Background: Human chorionic gonadotropin-beta (β-hCG) is an important index used to monitor embryonic development following embryo transfer. Architect i2000sr and Cobas e601 are widely used automated immunoassay systems used to measure serum β-hCG concentrations; however, the correlations between serum β-hCG levels measured with these two immunoassays and the accuracy of the immunoassays have not been fully evaluated.
Methods: Serum β-hCG levels were measured in 133 serum samples using the Architect i2000sr and Cobas e601 automated immunoassay systems. Passing-Bablok regression analysis was used to compare the correlation in serum β-hCG levels obtained using the two immunoassays. A Bland-Altman plot analysis was used to identify mean ratios and 95% CIs of the mean ratios of the β-hCG results between the two immunoassays. In this graphical method the mean ratios between the two techniques were plotted against the averages of the two techniques.
Results: The total coefficients of variations (CVs) of serum β-hCG ranged from 3.12 - 4.66% for Cobas e601 and 3.18 - 4.99% for Architect i2000sr. The measured value of serum β-hCG detected by the two immunoassays was statistically significant (p < 0.001). The Passing-Bablok regression analysis showed good correlation between the serum β-hCG values measured using the two systems. At a low concentration of serum β-hCG (< 10000 IU/L, n = 52), the correlation coefficient r was 0.9628. At a high concentration of serum β-hCG (> 10000 IU/L, n = 81), the correlation coefficient r was 0.8076. The Bland-Altman plot analysis showed that the measured value of serum β-hCG detected by Architect i2000sr was about 1.25 times higher than that of Cobas e601. The mean ratio was 1.12 at a low concentration of serum β-hCG, and it was 1.33 at a high concentration.
Conclusions: Architect i2000sr and Cobas e601 have good concordance for determining serum β-hCG. However, the β-hCG values measured with Architect i2000sr were 25% higher than those obtained using Cobas e601.