In an attempt to determine the significance of low plasma thyrotropin (TSH) concentrations in internal medicine and the usefulness of systematic TSH assays in hospitals, 732 consecutive TSH measurements were performed in first-admission patients. TSH concentrations below 0.15 mU/l were found in 33 patients (4.5%) divided into 4 groups: a) in 5 patients a second assay made within 10 days of the first one showed no fall in TSH levels; b) 5 patients had known endocrine disease; c) in 8 patients hyperthyroidism could be asserted; the diagnosis had not been suspected in 3 elderly women and 1 pregnant women; d) 15 patients remained with low TSH concentrations but had normal free T3 and free T4 levels; in this group a goitre was detected in 7 patients and 8 had a severe chronic disease. These results showed that a TSH concentration below 0.15 mU/l corresponded to hyperthyroidism in less than one out of three patients in this population and that the 0.07 to 0.15 mU/l range is particularly misleading. A second TSH assay, free T3 and free T4 measurements ant thorough investigations in search of a goitre must be made. Severe organic diseases and several drugs may induce a fall in TSH. All considered, the 1% prevalence of hyperthyroidism in this population does not justify systematic TSH assays, but in subjects over 60 years of age, the clinical manifestations of hyperthyroidism may be misleading or unrecognized, and TSH assays should be widely performed.