RESUMEN
OBJECTIVE@#To explore the reasons causing the false positive of HBsAg single-ELISA-reactive in blood donors of Jiangsu province so as to provide reference data for the return of blood donors.@*METHODS@#Serological test: HBsAg ELISA parallel detection was performed on 319 444 samples of blood donors from 2014 to 2017; the ECLIA was employed to confirm the single-ELISA-reactive (S/CO≥0.5) samples, the nucleic acid test was used to detect the HBV DNA on the all single-ELISA-reactive samples in 6/8 people mixed/single. Reagent evaluation: the Receiver-Operating-Characteristic curve (ROCC) was drawn by the ECLIA/NAT results as the gold standard, and the diagnostic performance of reagents A and B under different cut-off was evaluated.@*RESULTS@#A total of 227 (0.71‰) single-ELISA-reactive samples were detected among 319 444 blood donors, including 39 cases (17.2%) of positive HBsAg and 12 cases (5.3%) of positive HBV DNA; Under the maximum YI, the COI (1.0) employed by the manufacturer recommendation has a better diagnostic value than laboratory COI (0.5), and the capability of reagent A was better than that of reagent B (AUC: 0.661 vs 0.632; Youden: 0.329 vs 0.297), but the specificity of both reagents was restricted (＜60%). Under the maximum YI, the best cut-off value of reagents A and B were 2.4 and 1.4 COI, respectively. Compared with the cut-off value of manufacturer, the sensitivity of reagents A decreased by 33% and the false positive rate decreased by 60% while the sensitivity of reagent B increased by 140% and the false positive rate increased by 36%, respectively.@*CONCLUSION@#The false positive of HBsAg single-ELISA-reactive in blood donors is caused by the limited specificity of ELISA reagent and the setting of COI values. According to ROCC maximum YI method, the COI can be set as 2.4 COI and (0.5-1.4) COI for reagent A and B to reduce false positive rate.