Objective To evaluate the performance of HCV RNA detection in the first EQA program in 2012 and analyze possible problems in clinical laboratories. Methods The panel consisting of 5 samples was distributed to 927 laboratories.Each panel contains one negative sample and 4 positive samples,which were virus-like particles calibrated by international standard.The percent agreements of all the laboratories for qualitative and quantitative results were calculated.Genomic means (GM) and standard deviations (s) of all laboratories and each reagent were calculated.The overall GM and the GM of each reagent were compared with expected results and correlation curves were calculated. Results The percent agreements of sample 1211, 1212, 1213, 1214 for qualitative results were 99.5% (403/405), 98.5% (400/406), 100.0% (405/405), 100.0% (406/406), respectively. The percent agreement of the negative sample was 99% (401/405). The percent agreements of sample 1211, 1212 and 1213 for quantitive results were similar,which were 93.8% (549/585),92.3% (541/586) and 94.5% (554/586). However,the agreement of sample 1214 was only 87.7% (514/586) and the agreement of sample 1214 for reagent A was 67.2% (92/137). The overall GM agreed with expected results, while GMs of reagent C, E and G deviated from expected results.GMs of sample 1211,1212,1213 and 1214 reported by labs using reagent C were 4.22, 3.56, 5.16 and 5.90, respectively.GMs of sample 1211,1212,1213 and 1214 reported by labs using reagent E were 4.52, 3.78, 5.55 and 6.29, respectively. GMs of sample 1211, 1212, 1213 and 1214 reported by labs using reagent G were 4.83, 4.36, 5.72 and 6.56, respectively. Conclusions The overall results of HCV RNA qualitative and quantitative detection are satisfactory.However,some problems still exist,such as deviation of GM of some reagents,the interassay variability,systematic deviation and accidental deviation,which show that the quality of reagents should be improved.(Chin J Lab Med,2013,36:271-276) Key words: Hepacivirus; RNA; viral; Qualitycontrol
Read full abstract