The prediction and/or detection of ovulation by means of urinary steroid assays

Twenty normally menstruating women volunteered for a study in which plasma samples were collected daily during an entire menstrual cycle. On the same days, samples of morning urine were also collected, as well as random samples of urine voided at the visit to the Outpatient Clinic. Progesterone (P), estradiol (E2) and lutropin (LH) were assayed in plasma, and pregnanediol-3-glucuronide (PdG), estrone-glucuronide (E1G), estriol-16-glucuronide (E3G), P, and E2 were measured in urine using radioimmunoassays. Progesterone in urine was assayed both with and without preceding chromatography. All urinary glucuronides and progesterone exhibited cyclic patterns similar to those of E2 or P in plasma. Seven-fold increases from early follicular to luteal phase values (for PdG and urinary P; the latter both with and without chromatography), or to peak levels (for E1G and E3G) were observed. The difference between the baseline and peak levels was less distinct (approximately 5-fold) for E2 in urine. The day-to-day coefficient of variation of early follicular phase values decreased from 40% to 25% by calculating the ratios of the glucuronides or P to creatinine (C). The peaks of estrogen glucuronides were delayed mostly by 1 day in comparison to the peaks of E2 in plasma. The urinary peaks of estrogens were in most cases more closely clustered around the day of the LH-peak when the measurements were corrected for C. For the determination of the first significant rise of steroid levels in a cycle, the calculation of a sustained rise (leading to a significant cumulative sum - CUSUM) was found superior when compared to other recommended indices, such as a 50% increase over the mean of 3 preceding values, or the increase over the baseline level plus 2 standard deviations. Sustained rises were calculated for all indices studied (including the ratio of urinary E1G to PdG). The ratio of E1G to C in morning urine gave consistently the most compact distribution of sustained rises. It is concluded that daily measurements of urinary PdG (or P) and E1G (or, possibly, E2) could substitute the serial assays of P and E2 in peripheral blood in the retrospective assessment of the ovarian functionn. The day-to-day variation can be significantly reduced, if results are expressed per concentration of C. For the prediction of ovulation or fertile period, the best index of urinary steroids appears to be the sustained rise in the ratio of E1G to C. However, this "best" method is still not good enough in terms of overall reliability and practicability.