Endocrine Abstracts

Background: Urinary concentrations of the major progesterone metabolite pregnanediol glucuronide (P3G) are used clinically and in research to monitor ovulation. This ELISA is laborious and costly. If it could be replaced by the automated sensitive chemiluminescence assays routinely used for serum this would be of great value.

Objective: We aimed to determine the validity of urine progesterone as measured by two widely used automated assays in comparison to the standard P3G assay.

Methods: Daily urine aliquots were obtained across 20 cycles (median (range) length 28(25–37) days in 14 women mean (SD) age 33.5(6.6) years. Ovulation was confirmed in all cycles by transvaginal ultrasound and serial LH measurement. Urine progesterone was measured and corrected for creatinine (uP4-Cr), by automated Abbott chemiluminescent microparticle immunoassay (CMIA) or Roche electrochemiluminescence immunoassay (ECLIA), compared with an in-house P3G ELISA (uP3G-Cr). Midfollicular (LH surge day-3 to day-10) and midluteal (LH surge day+3 to +10) phases were compared. Sensitivity and specificity were calculated by comparing midfollicular and midluteal samples, using a cutoff 1.5-fold increase in concentration.

Results: There was a luteal rise in all cycles in uP4-Cr (both assays) and in 19 cycles (95%) in uP3G-Cr. The median (range) luteal rise was 3.52 (1.16–7.92) for CMIA, 1.66 (1.09–2.45) for ECLIA and 4.81 (0.78–11.70) for uP3G-Cr. Aberrant rises in uP4-Cr above the luteal threshold were seen around the follicular phase with CMIA but not ECLIA.

Conclusions: Automated CMIA but not ECLIA progesterone assay demonstrated marginally inferior sensitivity to detect ovulation and superior specificity, compared with P3G manual ELISA. The reasons ECLIA showed spurious follicular rises, poorer correlation with P3G, sensitivity and specificity are not clear but may include a matrix effect. Serial urinary progesterone using ECLIA chemiluminescence demonstrates potential as an alternative to P3G.