Abstract:

Objective To evaluate the consistency of electrochemiluminescence and 2 domestic chemiluminescence determination systems in carbohydrate antigen 72-4（CA72-4）determination. Methods According to the Clinical and Laboratory Standards Institute（CLSI）EP09-A3 document,42 clinical serum specimens in linear range were determined by cobas e601 electrochemiluminescence immunoassay system（reference method,X）,MAGLUMI 4000 chemiluminescence immunoassay system（experimental method 1,Y₁）,CL-2000i chemiluminescence immunoassay system（experimental method 2,Y₂）. The results of outliers were determined by generalized extreme studentized deviate（ESD）method. The optimal regression model was fitted to the regression equation. The biases were calculated at the medical decision level,and the comparability were judged by 1/2 external quality assessment bias（±7.65%） as acceptance standard. Results The correlations between e601（X） and MAGLUMI 4000（Y₁） and between e601（X） and CL-2000i（Y₂）were good by plotting scatter plot. One outlier was visually found in each group. No outlier of CA72-4 was found using generalized ESD method. Deming（X and Y₁）and Passing-Bablok（X and Y₂）were used. The regression equations were Y₁ = 5.020 6+1.149 8X and Y₂= 0.726 3+0.830 0X,respectively. Assuming CA72-4 medical decision level of 6.9 U/mL to regression equation,the proportional bias was 60.96% at e601（X） and MAGLUMI 4000（Y₁）,which was higher than acceptable value（1/2 bias）. The proportional bias was -6.74% at e601（X） and CL-2000i（Y₂）,which was lower than acceptable value（1/2 bias）. Conclusions The results of CA72-4 determination by e601 and MAGLUMI 4000 are not comparable,while results of CA72-4 determination by e601 and CL-2000i are comparable. CLSI EP09-A3 document can be used for comparison between different labeled immunoassay determination systems.

Key words: Carbohydrate antigen 72-4, Chemiluminescence immunoassay, Electrochemiluminescence, Methodology comparison