Cognitive Function in Subclinical Hypothyroidism

Abstract

Over the past 20 yr, there have been major advances in understanding the neural basis for cognitive processes (1, 2). Cognitive domains include attention and concentration, language, memory, psychomotor function, and executive function. These domains map to neural systems that involve multiple, often overlapping brain regions. There are validated neurocognitive tests to measure these domains, mapped to critical brain regions by lesional studies and functional imaging. Because thyroid hormone has profound effects on the central nervous system, it is logical to ask how hypothyroidism affects cognition. The consequences of overt hypothyroidism (particularly congenital hypothyroidism) are well known and include widespread cognitive deficits that can affect all the domains listed above (3). In contrast, there is less evidence regarding cognitive effects of subclinical hypothyroidism. This unresolved issue is a problem in clinical practice because subclinical hypothyroidism is prevalent in older patients, many of whom already have some cognitive decline. It is not clear whether these patients should be treated to ameliorate further cognitive dysfunction. Two recent systematic literature reviews did not find enough evidence to recommend treatment of subclinical hypothyroidism based on cognitive effects (4, 5). The existing literature on subclinical hypothyroidism can be divided into larger cross-sectional or longitudinal studies and smaller interventional [levothyroxine (L-T4) treatment] studies. Many older studies of either type failed to find decrements in cognitive domains, although these were often limited by small sample sizes, heterogeneous subjects, or limited cognitive tests. More recent studies have been larger and/or have used more modern, validated cognitive tests. The most rigorous studies have also controlled for mood alterations, which can occur in thyroid dysfunction and which affect cognitive measures. Most of the recent cross-sectional or longitudinal natural history studies have failed to find significant cognitive effects of subclinical hypothyroidism at baseline or developing over time (6–11). The largest of these was reported by Roberts et al. (7) in 2006. This was a well-conducted, population-based study of almost 6000 subjects, all at least 65 yr old. They underwent an extensive screening battery for cognitive function. No differences were found between subjects with subclinical hypothyroidism and euthyroid subjects. However, the cognitive batteries in this and other large studies were often designed to detect gross impairment in elderly subjects. Therefore, they might not be expected to detect subtle deficits induced by mild thyroid disease. In other cases, more sensitive cognitive tests were used but did not necessarily target the cognitive domains most likely to be affected by subclinical hypothyroidism, based on animal studies of thyroid hormone and its receptor distribution in the brain (12, 13). There are only a few recent intervention studies that used highly sensitive tests to assess possible effects of L-T4 treatment on cognitive outcomes. The most influential is the study in Tromso, Norway (14). Subjects with subclinical hypothyroidism underwent detailed cognitive testing of multiple domains at baseline and after treatment with placebo or L-T4 for 12 months. There were no baseline differences between subclinical hypothyroid subjects and matched euthyroid controls, and there was no effect of L-T4 treatment on any of the measures. This study concluded that patients with subclinical hypothyroidism have no cognitive deficits, which is well supported by their data. The major limitation was the relatively mild degree of subclinical hypothyroidism, with TSH levels only up to 10 mU/liter. In a smaller recent intervention study, Correia et al. (15) reported that subjects with subclinical hypothyroidism had impaired spatial and verbal memory on detailed cog-