Reliability of estimates of pulsatile characteristics of luteinizing hormone and growth hormone release in women.

The test-retest reliability of estimates of pulsatile LH and GH release was evaluated in 23 eumenorrheic women during the early follicular phase of the menstrual cycle. Each subject was studied during two successive or near-successive menstrual cycles by repetitive blood sampling every 10 min for 24 h. Pulsatile parameters for LH and GH release were identified and characterized using the Cluster pulse detection algorithm. For LH, no significant differences existed in any parameter mean between the two 24-h admissions. Correlation coefficients for consecutive 24-h studies ranged from r = 0.22 (P less than 0.32) for number of LH peaks to r = 0.79 (P less than 0.0001) for 24-h integrated LH values (area under the concentration vs. time curve). No significant mean differences in any parameter were observed for consecutive 24-h GH evaluations. Correlation coefficients for 24-h GH ranged from r = 0.25 (P less than 0.34) for nadir to r = 0.71 (P less than 0.002) for incremental peak increase. Cosinor analysis was used to determine significant 24-h variations in LH and GH concentrations. Statistically significant differences existed between admissions for the amplitude of the nyctohemeral LH rhythm and its acrophase (time at which maximal hormone value was attained), but no mean differences were found for mesor (mean concentration). Correlation coefficients for LH were r = 0.10 (P less than 0.65), r = 0.43 (P less than 0.08), and r = 0.78 (P less than 0.0001) for phase, amplitude, and mesor, respectively. No significant mean differences existed for any parameter of nyctohemeral GH rhythms. Correlation coefficients were r = -0.18 (P less than 0.52), r = 0.49 (P less than 0.72), and r = 0.14 (P less than 0.80) for 24-h GH amplitude, phase, and mesor, respectively. We conclude that comparisons of mean and integrated LH and GH concentrations over a 24-h interval in the early follicular phase of the menstrual cycle are reliable; however, certain pulsatile properties responsible for the achievement of the mean daily concentrations of LH and GH may be nonuniform from menstrual cycle to menstrual cycle. In addition, nonuniformities may exist in the nyctohemeral rhythms of serum concentrations of LH and GH in the adult woman between cycles when a single 24-h time series is the basis for the analysis.