Studies on the pattern of circulating steroids in the normal menstrual cycle. Circadian variation in theperi-ovulatory period.

Circadian variations of the plasma levels of dehydroepiandrosterone, 17-hydroxypregnenolone, pregnenolone and testosterone were investigated by radioimmunoassay in 10 normally menstruating women during the periovulatory period. In seven of the subjects it was also possible to estimate androstenedione and dihydrotestosterone levels. Blood was withdrawn continuously over a period of 48 h at a rate of 4 ml/h by means of a non-thrombogenic pump. The circadian rhythm was studied during 13 3-hour sampling periods (36 h) which were identical in all subjects. Dehydroepiandrosterone, 17-hydroxypregnenolone and pregnenolone showed a marked circadian rhythm with highest mean levels in the morning between 06.00 and 0.900 h and lowest mean levels during the night between 21.00 and 24.00 h. The peak levels of individual subjects coincided completely with the highest mean levels in the case of dehydroepiandrosterone, in 9 out of 10 cases with regard to 17-hydroxypregnenolone and in 8 of 10 cases as far as pregnenolone levels were concerned. The lowest individual levels were more dispersed around the means than were the individual peaks. The difference between the highest (morning hours: 06.00-09.00) and lowest (evening hours: 21.00-24.00) geometeric mean values was 404% for 17-hydroxypregnenolone, 163% for dehydroepiandrosterone and 71% for pregnenolone. The mean testosterone levels also exhibited an elevation between 06.00 and 09.00 h which was significant (P less than 0.05). However, the individual peak values were scattered from 21.00 h of the first day to 18.00 h of the second day. The mean concentrations of androstendine showed a significant increase between 06.00 and 18.00 h. The individual peaks were widely dispersed. Hence a uniform circadian rhythm correlated to the levels of the 5-steroids mentioned above could not be demonstrated in the case of testosterone and androstenedione. An analysis of variance indicated no significant differences between sampling periods as far as dihydrotestosterone levels were concerned. The extent and regularity of the circadian variation in the plasma levels of the 5-steroids studied makes it mandatory to standardize very carefully the exact time of blood withdrawal in any longitudinal study. In view of the sharp changes in the plasma levels during the morning hours, it is suggested that sampling during the afternoon period may provide more constant values.