Serum thyroid-stimulating hormone and cognition in older people.

Background: high TSH concentrations and cognitive decline are both very common among older people and could be linked. Objective: to assess cognition in our cohort of 335 home-dwelling older people (75 years and older) and to cross-sectionally relate the results to thyroid-stimulating hormone (TSH) concentrations. Our special focus was on the upper normal TSH range and subclinical hypothyroidism. Methods: cognitive performance was evaluated using the Consortium to Establish a Registry for Alzheimer's disease neuropsychological battery (CERAD-nb). The Clinical Dementia Rating (CDR) scale was used to evaluate severity of cognitive disorder. The APOE epsilon 4 genotype was also defined. Subjects were divided into quartiles based on the TSH concentrations, and results were compared between these groups. Results: expected relations were observed between CERAD domains and both educational level and APOE epsilon 4 genotype. Female sex significantly associated with better performance in Boston naming (OR = 0.48; 95% CI = 0.27-0.85). In the whole cohort, higher TSH concentrations tended to associate with better scores in most parts of the CERAD-nb tests, but differences were not statistically significant. However, subjects with the highest TSH concentration (90th TSH percentile, range 4.14-14.4 mU/l) had better CDR scores compared with subjects with the lowest TSH concentration (10th percentile, range 0.001-0.63 mIU/l; OR 0.10; 95% CI 0.014-0.76). Conclusions: our results do not support the notion that higher TSH concentrations, not even in the range of subclinical hypothyroidism, would adversely affect cognition among older people.