermia, hypoventilation, decreased level of consciousness, and sometimes seizures and death.3

Table 1. Symptoms and signs of overt thyroid dysfunction.

Table 1

Symptoms and signs of overt thyroid dysfunction.

Symptoms of overt hyperthyroidism may include palpitations, heat intolerance, and sweating, weight loss, hyperactivity, and fatigue. Thyroid storm is a life-threatening condition that results from an acute illness superimposed on undiagnosed or under-treated hyperthyroidism. It is accompanied by fever, delirium, seizures, and coma.3

Subclinical thyroid dysfunction includes subclinical hypo- and hyperthyroidism. Since the development of sensitive TSH assays, these conditions have been defined as follows (Table 2):

  • High or low serum thyroid stimulating hormone (TSH) levels
  • Normal free T4 and T3 levels
  • The absence of signs and symptoms of overt thyroid dysfunction.3

Table 2. Classification of thyroid dysfunction: Biochemical definition.

Table 2

Classification of thyroid dysfunction: Biochemical definition.

For the purposes of this report, the term ‘subclinical thyroid dysfunction’ is used to define the state of having an abnormal TSH in the context of normal free T4 and T3 levels. It includes those with ‘sub-clinical thyroid disease,’ i.e. those who have a high risk of disease progression or other adverse consequences, but also those whose prognosis is not well understood.

Patients with subclinical hypothyroidism are further categorized into those with mildly elevated TSH (4.5–10 mIU/L), and those with markedly increased serum TSH levels (>10 mIU/L).4Similarly, TSH levels below the lower reference limit can be classified as “low but detectable” (serum TSH 0.1-0.4mIU/L) and “clearly low serum TSH” (less than 0.1mIU/L).

Although widely used in the literature, these thresholds are arbitrary. For example, longitudinal studies have demonstrated that the risk of progression to overt hypothyroidism increases as the initial serum TSH level increases, but do not support the common notion of a threshold at either 4.5 mIU/L or at 10 mIU/L.1 Nevertheless, because many guidelines and studies use these thresholds, we use them in this report.

The biochemical definition has several limitations. First, the term “subclinical” usually implies that symptoms and signs are absent, whereas, in actual practice, the more common situation is that patients have nonspecific symptoms such as cold intolerance or feeling tired. As Cooper has noted, “Although subclinical hypothyroidism is the term most frequently used…, it is not necessarily apt, since on close questioning many patients disclose mild nonspecific symptoms. Mild hypothyroidism may be a more appropriate term for this very common syndrome, which is defined by an isolated elevated serum TSH level in the setting of normal serum thyroid hormone levels, in the presence or absence of symptoms.”5 Importantly, the presence of such symptoms does not mean they are related to the finding of an abnormal TSH test. In epidemiologic studies, individuals who had one or two nonspecific symptoms, such as cold intolerance or feeling “a little tired,” were no more likely to have subclinical thyroid dysfunction, or be at increased health risk, than were asymptomatic individuals in the general population.6-10

Second, the biochemical, TSH-based definition of subclinical thyroid dysfunction does not take into account clinical factors that affect the natural history. Subclinical hypothyroidism can be caused by recent hospitalization for severe illness; previously treated Graves disease or nodular thyroid disease; thyroiditis and recovery from thyroiditis; or untreated adrenal insufficiency. The lack of clinical context has caused confusion about the applicability of evidence from patients with overt thyroid disease to asymptomatic, otherwise healthy individuals who have an abnormal TSH level. For example, a frequently cited randomized trial of treatment for subclinical hypothyroidism recruited patients who had undergone treatment for Graves disease.11 Such patients predictably and rapidly progress to symptomatic hypothyroidism if not treated. While the study demonstrated quite convincingly that early treatment can prevent symptoms in patients who have undergone thyroid ablation, patients who have no history of thyroid disease and no clinical findings or symptoms attributable to the thyroid are unlikely to progress as rapidly, so treatment has much less effect on symptoms in the short term. The vast majority of patients identified by screening in the clinic or population-based settings are in the latter group.

Third, there is no consensus on what TSH level should be used as the cut off to diagnose subclinical hypothyroidism or hyperthyroidism. Differences among assays make it difficult to establish a universal upper limit.12 Most studies define an abnormal TSH test result as the upper and lower limits of the assay’s 95 percent reference range, approximately 0.1 to 4.5 mIU/L.

This method, although widely used in laboratory medicine, is not appropriate for identifying a threshold for diagnosing or treating subclinical hypothyroidism.13 As for other clinical measures, such as blood pressure, bone density, or serum lipid levels, the threshold should depend on the risk associated with a particular level as well as the balance of benefits and harms of treatment. Some have argued that the threshold for subclinical hypothyroidism should be raised above the upper limit of the reference range. The rationale is that otherwise healthy individuals who have a TSH in the range 4.5 mIU/L to 8 mIU/L or 9 mIU/L have not been shown to benefit from detection and treatment. Other experts argue that, because a TSH within the upper end of the usual reference range (2.5 to 4.5 mIU/L) confers some additional risk of progressing to overt hypothyroidism over time, the threshold for diagnosing subclinical hypothyroidism should be lowered to 2.5 mIU/L.14-16 Approximately 9.7 percent of the U.S. adult population, representing 20.6 million Americans, has a TSH in this range and would be identified as having subclinical hypothyroidism if this change were made.4

The appropriate level for decisionmaking can only be decided by better evidence about the adverse consequences in untreated individuals and the benefits and harms of treatment at different TSH thresholds.1,4,13,17 Because it depends on the risk of complications in a particular population, the appropriate TSH cutoff for defining subclinical hypothyroidism might vary with age, gender, and race.18

Fourth, although the definition of subclinical thyroid dysfunction requires a “normal” free T4 level, the definition is not necessarily applicable to the individual patient. Some experts argue that if a patient’s TSH level is mildly elevated, then even if their thyroid hormone levels are within the normal range, they “are not truly normal for that individual.”19 Put differently, while higher levels of thyroid-stimulating hormone increase thyroid hormone production, the additional production does not fully compensate for the underlying deficiency. According to this view, subclinical hypothyroidism represents “early thyroid failure,” that is, less than full compensation for the diminished function of the thyroid gland.

Prevalence and Course of Mild Thyroid Dysfunction

Using the upper limit of the reference range as a cut off, approximately 5 percent of women and 3 percent of men have subclinical thyroid dysfunction (i.e., TSH > 4 mIU/L).1 Approximately one in four of these individuals has a markedly elevated TSH concentration (>10 mIU/L). Such patients are likely to progress to overt hypothyroidism over 20 years.20

The other 75 percent of individuals with subclinical hypothyroidism have mildly elevated TSH levels (4 mIU/L < TSH ≤ 10 mIU/L). In this group, age, sex, geographic region, and the presence of thyroid auto-antibodies are strong predictors of the rate of progression to overt hypothyroidism. From one-third to two-thirds of these individuals have antithyroid antibodies. Depending on age, sex, and TSH level, 50 percent to 70 percent of these individuals will progress to overt disease over 20 years. In those who do not have antithyroid antibodies, the risk of progression is lower.20

In general, prevalence increases with age, is higher among whites compared with blacks, and higher in women compared with men.21 Estimates of the prevalence of subclinical hypothyroidism vary based on demographic factors and differences in the defined upper normal limit for TSH. A systematic review and meta-analysis of good-quality cross-sectional studies1 estimated that the prevalence in women ranged from 1.2 percent among non-Hispanic, African-American women to 5.8 percent in non-Hispanic, white women and from 4 percent in women age 18-44 to over 17 percent in women over 75 years. In the NHANES sample, estimates ranged from 1.8 percent among non-Hispanic, African-American men to 2.4 percent among Mexican-American men.22 In a population-based epidemiological study in Whickham, England, the prevalence ranged from 1 percent among men 18-65 to 6.2 percent among men 65 or older.23

Strategies for Detecting and Managing Subclinical Thyroid Dysfunction

Screening Strategies

Screening can be defined as “the application of a test to detect a potential disease or condition in a person who has no known signs or symptoms of that condition at the time the test is done.”24 In case-finding, testing for thyroid dysfunction is performed among patients who come to their clinicians for unrelated reasons. When the test is abnormal, the patient is called back for a detailed thyroid-directed history and confirmatory testing. Subclinical hypothyroidism is diagnosed if the TSH remains elevated and the free T4 remains normal for a period of 3-6 months. While hypothyroidism and hyperthyroidism are distinctly different disorders, with different symptoms and potential complications, screening for both subclinical hypo- and hyperthyroidism is accomplished through testing of serum TSH, with testing of serum free T4 if the TSH is high, and of T3 as well as free T4 if the TSH falls below the normal range.3

Management Strategies

Subclinical Hypothyroidism

For patients with subclinical hypothyroidism, the alternative management strategies are treatment with thyroid replacement hormone versus watchful waiting. A detailed strategy for routine treatment is described elsewhere.12 Treatment strategies vary, but all begin with repeat testing to confirm that the TSH is still elevated. If the TSH is >10 mIU/L on repeat testing, treatment with levothyroxine is initiated. If the TSH is mildly elevated (above the reference range but below 10 mIU/L), some experts recommend routine treatment. Others recommend measurement of serum thyroid peroxidase antibodies and treatment with levothyroxine if antibody levels are high. All guidelines recommend levothyroxine rather than triiodothyronine or both as the drug of choice. The preference for levothyroxine is based on the results of randomized trials of levothyroxine alone versus combination therapy for patients with overt hypothyroidism that show no clear benefit of combination therapy over levothyroxine alone.25

“Watchful waiting” encompasses two different strategies—“expectant management” or “active surveillance.” Expectant management means closely watching a patient’s condition and treating if symptoms develop or laboratory results change. Active surveillance means repeating thyroid function tests and taking a thyroid-directed history at a particular interval; then, if symptoms or laboratory results change, beginning treatment. The appropriate time-interval for retesting individuals with subclinical thyroid dysfunction is unknown.

A recent British guideline26 offered the following typical strategy for active surveillance:

  • If screening is performed, and a high serum TSH concentration and normal free T4 is found, repeat measurement 3-6 months later after excluding nonthyroidal illness and drug interference
  • If the TSH is mildly elevated (above the reference range but below 10 mIU/L), obtain serum thyroid peroxidase antibodies
  • If antibody levels are high, repeat measurement of TSH annually. If they are low, repeat measurement of TSH every 3 years. Initiate treatment if the TSH level is greater than 10 mIU/L or the patient develops clinical findings of hypothyroidism.

Practice styles vary widely. A well-conducted chart review study of 500 patients with subclinical hypothyroidism seen at the Mayo Clinic in 1995-1996 found that clinicians treated 38.7 percent of patients who had a TSH between 5.1 and 10.0 mIU/L.27 Unfortunately, more recent data and primary care-based practice patterns are not available.

Subclinical Hyperthyroidism

For subclinical hyperthyroidism, some experts 21,28 recommend repeating thyroid function tests after 3 months and, if the TSH remains suppressed, obtaining ultrasonography and a 24-hour Radioactive Iodine Uptake (RAIU) thyroid scan. These guidelines recommend antithyroid treatment if the patient has a persistent TSH level less than 0.1 mIU/L and is found to have Graves or nodular thyroid disease. Treatments include medications, such as propylthiouracil, or ablation with radioactive iodine or surgery. The guideline recommended against routine treatment in those whose TSH was between 0.1 and 0.45 mIU/L.

Guidelines on Early Detection and Treatment

A review of the history of guidelines for subclinical thyroid dysfunction provides insight into the reason for practice variation (see also Appendix E). In 1990 and again in 1998, the American College of Physicians found “it is reasonable to screen women older than 50 years of age for unsuspected but symptomatic thyroid disease.” The guideline specified that the goal of routine testing was to find overt, but overlooked, thyroid dysfunction, not subclinical hypothyroidism. Because the clinical significance of mildly elevated TSH test results was uncertain, the guideline recommended obtaining a free T4 test only when the TSH level was undetectable or 10 mIU/L or more. The ACP guideline panel used a systematic review of the literature to arrive at its recommendations.10These guidelines expired in 2003.

In 1999, the American Association of Clinical Endocrinologists recommended screening asymptomatic women over the age of 60.29In 2000, the American Thyroid Association recommended screening all patients over 35 years of age every 5 years (more frequently if the patient is at increased risk).30 These organizations used a consensus process to develop guidelines and did not use systematic reviews in arriving at their recommendations.

In 2003, the Institute of Medicine (IOM) published a volume entitled “Medicare Coverage of Routine Screening for Thyroid Disease,” which examined the issue of screening for thyroid dysfunction 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.”31 In 2004, the USPSTF determined that the evidence was poor that treatment of screen-detected adults, in either the general population or in specific high-risk groups, improves clinically important outcomes and concluded that there was insufficient evidence to recommend for or against screening for thyroid disease in adults. These groups used essentially identical systematic reviews to arrive at their recommendations.1,31 The USPSTF issued an “I” recommendation, indicating that the evidence was insufficient to recommend for or against routine screening for thyroid disease in adults.32The conclusions about the evidence were:

  • There is fair evidence that the thyroid stimulating hormone (TSH) test can detect subclinical thyroid dysfunction in people without symptoms of thyroid dysfunction, but poor evidence that treatment improves clinically important outcomes in adults with screen-detected thyroid disease
  • Although the yield of screening is greater in certain high-risk groups (e.g., postpartum women, people with Down’s syndrome, and the elderly), there is 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 good 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 asymptomatic adults for thyroid disease could not be determined.

In 2004, a panel sponsored by the American Association of Clinical Endocrinologists, the American Thyroid Association, and the Endocrine Society evaluated data regarding the management of subclinical thyroid dysfunction.21 Unlike the older AACE and ATA guidelines, the panel used a systematic review of the evidence to arrive at its recommendations.33 The panel adopted the strength of evidence rating of the USPSTF in its assessment. The panel found insufficient evidence to support population-based screening and recommended against population-based screening for thyroid disease, though it did advocate aggressive case-finding in those considered high-risk, including pregnant women and women older than 60. They also recommended against routine treatment of patients with subclinical hypothyroidism with serum TSH levels of 4.5 - 10.0 mIU/L. Specifically, the panel found insufficient evidence to support routine treatment of individuals who have a mildly elevated TSH (Table 3). The findings about the evidence regarding the complications of subclinical hypothyroidism and the effects of treatment agreed with those of the IOM and USPSTF.

Table 3. Evidence for the association of subclinical hypothyroidism (TSH 4.5-10 mIU/L) and adverse health outcomes and quality of evidence for risks and benefits of treatment: Findings of the American Association of Clinical Endocrinologists, the American Thyroid Association, and the Endocrine Society Panel.

Table 3

Evidence for the association of subclinical hypothyroidism (TSH 4.5-10 mIU/L) and adverse health outcomes and quality of evidence for risks and benefits of treatment: Findings of the American Association of Clinical Endocrinologists, the American Thyroid (more...)

In 2005, the three professional societies that sponsored the evidence-based panel issued a consensus statement rejecting its recommendations.34 While acknowledging that the independent review panel found insufficient evidence to recommend treatment of patients with subclinical hypothyroidism in the range of 4.5-10.0 mIU/L, the counter argued that “lack of definitive evidence for a benefit does not equate to evidence for lack of benefit” and recommended that most patients with TSH levels between 4.5 and 10.0 mIU/L should be considered for treatment. Their rationale for recommending screening despite gaps in the evidence about treatment is discussed in detail in the next section of this report (“Rationale for Screening and Treatment”).

In 2006, three British professional associations (Association for Clinical Biochemistry, the British Thyroid Association, and the British Thyroid Foundation) published guidelines for using thyroid function tests.26 The guideline panel found that “screening for thyroid dysfunction in a healthy adult population is not warranted.” It recommended against routine treatment of patients who have a TSH concentration above the reference range but below 10 mIU/L but said that “a therapeutic trial of thyroxine [should be considered] on an individual patient basis.” The British panel noted that that there was growing evidence against the use of thyroxine in elderly patients, in whom an elevated TSH concentration “may reflect an adaptive mechanism to prevent excessive catabolism.” Like the ACP and AACE guidelines, the British panel recommended aggressive case-finding in women with non-specific symptoms. The panel used systematic reviews in their process for developing guidelines.35

In 2007, The American College of Obstetricians and Gynecologists (ACOG) recommended against routine screening in pregnant women. It stated that there is no evidence that identifying and treating pregnant women with subclinical hypothyroidism improved either maternal or infant outcomes.36

Rationale for Screening and Treatment

Subclinical Hypothyroidism

Although there is wide agreement that the long-term benefits of early treatment of subclinical thyroid dysfunction have not been proven, there is disagreement about what to do until better evidence is available. This disagreement reflects differing views of the clinical relevance of research about the complications of subclinical thyroid dysfunction.

Proponents argue that thyroid dysfunction is common and associated with significant morbidity. Additionally, a serum TSH test is relatively inexpensive, accurate, readily available, and generally a very acceptable test for patients to undergo.30,34Symptoms of overt thyroid dysfunction also can be vague and at times difficult to diagnose, and therefore, thyroid screening may allow the diagnosis of overt disease earlier in the clinical course, thus reducing morbidity.30,34,37 For subclinical hypothyroidism, treatment with levothyroxine is noninvasive and inexpensive. Finally, proponents argue that the potential harms of screening are small in relation to the potential benefits: “Because the potential harm of early detection and treatment appear to be so minor and preventable, it seems prudent to err on the side of early detection and treatment until there is sufficient data to address these issues definitively.”34

Since 2005, the most important, widely cited argument for early treatment of thyroid dysfunction is the association of untreated subclinical hypothyroidism with other risk factors for heart disease and with incident coronary disease or heart failure later in life. Subclinical hypothyroidism may be a risk factor for atherosclerosis and myocardial infarction, but epidemiologic studies of this question have had mixed results.38-42 Four recent meta-analyses have evaluated the association of subclinical thyroid disorder on cardiac mortality or all-cause mortality.

One meta-analysis included six cohort studies and found that those with subclinical hypothyroidism had a relative risk (RR) of 1.53 (1.31-1.79) of having coronary artery disease (CAD) at baseline.43 Subclinical hypothyroidism was associated with a RR of 1.188 (1.024-1.379) of developing CAD in followup (included studies followup ranged from 4 to 20 years), but was not found to be associated with all-cause mortality.43 The second review (of 14 cohort and two case-control studies) found no association with subclinical hyperthyroidism and circulatory or all-cause mortality.44 With regard to subclinical hypothyroidism, the association with circulatory mortality was not statistically significant, but a hazard ratio of 1.25 (1.3-1.53) was found for all-cause mortality.44 The third review, which included 10 population-based prospective studies and two studies of convenience samples of cardiac patients or patients with a history of stroke or hip replacement, found no overall statistically significant association between either subclinical hypothyroidism or subclinical hyperthyroidism and total mortality or coronary morbidity or mortality.45 A statistically significant, but small, increased risk for heart disease was found for individuals younger than 65.45 The last meta-analysis included 15 studies (six population-based cross-sectional studies and nine longitudinal cohort studies) with 2,531 subclinical hypothyroid participants and 26,491 euthyroid individuals.46 Five studies provided longitudinal data on the risk of coronary events; in these, overall there was no difference in incident ischemic heart disease (IHD) in subclinical hypothyroid participants and euthyroid participants (OR 1.27 [0.95–1.69]). In a subgroup analysis, in studies of subjects younger than 65 years, the odds ratio was 1.68 (1.27–2.23), but for older subjects, there was no relationship between subclinical hypothyroidism and incident IHD (OR 1.02 [0.85–1.22]).46

The epidemiological studies included in these reviews had serious limitations. Many included individuals who had known thyroid disease, ischemic heart disease, or TSH levels within the reference range. Many included subjects who underwent treatment with levothyroxine during the followup period. In general, the studies did not adequately control for potential confounders, such as lipid levels and blood pressure.

The most recent research addresses some of the limitations of previous studies.40,47One of these, from the Cardiovascular Health Study, found no relationship between an elevated TSH level and cardiovascular outcomes among 3, 233 individuals aged 65 years or older enrolled in 1989-1990 and followed an average of 12.5 years.40 The other was an reanalysis of the Wickham Study, a survey of 2,779 adults sampled from a mixed urban and rural area of England conducted between July 1972 and June 1974. The goal of the study was to assess the prevalence of thyroid disease in a cross-section of the community. Personal and family history of thyroid disease and thyroid-related symptoms was collected along with history of diabetes and cardiovascular-related diseases, and fasting blood samples included assessment of levels of TSH, T4, T3, and cholesterol.23 Unlike the earlier report from the Whickham study,48 the new analysis included only subjects who had mildly elevated TSH levels, excluded subjects who had known thyroid disease or ischemic heart disease, stratified subjects according to whether they were treated with thyroxine during the followup period, and adjusted for several cardiovascular risk factors as well as socioeconomic status. After adjustment for baseline age, gender, social class, body weight, history of cerebrovascular disease, diabetes mellitus, smoking, systolic and diastolic blood pressure, serum cholesterol levels, and levothyroxine use during followup as covariates, subclinical hypothyroidism was associated with a higher risk of coronary events over 20 years (hazard ratio 1.76 [1.15–2.71]) When levothyroxine use during followup was not controlled for, the relationship was weaker (hazard ratio 1.53 [0.97–2.45]). Among the 91 subjects who had subclinical hypothyroidism, two of 20 who were treated with levothyroxine during the course of the study died, versus 22 of 71 who did not receive levothyroxine. Despite the low number of events, the authors reported a hazard ratio of 0.20 (0.05–0.89) for all-cause mortality after adjustment for age, gender, and total serum cholesterol.

While it is stronger than previous evidence, this analysis also had weaknesses. First, the description of the methods for ascertaining cardiac events is not clear. Specifically, it is not clear what events were considered in addition to documented myocardial infarction or death from coronary disease. In the context of risk factor epidemiology, for a broadly defined composite endpoint, the hazard ratios are low, making it more likely that confounding or methodological factors account for the observed differences. Second, the study did not control for the use of lipid-lowering medications during the followup period. It is possible that the better outcomes of subjects treated with levothyroxine could be due to other interventions introduced by their clinicians. Finally, the study was conducted between 1972 and 1974, before the era of statins and other aggressive cardiac risk management techniques. It is possible that the association between subclinical hypothyroidism and subsequent cardiovascular outcomes would become negligible in the context of current cardiac disease management.

The 2004 USPSTF review by Helfand found that evidence regarding mildly elevated TSH levels and hyperlipidemia and atherosclerosis was inconsistent.1 In the recent reanalysis of the Whickham study data, subjects with subclinical hypothyroidism had higher baseline systolic blood pressure (146.9 ±26.4 mm Hg vs.139.5± 24.7 mm Hg) and total cholesterol levels (6.2 ±1.3 mmol/L [239.8 ± 50.3 mg/dL]) vs. 5.9 ±1.2 [228.2 ± 46.4 mg/dL]) and were slightly less likely to smoke than euthyroid subjects (see Appendix A for lipid conversion factors). After adjustment for age, gender, weight, smoking, and relevant medications, however, systolic blood pressure was associated with subclinical hypothyroidism, but total cholesterol was not. In other cross-sectional studies, subclinical hypothyroidism was weakly associated with total and LDL cholesterol, blood pressure, abnormalities of cardiac function, and subcutaneous fat.49,50

Recently, two prospective studies have demonstrated a correlation between subclinical hypothyroidism and congestive heart failure (CHF). In the Health, Aging, and Body Composition Study cohort, 338 of 2,730 individuals, ages 70 to 79 were found to have subclinical hypothyroidism.41 After a 4 year followup, a total of 178 individuals had a CHF event.41 While those with a TSH of 4.5 to 6.9 did not have a statistically significant higher rate of CHF events compared to those without subclinical hypothyroidism, those with a TSH of 7 to 9.9 had a hazard ratio of 2.58 (95% CI, 1.19-5.60) and those with a TSH >10 had a hazard ratio of 3.26 (95% CI, 1.37-7.77).41 Additionally, of the 2,555 without a history of CHF, those with a TSH of > 7 had a hazard ratio of 2.33 (95% CI, 1.1.0-4.96) of developing a incident CHF event.41 In the Cardiovascular Health Study cohort, 474 of 3044 subjects older than 65 years who had no history of CHF had subclinical hypothyroidism, 46 of whom had a TSH of > 10.51 After a median of 12 years of followup, those with a TSH >10 had a hazard ratio of 1.88 (95% CI,1.05-3.34) of developing a CHF event, while those with a TSH of 4.5 to 9.9 were not more likely to have a CHF event.51

These two relatively large prospective cohort studies do indicate that an isolated TSH might be a risk factor for the development of CHF, particularly for individuals with a TSH of > 10. This association deserves further study. However, it is still unknown if thyroid replacement therapy would modify this potential risk.

Subclinical Hyperthyroidism

Cross-sectional studies have shown untreated subclinical hyperthyroidism to be associated with tachycardia, increased left ventricular mass leading to diastolic dysfunction, atrial arrhythmias, and a decline in bone mass density increasing the risk of fractures.1,21 Only atrial fibrillation and disease progression have been associated with subclinical hyperthyroidism in longitudinal studies.1 In a recent longitudinal study of 102 women who had a TSH between 0.1 to 0.4 mIU/L, three progressed to overt hyperthyroidism and 24 reverted to a normal TSH.52

In summary, progression to overt disease is the best established complication of subclinical thyroid dysfunction. Epidemiologic data suggest that subclinical hypothyroidism is associated with cardiovascular disease in subjects younger than 65 years, but the magnitude of risk is low. Evidence about the relationship of a mildly elevated TSH to symptoms and to other cardiac risk factors, including the metabolic syndrome, is weak. Subclinical hyperthyroidism is less common and less studied, and thus, its natural history and effects are less clear.

Scope of Review

The main question addressed in this review was whether individuals who have mildly abnormal TSH values will either benefit from or be harmed by screening and potential subsequent treatment. We also addressed the question of whether the primary care clinician should screen for thyroid function in patients seen in general medical practice who have no specific indication for thyroid testing and who come to the clinician for other reasons. For the purposes of this review, we considered overt thyroid disease to be a well-defined clinical entity that has clear signs and symptoms, and thus, outside the scope of our review.

In order for a condition to be a good candidate for screening in the general population, several conditions need to be met. First, the condition needs to be relatively prevalent, having a significant impact on the health of the population or an easily identified special population. Second, there needs to be a test that is readily available to the general population that is of reasonable cost and accuracy and is acceptable to individuals to undergo. Finally, there needs to be an intervention that is of reasonable cost and tolerability that when administered in a timely fashion will alter the disease state to prevent morbidity and/or morality.

The Helfand (2004) review established that subclinical thyroid disease is quite prevalent; may be responsible for morbidity; and that the serum TSH test is a readily available, reliable, and acceptable test to detect the condition with a sensitivity above 98 percent and specificity greater than 92 percent.1 However, in 2004, it remained unclear whether, if detected, treating patients with subclinical thyroid disease would reduce morbidity.

As evidence of prevalence, test yield, and test performance have already been adequately established,1 this current review focuses on whether new evidence demonstrates that treatment improves clinically important outcomes in adults with screen-detected thyroid disease.

Key Questions and Analytic Framework

We reviewed published studies to answer the following key questions:

Key Question 1

Does screening for subclinical thyroid dysfunction reduce morbidity or mortality?

Key Question 2

What are the harms of screening? Specifically, how frequently and how severely do patients screened for subclinical thyroid dysfunction experience adverse psychological impacts or other harms of work-up from screening?

Key Question 3

Does treatment of patients with subclinical hypothyroidism or subclinical hyperthyroidism detected by screening affect outcomes?

Key Question 4

What are the harms of treatment of subclinical hypothyroidism and subclinical hyperthyroidism? Specifically, what are the consequences of over-treatment, including effects on bone mineral density and incidence of atrial fibrillation, and how frequently do they occur?

We developed the final analytic framework shown in Figure 1 to guide the literature review. The analytic framework shows the populations, classification, intermediate outcomes, and health outcomes we examined in the review. We defined the target population as community-living, non-pregnant adults, without a history of thyroid disease or symptoms of overt hypothyroidism or hyperthyroidism, who might be representative of those seen in primary care settings.

Figure 1 is an analytic framework that shows the populations, classification, intermediate outcomes, and final outcomes examined in the review. The target populations are asymptomatic nonpregnant adults without known thyroid disease and adults with mild nonspecific symptoms (fatigue, cold intolerance, etc.).  Arrow 1 represents studies directly comparing effects of screening vs. not screening on important health outcomes (Key Question 1). Arrow 2, corresponding to Key Question 2, represents evidence about potential harms of screening. Arrow 3 represents studies comparing different strategies (treatment vs. monitoring) for managing individuals who are found to have borderline high or low TSH levels (Key Question 3). Arrow 4 represents studies of harms of treatment. Outcomes of interest for subclinical hypothyroidism were cardiovascular morbidity and mortality, measures of well-being, weight change, and progression to overt disease. Outcomes of interest for subclinical hyperthyroidism were fracture, atrial fibrillation, measures of well-being, weight change, and progression to overt disease. Harms of screening are  psychological impact, cost, and harm of workup. Harms of treatment are number needed to treat, cost, and overtreatment resulting in atrial fibrillation and bone density problems.

Figure 1

Analytic framework. Note: A. fibrillation=atrial fibrillation; CAD=coronary artery disease; CHF=congestive heart failure; DEXA=dual-energy x-ray absorptiometry; NNS=number needed to screen; NNT=number needed to treat; S=subclinical; TSH=thyroid stimulating (more...)

In the framework, Arrow 1 represents studies directly comparing the effects of screening versus not screening on important health outcomes (Key Question 1). Arrow 2, corresponding to Key Question 2, represents evidence about the potential harms of screening. Arrow 3 represents studies comparing different strategies (treatment vs. monitoring) for managing individuals who are found to have borderline high or low TSH levels (Key Question 3). Arrow 4 represents studies of the harms of treatment. Outcomes of interest for subclinical hypothyroidism were cardiovascular morbidity and mortality, measures of well-being including but not limited to cognition and memory, weight change, blood pressure changes, and changes in lipid levels. For subclinical hyperthyroidism, outcomes of interest were cardiovascular morbidity and mortality, osteoporotic fractures, measures of well-being including but not limited to cognition and memory, weight changes, blood pressure changes, changes in bone mineral density, and changes in lipid levels.


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.4M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...