Context: Both hypothyroidism and subclinical hyperthyroidism hinder cognitive function.
Objective: We aimed to determine whether more subtle alterations of thyroid hormone levels predict increased incidence of dementia in aging men.
Participants and design: Community-dwelling men aged 70-89 yr participated in this prospective longitudinal study.
Main outcome measures: The Standardized Mini-Mental State Examination was performed at baseline (2001-2004), and circulating TSH and free T(4) (FT(4)) were assayed. Men with known thyroid disease or dementia, or Standardized Mini-Mental State Examination scores below 24 were excluded from follow-up. New-onset dementia, defined by International Classification of Disease (ICD) codes, was ascertained using data linkage (2001-2009).
Results: During follow-up, 145 of 3401 men (4.3%) were diagnosed for the first time with dementia. Men who developed dementia had higher baseline FT(4) (16.5 ± 2.2 vs. 15.9 ± 2.2 pmol/liter, P = 0.004) but similar TSH (2.2 ± 1.4 vs. 2.3 ± 1.6 mU/liter, P = 0.23) compared with men who did not receive this diagnosis. After adjusting for covariates, higher FT(4) predicted new-onset dementia (11% increased risk per 1 pmol/liter increase in FT(4), P = 0.005; quartiles Q2-4 vs. Q1: adjusted hazard ratio = 1.76, 95% confidence interval = 1.03-3.00, P = 0.04). There was no association between TSH quartiles and incident dementia. When the analysis was restricted to euthyroid men (excluding those with subclinical hyper- or hypothyroidism), higher FT(4) remained associated with incident dementia (11% increase per unit increment, P = 0.03; Q2-4 vs. Q1: adjusted hazard ratio = 2.02, 95% confidence interval = 1.10-3.71, P = 0.024).
Conclusions: Higher FT(4) levels predict new-onset dementia in older men, independently of conventional risk factors for cognitive decline. Additional studies are needed to explore potential underlying mechanisms and to clarify the utility of thyroid function testing in older men at risk of dementia.