Serum TSH and free T4 were determined by chemiluminometric assays in 601 women and 285 men aged 85 years from the population study "70-year-old people in Gothenburg, Sweden". For individuals with serum TSH concentration above 6.0 mU/l, "antimicrosomal" antibodies were determined, to assess the etiology of the elevated TSH concentration. Clinical follow-up was done of survivors until the age of 88, and records were inspected also for individuals who died before that age. On the basis of these evaluations the prevalence of previously undetected hypothyroidism was estimated to 4.0% in women and 2.5% in men. Previously undetected hyperthyroidism was found in 2, at the most 4, out of 601 women; in one out of 285 men the diagnosis could not be excluded. Possible confounding factors for the evaluation of TSH and/or free T4 concentrations were analysed by permutation t-test followed by multiple regression analysis, which revealed correlations of log TSH concentration to body mass index (p<0.05), serum creatinine concentration (p<0.05), and diabetes mellitus (inverse relationship, p<0.01). Correlation was found of free T4 concentration to treatment with non-selective β-blocking agents (p<0.001) and digitalis glycosides (p<0.01). However, none of the factors influencing TSH and/or free T4 concentrations had any major influence on reference limits, nor could they account for individuals with "outlier" values.