RESUMO
【Objective】 To explore the significance of blending internal controls by automatic sample processing instruments in the enzyme linked immunosorbent assay (ELISA). 【Methods】 The internal controls were vortexed and mixed before the test, and then were added to the same ELISA plate by the STAR automatic sample processing instruments under the same detection conditions. The difference of S/CO value of internal controls with and without sufficient blending via the sampling needle and their frequency distribution were compared. Internal controls that were greatly affected by mixing parameters were submitted to the same test with different batches of reagents from the same manufacturer, and the results were analyzed for consistency. 【Results】 The S/CO value of anti-HCV internal controls without blending using adding sample needle was significantly lower than that of quality control samples with sufficient blending (P<0.000 1). The S/CO values of unmixed internal controls concerning anti-TP and anti-HIV detection given by some detection systems were also different from the values of mixed internal controls (P<0.05). Some of the S/CO values of the anti-HCV internal controls without mixing were distributed within the interval of less than 2. Different batches of reagents from the same manufacturer were used to detect anti-HCV internal controls, and there were differences in the partial detection values between the mixed and unmixed internal controls (P<0.05). 【Conclusion】 Although the internal controls were mixed by vortex shock before the test, the detection results of some internal controls will still be affected when the STAR automatic sample processing instruments does not set the mixing parameters for internal controls.
RESUMO
【Objective】 To evaluate the necessity and rationality of setting 0.8 CO of the gray area of hepatitis B surface antigen(ELISA) reagent in our laboratory. 【Methods】 1) 792 samples of serum plates from the Clinical Laboratory Centre, Ministry of Health (NCCL) were tested by two HBsAg ELISA reagents. The true positive rate, and confirmed positive rate of gray area samples revealed by 2 reagents were calculated. ROC curve was drawn to determine the best CO value of 2 reagents. The changes in sensitivity and specificity under different CO values were compared. 2) Based on previous data, the HBV-DNA-yield rate among HBsAg gray area samples was analyzed, and the relationship between the distribution of ELISA results of solo HBV-DNA positive samples and gray area was analyzed. 【Results】 Of the 792 samples that form NCCL, 587 were positive, 197 negative and 8 indeterminate. The true positive detection rates of reagents A and B were 82.45% and 71.89%. The confirmed positive rate of gray area samples given by 2 reagents were 94.74%(18/19)and 93.10%(27/29). The best CO values of reagents A and B are 0.49 and 0.27, which are both lower than the 0.8. The specificity corresponding to the best CO values of the two reagents decreased slightly, but the sensitivity increased greatly. From January 2015 to January 2019, 183 551 samples were tested. Of the 13 cases of HBsAg gray area samples, 7 were revealed by reagent A and 1 was positive for NAT; 6 were revealed byreagent B and all negative for NAT. Out of 134 cases of solo HBV-DNA positive samples, 96.27% (129/134) samples had S/CO values below 0.4, overlapped with negative samples, and were far from 0.8. 【Conclusion】 It is necessary to set gray area for these two HBsAg ELISA reagents. The gray area value setting to 0.8 CO corresponds to poor reagent sensitivity. The best cutoff value would be selected according to the ROC curve: 0.49 CO in Reagent A and 0.27 CO in Reagent B. Gray area has no obvious effect on screening of single-virus NAT-yield sample.