Redefining "Normal" for hCG in postmenopausal women: 97.5th percentile reference intervals and 99th percentile thresholds across four chemiluminescence immunoassay systems.

The study provides system-specific reference intervals for serum hCG in postmenopausal women, which may help reduce misclassification of hCG elevations.

Summary and findings

This study measured serum hCG levels in 460 healthy postmenopausal women using four different immunoassay systems to establish reference intervals and thresholds. The 97.5th percentile reference intervals were found to be < 7.1, < 8.2, < 5.7, and < 5.9 IU/L for Systems A-D, respectively, while the 99th percentile thresholds were < 8.8, < 9.3, < 7.1, and < 7.8 IU/L. The findings aim to reduce misclassification of hCG elevations in this population.

Abstract

Background and objective Physiological human chorionic gonadotropin (hCG) elevations in postmenopausal women are prone to misclassification as pathological. Substantial inter-system variation exists among immunoassays, and system-specific reference intervals (RIs) for this population have not been established. This study aimed to establish 97.5th percentile RIs and 99th percentile thresholds for serum hCG in healthy postmenopausal women across four chemiluminescence immunoassay (CLIA) systems. Methods A total of 460 postmenopausal women were enrolled. Serum hCG was measured using four immunoassay systems: Beckman UniCel DxI 800 (A), Siemens Atellica IM 1600 (B), Roche Cobas e 801 (C), and Abbott ARCHITECT i2000SR (D). Analytical precision was evaluated according to Clinical and Laboratory Standards Institute (CLSI) EP15-A2. Samples with hCG > 5.0 IU/L (> 6.0 IU/L for System B) underwent polyethylene glycol (PEG) precipitation to exclude macro-complex interference. Spearman rank correlation was used to assess correlations between hCG and age, follicle-stimulating hormone (FSH), and luteinizing hormone (LH). One-sided upper reference limits (97.5th and 99th percentiles) were estimated using the non-parametric method. Results 355 subjects were included in the final analysis. The imprecision of all four systems met the analytical performance criterion. Serum hCG showed poor positive correlations with age (Spearman r = 0.146-0.198, all P < 0.01). Fair correlations were observed between hCG and FSH (r = 0.341, 95% CI: 0.246-0.430) and between hCG and LH (r = 0.428, 95% CI: 0.339-0.509). The 97.5th percentile RIs were < 7.1, < 8.2, < 5.7, and < 5.9 IU/L for Systems A-D, respectively; the 99th percentile thresholds were < 8.8, < 9.3, < 7.1, and < 7.8 IU/L. System B consistently yielded the highest thresholds, while System C produced the lowest. Conclusion The established 97.5th percentile RIs and 99th percentile thresholds provide reliable, system-specific quantitative criteria for differentiating pituitary-derived from pathological hCG elevations.

Limitations

Not reported in abstract.

Elsewhere in the HCG (Human Chorionic Gonadotropin) corpus