Thyroid disease in the elderly. Part 2. Predictability of subclinical hypothyroidism.

BACKGROUND The purpose of this study was to determine how well hypothyroidism could be predicted from clinical signs and symptoms in elderly patients with no known history of thyroid disease. METHODS This retrospective study was conducted in a primary care geriatrics clinic. Two hundred eighty-three patients (205 women, 78 men), who were between the ages of 60 and 97 years of age and had no history of thyroid disease, were included in the study. Medical records of patients were reviewed for data obtained at their first visit to the clinic: serum thyroid-stimulating hormone (TSH), free thyroxine (T4), height, weight, demographic variables, clinical signs and symptoms of hypothyroidism, previous history of thyroid disease and treatment with thyroid medications, and current medications. RESULTS Of the 283 patients with no previous history of thyroid disease, 15.4% of men and 14.6% of women had subclinical hypothyroidism (TSH levels of 5.0 to 14.9 microU/mL, and normal free T4 levels [0.7 to 2.0 ng/dL]). Overt hypothyroidism (TSH > or = 15.0 microU/mL, low free T4 < 0.7 ng/dL) was discovered and subsequently treated in one male and two female patients. There were no significant differences (P > .05) in the frequencies of any of the clinical signs and symptoms of hypothyroidism between euthyroid and hypothyroid patients. There was no significant relationship between TSH levels and the total number of hypothyroid symptoms experienced by all patients (r = -.004, P = .99). Logistic regression analyses determined that clinical signs and symptoms were poor predictors of subclinical hypothyroidism in these elderly patients. CONCLUSIONS Thyroid status could not be predicted from clinical signs and symptoms in this sample of elderly community-dwelling patients. This finding substantiates the difficulty of diagnosing subclinical hypothyroidism in the elderly based solely on clinical features.