In 1998, the American College of Physicians¹ recommended that a routine thyrotropin-stimulating hormone (TSH) measurement be obtained to screen women older than 50 years for thyroid disease. According to this guideline, screening could benefit patients by detecting symptomatic, overt hypothyroidism and hyperthyroidism that had otherwise escaped attention. A review of the guideline, published in Journal of Family Practice,² endorsed the rationale for the new guideline but recommended that screening should begin at age 60.

In the course of screening to detect thyroid dysfunction, physicians will also find laboratory abnormalities in otherwise healthy, asymptomatic persons. In such cases, when the TSH level is elevated (subclinical hypothyroidism), physicians must decide whether treatment with levothyroxine (Synthroid) is indicated. When the TSH level is low (subclinical hyperthyroidism), decisions are more difficult because treatments for hyperthyroidism have potentially serious complications, and minimal evidence about the benefits and risks of early treatment exists.

In this issue of American Family Physician, Shrier and Burman³ provide a balanced review of existing evidence and suggest a plan for the diagnostic evaluation and treatment of asymptomatic persons with a low TSH level. Their treatment recommendations are sensible. They recommend monitoring patients who have a low but detectable TSH level (more than 0.01 μU per mL [0.01 mU per L] but below the lower limit of the normal range). For patients with an undetectable TSH level (less than 0.01 μU per mL for most assays), they favor treatment with low-dose methimazole (Tapazole) or propylthiouracil, after taking into account the patient's general health, preferences, and ability to pay for medication.

Most published studies of subclinical hyperthyroidism lack controls and use small, convenience samples of patients who were referred to an endocrinology clinic at a tertiary medical center. Many of the patients in these studies had clinical signs of thyroid disease (e.g., a visible goiter, ophthalmic findings) that prompted testing. Such patients probably have a high rate of progression to overt hyperthyroidism and nonspecific symptoms (e.g., nervousness) that may improve after treatment.

Unfortunately, this literature has little relevance to asymptomatic patients who are found to have a low TSH level. When a low TSH level is found in asymptomatic patients who are screened in the primary care setting, it returns to the normal range in about 40 percent of cases.^4–6 In these patients, Shrier and Burman³ recommend repeating the TSH measurement and pursuing additional tests only if the TSH level is low in four consecutive measurements. They recommend a 24-hour radioactive iodine uptake and thyroid scan if the TSH level is persistently low.

Shrier and Burman's recommendations³ are practical because they offer guidance for physicians and patients who would be reluctant to simply ignore TSH values that are borderline low. No studies to date have examined the benefits of this approach. Evaluating all patients who have low TSH levels adds cost and complexity to a primary care screening program. In asymptomatic patients who have detectable, borderline-low TSH values (more than 0.01 μU per mL but below the lower limit of the normal range), the chance of finding underlying thyroid disease is negligible. Even if a multinodular goiter or diffuse increased uptake is found, monitoring rather than treatment still would be recommended.

Ironically, the complications of subclinical hyperthyroidism—osteoporosis,⁷ atrial fibrillation⁶ and, possibly, cardiac hypertrophy^8,9— are best documented in patients who take high doses of levothyroxine and are least well documented in asymptomatic persons identified through screening. Wider application of screening will increase the number of patients who are prescribed levothyroxine, either for endogenous hypothyroidism or for hypothyroidism caused by treatment for hyperthyroidism. From a societal perspective, the greatest potential for reducing the burden of disease comes not from screening to identify people with subclinical hyperthyroidism, but from prudent use and monitoring of levothyroxine therapy.