How Selective Are the New Guidelines for Treatment of Subclinical Hypothyroidism for Patients with Thyrotropin Levels At or Below 10 mIU/L?

Introduction

S ubclinical hypothyroidism (SCH) is highly prevalent among adults (∼5–10%) (1 –5). This fact requires careful treatment recommendations, since a small reduction in the thyrotropin (TSH) cut-off or inclusion of a certain condition as an indication for treatment means offering levothyroxine therapy (L-T4) to millions more people. Unfortunately, the indications for hormone replacement therapy in SCH are rarely based on the demonstration of the benefit of this approach in large randomized studies (2 –4).

By consensus, TSH persistently >10 mIU/L is an indication for the treatment of SCH (2 –4). Controversy exists regarding the recommendation of L-T4 replacement therapy for subjects whose TSH is elevated but <10 mIU/L (approximately 85% of patients with SCH) (5,6). Recently, the American Thyroid Association (ATA) and the American Association of Clinical Endocrinologists (AACE) stated that “treatment based on individual factors for patients with TSH levels between the upper limit of a given laboratory's reference range and 10 mIU/L should be considered particularly if they have symptoms suggestive of hypothyroidism, positive [anti]thyroperoxidase antibodies (TPOAbs), or evidence of atherosclerotic cardiovascular disease [or] heart failure, or have associated risk factors for these diseases” (4).

The objective of the present study was to evaluate the frequency of these conditions in adult (non-pregnant) women with SCH and TSH ≤10 mIU/L in order to define the impact of the new guidelines proposed by ATA and AACE (4).

Patients and Methods

Among women sequentially referred to the lead author (P.W.R.) by primary care physicians, cardiologists, and gynecologists between April 2005 and April 2010 because of elevated serum TSH concentrations, 252 women who had a confirmed diagnosis of SCH and TSH ≤10 mIU/L were studied. SCH was defined based on two TSH measurements between 4.5 and 10 mIU/L obtained at an interval of approximately 12 weeks, in the presence of normal free T4 and in the absence of any other cause of hyperthyrotropinemia (hospitalization or major surgery up to two months prior to evaluation, antidopaminergic drugs, untreated adrenal insufficiency, and chronic renal failure) (7). Women who were pregnant were excluded. Women treated with L-T4, antithyroid drugs, amiodarone, lithium, or interferon were also excluded.

The following data were available for all women: (i) previous diagnoses (atherosclerotic disease, heart failure); lifestyle habits (smoking); (ii) use of medications; (iii) family history (coronary disease); (iv) symptoms or signs of hypothyroidism (physical tiredness, mental lethargy, slow cerebration, reduced sweating, dry skin, cold intolerance, dry hair, weight gain, constipation, hoarseness, paresthesia, muscle pain, hearing impairment, slow movements, coarse skin, cold skin, periorbital puffiness, pulse rate, and slow relaxation phase of the ankle jerk) (8); symptoms suggestive of atherosclerotic disease (intermittent claudication, dyspnea, or chest pain during exercise); (v) general physical examination (body–mass index, blood pressure, pulse palpation, auscultation of the carotid and abdominal arteries); (vi) basic laboratory tests (TPOAbs, oral glucose tolerance test [in non-diabetic subjects], glycated hemoglobin [HbA1c], total cholesterol and fractions). These data were obtained by physicians as part of a larger prospective study started in 2005 and involving patients with SCH.

A high cardiovascular risk was defined in women with known atherosclerotic disease (cerebrovascular accident, acute myocardial infarction, stable or unstable angina, ischemic cardiomyopathy, aortic aneurysm, renovascular disease); women with a suspicion upon anamnesis or physical examination of atherosclerotic disease and subsequently confirmed by adequate imaging methods; diabetic women; and women with an estimated 10-year risk of coronary events >20%, calculated using the Framingham risk score (9). The following cardiovascular risk factors were considered (except for diabetes and age): current smoking; blood pressure ≥140/90 mmHg (on two occasions) or use of antihypertensive medication; high-density lipoprotein–cholesterol <40 mg/dL or low-density lipoprotein–cholesterol >160 mg/dL (in two measurements); family history (first-degree relative) of premature coronary artery disease (men <55 years and women <65 years) (9). Diabetes mellitus was diagnosed if fasting blood glucose was ≥126 mg/dL on two occasions, or ≥200 mg/dL after an oral glucose load, or if HbA1c was >6.5%.

The study was approved by the Research Ethics Committee of our institution.

The presence of the conditions that should be considered for the treatment of SCH according to the ATA and AACE (4) were evaluated in this study. Symptoms were only considered when they could not be explained by another associated condition, when they were persistent, and when they bothered the patient.

Results

The characteristics of the patients and frequency of the parameters studied are shown in Table 1. TPOAbs were detected in 137 (54.3%) women. A high cardiovascular risk was observed in 43 (17%) women. Eighty (31.7%) women who were not at high cardiovascular risk presented at least one classical risk factor. At least one symptom or sign of hypothyroidism that could not be explained by another condition was observed in 180 (71.4%) women. According to the new guidelines of the ATA and AACE (4), L-T4 therapy would be considered for 92% of women with SCH and TSH ≤10 mIU/L, since they had positive TPOAb or at least one classical cardiovascular risk factor or at least one symptom or sign of hypothyroidism.

Discussion

The data show that, according to the new guidelines of the ATA and AACE (4), L-T4 therapy would be considered for more than 90% of female patients with SCH and TSH ≤10 mIU/L.

First, the existence of underlying (autoimmune) thyroid disease is very likely in an adult with a persistently elevated TSH >4.5 mIU/L (although still ≤10 mIU/L) in the absence of any apparent nonthyroid-related cause. Therefore, a high frequency of TPOAbs is not surprising. In fact, previous studies have shown the presence of TPOAbs in 55–68% of subjects with elevated TSH, even those with ≤10 mIU/L (5,7,10 –13). Using only this criterion (positive TPOAbs), which is included in the new guidelines of the ATA and AACE (4), L-T4 therapy would be considered in >50% of cases of SCH and TSH ≤10mIU/L. The presence of TPOAbs does not change the diagnosis of SCH or the expected efficacy of treatment; it only identifies an autoimmune etiology and predicts a higher risk of developing overt hypothyroidism, with that rate still being low (4.3% per year) (10). In contrast to the new guidelines of the ATA and AACE (4), the importance of TPOAb measurement in patients with SCH has been questioned by other authors (2).

Second, symptoms of hypothyroidism are nonspecific and highly prevalent in the general population. In the Colorado Thyroid Disease Prevalence Study (1), at least one symptom was reported by 60% of euthyroid subjects. A similar rate was found in the present study, in which 70% of women with SCH had at least one symptom that could not be attributed to another condition. Previous studies reported symptom(s) in ∼40% (12), ∼65% (1), and ∼85% (14) of the patients evaluated (with SCH and TSH ≤10 mIU/L). Thus, if any of the classical symptoms of hypothyroidism is sufficient to indicate hormone replacement therapy, it will be considered in most patients with SCH. The adoption of more restricted criteria, such as the presence of between three and five symptoms (15) or a high score (8), may be a better approach. It should be noted that many studies did not find a higher prevalence of symptoms in patients with SCH, particularly those with TSH ≤10 mIU/L, when compared to euthyroid subjects (1 –3,16 –18). The validity of the use of this parameter is therefore questionable, especially for the elderly population (3,16,17) in which SCH is more frequent (1 –4). Moreover, there is no evidence that L-T4 therapy significantly improves these symptoms (19).

Finally, according to the ATA and AACE, another factor that should be considered for the treatment of SCH with TSH ≤10 mIU/L is the presence of cardiovascular disease or cardiovascular risk factors (4). Conditions such as diabetes mellitus, arterial hypertension, smoking, and dyslipidemia are common in the general adult population. In fact, the criterion of the presence of at least one of these risk factors was met by about 45% of the patients with SCH studied here, and this frequency does not seem to be overestimated (12,13,20 –23). These data only refer to classical risk factors, and this frequency would be even higher if the clinician considered nonclassical risk factors. Therefore, when including cardiovascular risk as a parameter in the treatment decision of SCH with TSH levels ≤10 mIU/L, the restriction to high-risk patients instead of simply “with risk factors” may be a better approach. This restriction is not exaggerated, since there are no randomized (treated vs. untreated) studies demonstrating a benefit of hormone replacement therapy in reducing cardiovascular events and death (2 –4,24). Furthermore, the association of SCH with cardiovascular disease is poorly consistent with TSH levels <10 mIU/L (24).

In conclusion, according to new guidelines for the treatment of SCH in the presence of TSH levels ≤10 mIU/L (4), follow-up without L-T4 therapy represents an exception (<10% of cases), despite the lack of demonstration of its benefits in large randomized studies and of its long-term safety, since periods of subclinical thyrotoxicosis are not uncommon during L-T4 therapy (1). Moreover, if the clinician is flexible when interpreting the criteria, which are not reported in detail (4), including nonclassical risk factors for cardiovascular disease, less rigor in the attribution of symptoms to hypothyroidism, or even adopting cut-off values <10 mIU/L, it is likely that none of the patients with SCH escapes treatment.