Abstract
Hyperthyroidism and subclinical hyperthyroidism, the latter defined as a subnormal serum thyrotropin (TSH) and normal serum-free thyroxine (fT4) and triiodothyronine (T3), are both common in the general population (1). Both conditions are caused by the usual entities that are seen daily in endocrine practices: Graves' disease, toxic multinodular goiter, and solitary autonomously functioning nodules (2). Many studies have confirmed that thyroid over activity can lead to adverse health outcomes, especially in older persons, including cardiovascular events and atrial fibrillation (3), low bone mass and fractures (4), and dementia (5,6). While a link between thyroid over activity and arrhythmias and increased bone loss seems logical given the known effects of thyroid hormones (THs) to increase cardiac function and bone turnover, the reasons for a possible connection to dementia remain obscure. Furthermore, a relationship between cognition and thyroid over activity has, until now, only been observed in persons with subclinical hyperthyroidism (5,6), and not in overtly hyperthyroid patients, which seems somewhat illogical.
In this issue of Thyroid, investigators from the Odense University Hospital in Denmark utilized two cohorts to further explore the relationship between hyperthyroidism and dementia (7). One cohort, the Danish National Patient Registry (DNPR), includes >99% of all hospitalizations and outpatient visits in Denmark since 1977 and 1995, respectively, and includes ∼56,000 individuals with a diagnosis of either Graves' disease or toxic nodular goiter. In this cohort, thyroid function data were not available. The other cohort (called the OPENTHYRO registry) consists of all persons who had at least one TSH measurement in the Clinical Biochemistry Laboratory of the Odense University Hospital between 1995 and 2012. For this particular study (7), ∼2700 patients with at least 2 serum TSH measurements <0.3 mU/L at least 14 days apart, and within 6 months of each other, were defined as being hyperthyroid. Overt hyperthyroidism was defined as a subnormal serum TSH and fT4 and/or T3 above their respective reference ranges, and subclinical hyperthyroidism was defined as a subnormal serum TSH and fT4 and T3 levels that were within their respective reference ranges. For both cohorts, persons were considered to have dementia if they had been given a diagnosis of dementia or were prescribed a cholinesterase inhibitor used to treat dementia at a time after the diagnosis of thyroid dysfunction had been made.
Compared with a matched euthyroid control population from the DNPR registry, after a median follow-up time of 7.9 years, and after adjusting for pre-existing morbidity, the hazard ratio (HR) for a dementia diagnosis in the hyperthyroid group was 1.13 [CI: 1.07–1.19]. There was no difference in the risk according to the etiology of hyperthyroidism and no differences in a subanalysis according to type of dementia (vascular dementia vs. Alzheimer's disease). In contrast, in the OPENTHYRO registry, after adjusting for pre-existing morbidities, there were no differences in the risk of dementia in patients with hyperthyroidism compared with euthyroid persons (HR 1.06 [CI 0.89–1.26]). However, the cumulative incidence of dementia over time was higher in hyperthyroid patients than in euthyroid controls, especially after prolonged follow-up (>10 years). In the OPENTHYRO subgroup with overt hyperthyroidism, the risk of dementia was the same as in euthyroid controls (5.5% vs. 4.4% in controls; HR 0.82 [CI 0.66–1.10]), whereas it was higher in patients with subclinical hyperthyroidism (9.1%; HR 1.57 [CI 1.16–2.13]). Unfortunately, the study did not present dementia diagnosis data in persons with “mild” subclinical hyperthyroidism (serum TSH 01.–0.45 mU/L) versus those with “severe” subclinical hyperthyroidism (serum TSH <0.1 mU/L).
The authors conjecture that this discrepancy may be due to the fact that overt hyperthyroidism is almost always promptly recognized and treated, while subclinical hyperthyroidism may not be treated for prolonged periods of time, thereby exposing the patient longer to an abnormal, although less deranged hormonal milieu. This is a reasonable, although unproven, explanation for the discrepancy between overt and subclinical hyperthyroidism reported in this, as well as in previous, studies. It is also possible that some patients with overt hyperthyroidism in these cohorts are undertreated for various reasons and would then be classified as having subclinical hyperthyroidism. The observation that the risk of dementia increased with duration of low TSH levels is concordant with this scenario.
Almost all of the studies of this topic published to date utilized clinical criteria or global cognitive screening tests to diagnose dementia (the Mini-Mental State Exam). These instruments are useful in large groups of subjects but may miss subtle defects in specific cognitive domains. Complementing these reports are more intensive smaller scale studies in patients with altered thyroid status. These studies generally show decrements in attention/concentration and executive function in overt hyperthyroidism (reviewed in ref. 8). Structural and functional brain imaging studies support these findings, as hyperthyroid patients have decreased gray matter volume, resting state functional connectivity, and metabolism of glucose, glutamate and choline in brain regions associated with attention, memory, decision making, visual processing, and affective modulation (9 –13)—all cognitive domains affected in dementia. A few studies examined patients after treatment of overt hyperthyroidism and reported improvements in cognitive measures and functional brain imaging (8,13). Unfortunately, there is a paucity of data regarding cognitive function in subclinical hyperthyroidism. Several observational studies have reported cognitive decrements, while others have failed to find any associations (8). There are no treatment studies of endogenous subclinical hyperthyroidism and cognitive outcomes.
A number of possible mechanisms may underlie the putative effects of excess TH on brain function, although none are proven. Depletion of acetylcholine and presynaptic cholinergic metabolites in the cerebral cortex and hippocampus has been described in both Alzheimer's disease and cognitively impaired hyperthyroid patients (reviewed in ref. 14). In addition, TH may increase oxidative stress in the brain (reviewed in ref. 15). TH has complex tissue-specific effects on the formation of reactive oxidation products and antioxidant processes, which can lead to increased or decreased oxidative stress. In the brain, administration of TH increases the respiratory rate of neuronal mitochondria, which would be expected to increase oxidative stress. Abnormal levels of oxidative stress biomarkers have been reported in the brain and peripheral circulation in Alzheimer's disease, providing a possible link between thyrotoxicosis and the development of dementia (reviewed in ref. 15).
However, arguing against these possible mechanisms, T4 administration actually enhances spatial learning, memory, cholinergic function, antioxidant enzyme levels, adenosine triphosphate content, and neuronal survival in a mouse model of Alzheimer's disease (reviewed in refs. 15,16). A third possible mechanism is through modulation of β-amyloid deposition by TH. However, T3 negatively regulates transcriptional activity of β-amyloid precursor protein, arguing against this mechanism (reviewed in ref. 15). Taken together, these studies do not yet provide a clear mechanistic rationale for the observational studies linking subclinical hyperthyroidism to the development of dementia.
Current clinical practice guidelines from the American and European Thyroid Associations (17,18) do not recommend treatment of subclinical hyperthyroidism in asymptomatic individuals aged <65 years, especially with serum TSH levels between 0.1 and 0.45 mU/L. The studies by Folkestad et al. and others (5 –7) raise the question as to whether treatment to prevent dementia might be a consideration. While the evidence continues to accumulate that subclinical hyperthyroidism is a risk factor for the development of dementia, its mechanism(s) are not clear and, therefore, it is not clear whether treating subclinical hyperthyroidism will have any beneficial effect. In the absence of those data, and the unlikelihood of conclusive treatment trials, it is reasonable to screen people with cognitive impairment for thyroid dysfunction, and to consider treatment of subclinical hyperthyroidism if it can be done safely. For now, treating younger asymptomatic persons who have subclinical hyperthyroidism with the goal of preventing cognitive impairment seems premature.