The magnitude and temporal pattern of serum TSH suppression after single or multiple doses of thyroid hormone (T-3, T-4, or triiodothyroacetic acid) were studied using third and fourth generation TSH assays (sensitivities, 0.01 and 0.001 mU/L, respectively). A constant T-3 dose (263 mu g iv) administered at a uniform clock time (1200 h) produced identical serum TSH suppression patterns, (percent of control TSH vs. hours) in euthyroid and hypothyroid subjects. The percent log TSH vs. log time plot revealed three temporally distinct linear suppression phases: phase 1, a rapid TSH suppression, onset 1 h and lasting for 10-20 h; phase 2, slower suppression, onset between 10 and 20 h and lasting for 6-8 weeks; and phase 3, an invariable low TSH level (<0.01 mU/L) with chronic T-3 suppression (100 mu g four times a day). TSH escaped maximal suppression at a similar serum T-3 level in both euthyroid and hypothyroid subjects (2.9 +/- 0.2 vs. 3.5 +/- 0.5 nmol/L, respectively; P > 0.9), despite different basal serum T-3 values (2.0 +/- 0.1 vs. 0.6 +/- 0.1 nmol/L, respectively; P < 0.01). Two milligrams of triiodothyroacetic acid or 2 mg T-4 given iv at 1200 h produced TSH suppression patterns similar to T-3. The phase 1 suppression varied with the clock time of T-3 administration, (steeper responses were seen at 2400 vs. 1200 h), whereas phase 2 responses were unaltered. This study shows that thyroid hormone suppression of TSH is a complex, biphasic, nonlinear process, which is reproducible and independent of thyroid status or the thyroid hormone analog used. It is hypothesized that phase 1 reflects inhibition of release of preformed hormone, whereas phase 2 likely reflects inhibition of de novo synthesis and/or thyrotroph storage of TSH. In contrast, phase 3 secretion seems to represent basal constitutive TSH release, which may have relevance to the role of thyroid hormone-suppressive therapy in the treatment of patients with benign or neoplastic thyroid disease.