Physiological profiles of episodic progesterone release during the midluteal phase of the human menstrual cycle: analysis of circadian and ultradian rhythms, discrete pulse properties, and correlations with simultaneous luteinizing hormone release.

To define the physiological relationships between episodic progesterone and LH release, we measured serum progesterone and LH concentrations in blood sampled at 10-min intervals for 24 h in seven young women in the midluteal phase of the menstrual cycle. The resultant time series were assessed further by Fourier transformation, Cluster analysis, and cross-correlation analysis with autoregressive modeling. These techniques permitted an examination of circadian rhythms, discrete (ultradian) pulse properties, and simultaneous or lagged correlations between progesterone and LH release. We found the following. 1) Both serum LH and progesterone concentrations had significant circadian periodicities, with similar acrophases (times of maximal nyctohemeral values). LH and progesterone also manifested multiple ultradian rhythms of similar periodicities (range, 48-241 min). 2) Discrete serum progesterone peaks occurred at a mean interpulse interval of 118 +/- 12 (+/- SE) min, had durations of 92 +/- 12 min, and had incremental amplitudes of 4.3 +/- 0.9 ng/mL (14 +/- 3 nmol/L). The frequency and duration characteristics of the progesterone and LH peaks were not significantly different, but progesterone fractional peak amplitudes were one quarter those of LH pulses. 3) Fractional progesterone peak amplitudes in the seven women correlated inversely (r = -0.811) with 24-h mean LH concentrations, suggesting a negative feedback relationship between progesterone and LH release. 4) LH and progesterone interpulse intervals both exhibited significant nyctohemeral variations, with diurnal amplitudes of 73 +/- 12 min for LH and 43 +/- 8.9 min for progesterone (P less than 0.01). 5) Significant positive cross-correlations existed in all seven women between serum LH and progesterone concentrations considered simultaneously and at progesterone time lags of 10-50 min. By autoregressive modeling, the later (20-50 min) cross-correlations could be accounted for by sustained autocorrelations in the individual progesterone and LH time series and significant cross-correlations between LH and simultaneous progesterone concentrations and between LH and 10-min lagged progesterone concentrations. We conclude that progesterone release occurs in a periodic (circadian and ultradian) fashion as well as in a discrete (episodic or pulsatile) mode. Moreover, both positive and negative feedback relationships operate to coordinate LH and progesterone secretion in the midluteal phase of the human menstrual cycle.