TRH stimulation tests (n = 1109) were performed on 1061 ambulatory and 43 hospitalized patients with varying thyroid status, using a TSH immunochemiluminometric assay with third and fourth generation sensitivity characteristics (functional sensitivity, 0.01 and 0.001 mU/L, respectively). TRH test results were analyzed as both absolute (stimulated minus basal TSH) and fold (stimulated/basal TSH) responses. The absolute TRH response varied 8-fold across the physiological TSH range, whereas the mean fold response remained almost constant (mean +/- SEM, 8.5 +/- 0.2). The fold response became progressively attenuated as basal TSH values declined below physiological levels, becoming essentially absent in clinically thyrotoxic patients with markedly depressed basal serum TSH levels (0.007 +/- 0.002 mU/L). Progressive attenuation also occurred at hypothyroid TSH levels; a markedly impaired fold response (2.5 +/- 0.4) was characteristic of primary hypothyroid patients with basal TSH values greater than 50 mU/L. In untreated central hypothyroid patients with near-normal basal TSH levels, the TRH fold response was impaired (1.7 +/- 0.2), whereas in T4-replaced central hypothyroid patients, fold responses were near normal (5.6 +/- 1.2). Neither nonthyroidal illness, age, or sex appeared to influence the pattern of fold TRH response in the populations evaluated. When using third and fourth generation TSH methodology, the TRH-stimulated TSH fold response is more diagnostically useful than the absolute TRH response. However, if patients have an intact hypothalamic-pituitary axis, there appears to be no diagnostic advantage gained by TRH testing over an accurately measured basal TSH value.