An ultrasensitive immunoradiometric assay (IRMA) using two monoclonal anti-TSH antibodies has been used for TSH measurements in basal conditions and after TRH stimulation. The results have been compared with those obtained by conventional radioimmunoassay (RIA). The IRMA method had very high sensitivity (0.07 microU/ml). Detectable serum TSH concentrations were found in all normal subjects by IRMA, but in only 76% by RIA. No overlap was observed with the results obtained by IRMA in untreated overtly hyperthyroid patients, in whom serum TSH was below the limit of detection. The relationship between basal and TRH-stimulated serum TSH concentrations by IRMA and RIA was evaluated in 176 subjects including normals and patients with untreated and treated hyperthyroidism, functioning thyroid adenoma, nontoxic goitre and patients on L-thyroxine therapy. A normal TSH response to TRH was observed in virtually all patients with detectable basal serum TSH by both methods. When patients with undetectable basal serum TSH levels were considered, all but one (98%) had no TSH response to TRH by IRMA. On the contrary using RIA, an absent response was found only in 47% of subjects, a blunted responses in 10% and a normal response in 42%. These data indicate that basal serum TSH measurements by IRMA allows a complete discrimination of normal from hyperthyroid patients and can avoid the need for TRH stimulation tests.