Interpopulation, interindividual, intercycle, and intracycle natural variation in progesterone levels: a quantitative assessment and implications for population studies.

Methodological challenges in studying sex steroid hormones in premenopausal women result from the existence of variation at three levels: among women from the same population, among menstrual cycles recorded for women at different times of the year, and among days of the same cycle. We partitioned, for a Polish rural population, the natural, nonpathological, variation in salivary progesterone concentrations (measured during 14 days of the luteal phase) into the intracycle component (which accounts for 65% of the total variation) and the among-cycle component (the remaining 35% of the total variation). Out of the among-cycle variation in salivary progesterone, as much as 46% is expressed as differences among individual women (interindividual component); the remaining 54% of variation is due to differences among cycles of individual women (intercycle, within-women component). Such intercycle variation is probably caused by a seasonality of agricultural workload and is much higher than in nonseasonal, industrial populations. We also used bootstrap analyses to generate heuristic recommendations for choosing sample sizes of the number of subjects, number of cycles per woman, and number of days per cycle. Studies in populations with seasonal lifestyles should rely on measurements of at least three cycles per woman. Given the substantial intracycle amplitude in hormone levels to reliably assess biologically and medically relevant variation in ovarian function, at least 7-8 days/cycle should be measured. (c) 2007 Wiley-Liss, Inc.