Abstract
The analytical performance of four radioimmunoassays for the determination of progesterone in unextracted serum is evaluated. The methods include kits from Radioassay Systems Laboratories (RSL), Nordiclab, Behringwerke, and an in-house assay. Extraction of progesterone from serum is avoided in these methods by employing various agents to displace progesterone from binding proteins. The radioimmunoassays employ 125I-labelled progesterone derivatives, and bound and free fractions are separated by double antibody, polyethylene glycol or charcoal methods. The in-house method gives essentially quantitative recovery of progesterone, and results of patients' specimens correlate well with those obtained by a reference method involving extraction of serum. The Nordiclab kit gives progressive under-recovery of progesterone with increasing progesterone concentration, and results of patients' specimens correlate poorly with those obtained by other methods. Gross over-recovery of progesterone is obtained with the RSL kit, and results of patients' specimens are approximately three-fold higher than those of the reference method. The Behringwerke kit is precise but biased, progesterone levels being over-estimated in both recovery pools and patients' specimens. The performance of these methods and their suitability for the assessment of luteal function are discussed.