The Treatment of Subclinical Hypothyroidism Is Seldom Necessary

Abstract

Subclinical hypothyroidism (SH) is a common disorder with a prevalence ranging from 1–10% of the adult popula-tion in most community studies (1–14). The risk of develop-ing SH increases with female gender, advanced age, and greater dietary iodine intake. The view that most subjects with SH should be treated with l-thyroxine has gained in-creasing credence (15–17). A limited number of placebo-controlled randomized trials involving a small number of patients with SH have been performed (18 –22), and several of these studies show that l-thyroxine therapy may reduce symptoms of hypothyroidism (18, 22). Other studies in sub-jects with SH have shown that l-thyroxine lowers low-density lipoprotein (LDL) cholesterol, improves cardiac function, and diminishes neuropsychiatric symptoms (15– 17). A recent cross-sectional observational study noted an association between SH and atherosclerotic disease (13). On initial assessment then, it appears that the position advocat-ing l-thyroxine therapy for most subjects with SH enjoys strong experimental support. However, careful scrutiny of the entire spectrum of pri-mary data bearing on the question of whether or not SH should be treated with l-thyroxine yields a legitimate con-trary view. There are as many placebo-controlled random-ized trials that observed no reduction in symptoms of SH (20, 21) as there are that note benefits of treatment. Other reports have noted the doubtful clinical significance and frequent statistical nonsignificance of l-thyroxine therapy on changes in LDL cholesterol, myocardial performance, and neuropsy-chiatric parameters. No association between SH and isch-emic heart disease was shown in the Whickham survey (23), the most extensive longitudinal study of thyroid disease ever conducted. Such conflicting findings stem from inconsisten-cies of the reports in the variable definition of SH, the wide degree of thyroid failure examined, as well as the heteroge-neous age, gender, and ethnicity of the subjects tested. There is considerable evidence suggesting that the subjects with SH who would benefit most from l-thyroxine therapy are those with TSH levels exceeding 10 mU/liter. Such in-dividuals constitute the minority of those with SH in all large-scale epidemiological studies that have stratified TSH levels (1, 2, 4 –9, 12, 14). The majority of subjects with SH, however, have slight elevations of TSH ranging between 5 and 10 mU/liter, and they have minimal, often nonsignifi-cant, metabolic abnormalities. They are either affected by mild incipient thyroid failure for which l-thyroxine therapy has not been shown to convey recognizable benefits, or they may simply represent " euthyroid outliers " in the 2.5% tail above the upper limit of the normal TSH reference range, in which case l-thyroxine treatment would be inappropriate. The available evidence also calls into question the need to treat men with SH, who are much less prevalent than women with SH (1, 4, 6, 7, 10, 11, 14), and who manifest almost no metabolic differences compared with men with normal TSH levels. Unfortunately no large-scale multicenter randomized trial has been performed to address and settle these issues. As a result, we are left to struggle with and sort through the conflicting results from a variety of small studies. Given this lack of definitive data, clinicians must weigh the benefit to risk ratio of initiating l-thyroxine therapy in subjects with SH, realizing that lifelong treatment is not without incon-venience or potential morbidity. In the absence of definitive outcome studies, we will conclude that the decision not to treat the majority of SH subjects is as strongly or even more strongly supported by the existent data as is the decision to treat. Thus, the burden of proof rests squarely on the shoul-ders of clinicians who initiate lifelong l-thyroxine treatment for SH to demonstrate that the goals of such therapy are clear and compelling.