A-033 Comparison of the Roche hCG STAT and hCG+β Assays

Abstract

Abstract Background Roche Diagnostics offers hCG STAT and Elecsys hCG+β for the quantitative determination of human chorionic gonadotropin (hCG) in serum. While hCG STAT detects intact hCG after a 9-minute incubation, hCG +β; detects intact hCG, nicked hCG forms, the β-core fragment and the free β-subunit after 18 minutes. The manufacturer reported a strong correlation (R = 0.999) between the two assays using 81 samples with hCG concentrations ranging from 6 - 7675 mIU/mL. However, hCG concentrations during early pregnancy are low. This study aims to determine the correlation between the two assays with samples of < 50 mIU/mL hCG concentration. Additionally, we aim to determine the qualitative concordance between the two assays assuming the upper 95% confidence limit reported by the manufacturer as the positive cutoff value. Methods We compared hCG results from a cohort of remnant serum samples (N = 65; hCG 2 - 49 mIU/mL) previously analyzed using the Elecsys hCG + β assay on Roche cobas e 801 analyzers against the Elecsys HCG STAT assay on the cobas e 601 analyzers. A quantitative method comparison was also performed on a subset (N =33) of samples with hCG < 5 mIU/mL. Regression analysis was conducted using EP Evaluator software (version 12). A qualitative concordance between the two assays was performed assuming the upper 95% confidence limit in healthy, non-pregnant premenopausal women as the positive cutoff value (5.3 mIU/mL and 4.9 mIU/mL for the hCG+β and hCG STAT assay, respectively). Results Quantitative comparison using a wider range revealed an R value of 0.968, however, at hCG < 5mIU/ml concentration, the R value was 0.755. Qualitative comparison using each assay’s cutoff value identified seven discrepancies. Three samples that would be classified as positive with the hCG+β assay would be considered negative with hCG STAT assay, and four samples that would be considered negative with the hCG+β assay would be classified as positive with the hCG STAT assay. When a 4.9 mIU/mL cutoff value is used for both assays, eight samples that would be positive with the hCG+β assay would be considered negative with the hCG STAT assay, and one sample goes the other way. Conversely, if 5.3 mIU/mL is used as a cutoff value for both assays, three samples that would be classified as positive with the hCG+β assay would be considered negative with the hCG STAT assay, and one sample tested the other way. Conclusions The correlation between the two assays is strongest when comparison is made using a wide range of hCG values, consistent with the manufacturer's claims. However, the correlation weakens when comparing samples with low hCG values. Qualitative comparison of the assays using their respective cutoff values resulted in 89.2% overall agreement, 84.6% when a lower cutoff value of 4.9mIU/ml is used and 93.8% overall agreement when 5.3mIU/ml cutoff value used. We concluded that while the assays are comparable, minor differences were observed at low hCG values. Further study using a larger sample size and clinical correlation is needed to reinforce the conclusions.