Circadian and pulsatile TSH secretion under physiological and pathophysiological conditions.

In addition to the well known circadian rhythm of TSH secretion a pulsatile pattern of release has been shown. Analysis of the pulsatile release by different computer-assisted methods revealed systematic differences in the number and distribution of TSH pulses. Using the same approximation of false positive pulses (less than 1%) in any of the 21 healthy male volunteers tested a lower number of pulses was found by the Pulsar method (mean 5.1 +/- 2.0/24 h) than by the Cluster (10.6 +/- 1.8) or the DESADE program (13.6 +/- 4.6). The results of the Cluster and Desade analysis fit well to that of Fourier transformation which revealed a dominant frequency at 160 min. In addition dominant frequencies in comparison to a noise series were found at 24 h and at 33 min. Analysis of the data in 8 h segments between 2000 and 0400 h, 0400 and 1200 h and 1200 and 2000 h by Desade and cluster revealed that app. 50% of pulses occurred between 2000 and 0400 h, suggesting an important role of pulsatile TSH release in the generation of the circadian TSH rhythm. In 3 patients with TSH-induced hyperthyroidism the circadian and pulsatile pattern of TSH secretion was similar to that in healthy controls. In contrast, in a patient with a TSH producing pituitary tumor the circadian variation of TSH secretion was abolished. Patients with a non-toxic goitre revealed a significantly lower mean TSH serum level as the control group of healthy subjects. The number of TSH pulses was slightly but significantly lower when analyzed by the DESADE program but not when analyzed with any other method.