Poster Presentation ESA-SRB Conference 2015

Background: Progesterone (P4) secreted by the corpus luteum is essential for implantation and early pregnancy. Serum P4 measurement on day of hCG administration during IVF controlled ovarian stimulation has been proposed to identify premature ovulation and/or luteinisation with an adverse impact on pregnancy in that IVF cycle.

Objective: To evaluate the accuracy of serum P4 measured by direct (unextracted) immunoassay (IA) vs a liquid chromatography mass spectrometry (LC-MS) reference method.

Method: Serum samples were collected from 254 women (median age 38, range 20-49 yr) on hCG day during an IVF cycle. Serum P4 was measured by IA (Beckman Coulter Access) and by LC-MS with results compared by Bland-Altman [BA], Passing-Bablok [PB] and Deming [D] regression methods. For analysis, left-censored (undetectable) results in LC-MS were assigned a value half of the detection limit (0.05 ng/ml).

Results: IA over-estimated serum P4 in every sample (median 4.8 vs 1.5 nM; median difference 4.4 nM [interquartile range 3.5, 5.9 nM]). Serum P4 was detected in 252 (99%) by IA and in 215 (85%) of samples by LC-MS. By PB regression, the intercept was 3.2 nM (95%CI 3.1, 3.3 nM) with a slope of 1.0 (95%CI 0.9, 1.1). By D, the intercept was 3.6 nM (95%CI 3.5, 3.8 nM). The upward bias of IA increased exponentially at low serum P4 concentrations (IA <5 nM or LC-MS<2 nM). Age was unrelated to either assay result or their difference.

Conclusion: IA consistently overestimates serum P4 levels so that low measurements (IA<5 nM) are too inaccurate to be used quantitatively. The utility of higher serum P4 measurements by IA and serum P4 and other steroids measured by multiplex LC-MS profiling in predicting IVF pregnancy outcomes warrants further investigation.