Is Anti-Hepatitis C Virus Antibody Level an Appropriate Marker to Preclude the Need for Supplemental Testing?

Original Paper Received: September 1, 2015 Accepted: October 4, 2015 Published online: January 20, 2016 Is Anti-Hepatitis C Virus Antibody Level an Appropriate Marker to Preclude the Need for Supplemental Testing? Kuo Zhang Lunan Wang Guigao Lin Jinming Li National Center for Clinical Laboratories, Beijing Hospital, Beijing, PR China Key Words Hepatitis C Hepatitis C antibody Chemiluminescence immunoassay Signal-to-cutoff ratio Strategy Abstract Objectives: In the present study, we aimed to determine whether signal-to-cutoff (S/Co) ratios of reactive anti-hcv samples could be used as a basis for avoiding the need for supplemental testing in our study population. Methods: We analyzed 901 anti-hcv-positive sera from 8 institutions in China. The Ortho VITROS anti-hcv assay and Monolisa Plus anti-hcv version 2 were used as screening assays to detect anti-hcv antibodies. Recombinant immunoblot assay (RIBA) and quantitative tests for HCV RNA were performed to validate confirmed HCV infection status. Results: Receiver operating characteristic curve analyses demonstrated that 41.5% (114/275) of true-positive samples with S/Co ratios 3.0 would be missed and the negative predictive value was 63.9 and 87.06%, using real-time polymerase chain reaction (RT- PCR) and RIBA as supplemental testing, respectively. 29.8% (90/302) of those who tested positive by RIBA samples were missed when only RT-PCR was used as supplemental testing. Conclusions: We determined that very low anti-hcv levels (S/Co 3.0), as determined by chemiluminescence immunoassay, was not an appropriate marker to preclude the need for supplemental testing in our study population. A screening strategy employing a secondary HCV antibody assay using different HCV antigens from the first assay as the supplemental testing method should be studied further. The immunoblot assay, as a supplemental testing method, is still necessary. 2016 S. Karger AG, Basel Introduction Hepatitis C virus (HCV) is the causative agent of the contagious liver disease hepatitis C. The virus can cause both acute and chronic hepatitis infection, thus presenting an enormous health burden. A significant number of people with chronic hepatitis C infection are at high risk for consequently developing liver cirrhosis and hepatocellular carcinoma, which causes serious mortality and morbidity [1 3]. It is therefore essential to identify individuals infected with HCV. Enzyme immunoassays and chemiluminescence immunoassays (CIAs) are the two main screening immunoassays for detection of anti-hcv antibodies. Although the CIA screening method demonstrates improved specificity compared to enzyme immunoassays [4], these anti- HCV screening assays may generate false-positive results. E-Mail karger@karger.com www.karger.com/int 2016 S. Karger AG, Basel 0300 5526/16/0585 0310$39.50/0 Jinming Li National Center for Clinical Laboratories Beijing Hospital No. 1 Dahua Road, Dongdan, Beijing 100730 (PR China) E-Mail jmli @ nccl.org.cn

Therefore, it is essential to validate the specificity of these screening assays using a supplemental assay. Recombinant immunoblot assay (RIBA) or HCV RNA polymerase chain reaction (PCR) has been recommended for confirmation of positive anti-hcv screening tests [5]. In 2013, due to discontinuation of RIBA, the Centers for Disease Control and Prevention (CDC) [6] recommended that a positive result from an initial anti-hcv screening test be followed only by nucleic acid testing (NAT) for detection of HCV RNA. A second round of anti-hcv screening offers an alternative supplemental testing method for confirmation, according to the algorithm published by Vermeersch et al. [7]. Guidelines published by the CDC incorporated the level of the signal-to-cutoff (S/Co) ratio relative to the anti-hcv concentration into laboratory algorithms for anti-hcv testing in 2003. Samples with S/Co ratios 8.0, as determined by the VITROS anti-hcv assay (Ortho Clinical Diagnostics, Raritan, N.J., USA), are considered positive for anti-hcv antibodies [8]. Using receiver operating characteristic (ROC) curve analysis to predict appropriate S/Co ratios for anti-hcv testing, previous studies have shown that very low anti-hcv S/Co ratios of <3.0 or 4.5 are associated with a high diagnostic sensitivity and negative predictive value (NPV), indicating no risk of HCV infection and therefore requiring no additional tests. However, high anti-hcv S/Co ratios of 20.0, as determined by the Ortho VITROS anti-hcv assay, were an accurate serological marker of viremia [9 11]. In these studies, samples with high anti-hcv levels were further evaluated by HCV RNA testing to assess viremic status. Such strategies, using anti-hcv S/Co ratios as a measure of anti-hcv concentration minimizes the number of individuals that require supplemental testing to some extent. Therefore, the objective of this study was to determine whether S/Co ratios of reactive samples could be used as a basis of avoiding the need for supplemental testing, and whether secondary anti-hcv testing could offer an alternative to the supplement or confirmation of HCV detection. We also evaluated whether RIBA could address false-positive anti-hcv results obtained from initial screening assays if the HCV RNA results are negative. Materials and Methods Ethics Statement The study involved the use of leftover patient samples. The Ethics Committee of the National Center for Clinical Laboratories approved our use of these patient samples, and we adhered to the tenets of the Declaration of Helsinki. Since this study did not require the collection of detailed patient information, and the data were analyzed anonymously, participants did not provide their written informed consent. Samples Patient serum samples used in this study were collected from the General Hospital of Ningxia Medical University, Peking University People s Hospital, Fuzhou General Hospital of the Nanjing Military Area Command, Shanghai Ruijin Hospital, Shangdong Province Hospital, General Hospital of the Nanjing Military Region, East Hospital of the Affiliated Hospital of Qingdao University Medical College, and Fujian Blood Center in China. All serum samples were assessed for the presence of antibodies to HCV using VITROS ECi CIA (Ortho Clinical Diagnostics). S/Co ratios were recorded directly from the automated equipment. Samples with S/Co ratios of 1.0 were defined as reactive based on the manufacturer s recommendation. Serum samples that showed reactive anti-hcv results were shipped on dry ice to the National Center for Clinical Laboratories in Beijing for further testing. Screening Assays Serum samples shipped to the National Center for Clinical Laboratories were first retested using the VITROS ECi CIA. Sera that were negative with VITROS ECi CIA were considered anti-hcv negative, according to CDC guidelines [6]. Sera were subsequently tested with Monolisa Plus anti-hcv version 2 (Monolisa Plus; Bio-Rad, Marnes-la-Coquette, France) and samples with S/Co ratios 1.0 were considered reactive according to the manufacturer s instructions. These two screening assays use antigens from a different manufacturer. Confirmation Assays Quantitative HCV NAT was performed on all serum samples with positive ECi CIA results (S/Co ratios 1) using the Roche COBAS AmpliPrep/COBASTaqMan HCV Test (Roche Diagnostics, Branchburg, N.J., USA) as a confirmation assay. The sensitivity (lowest limit of detection) of the quantitative HCV NAT was 15 IU HCV RNA/ml. Testing and result interpretation was performed according to the manufacturer s instructions. However, due to insufficient sample volume, HCV NAT was not performed on 13 samples. A third-generation RIBA (RIBA HCV 3.0; Ortho Clinical Diagnostics) was used to detect the HCV recombinant proteins C100 (NS4), C33c (NS3), C22p (core), and NS5 in order to define the sample status. The results were interpreted according to the manufacturer s recommendations. Samples were deemed positive when 2 bands showed reactivity, indeterminate when only 1 band was reactive, and negative when no reactivity was observed. Definition and Statistical Analysis Individuals with HCV-positive RNA were considered viremic. Samples with a positive RIBA result and negative HCV RNA were recorded as true antibody positive, nonviremic. Samples with reactive anti-hcv screening test results but negative HCV RNA results and negative or indeterminate RIBA results were categorized as falsely positive [8]. ROC curves were constructed by plotting sensitivity versus 1 specificity, using HCV RNA and the third-generation RIBA Anti-HCV Level as a Supplemental Testing Marker 311

Table 1. Categories of hepatitis C antibody levels based on S/Co ratios and the results of supplemental testing Anti-HCV S/CO ratio by CIA Number of patients (n = 888) Viremic subjects (n = 586) Nonviremic subjects (n = 302) Subjects in whom viremia was not (n = 13) positive negative indeterminate positive negative indeterminate posi tive negative indeterminate RIBA RIBA RIBA RIBA RIBA RIBA RIBA RIBA RIBA 1.01 3.00 275 30 29 40 13 72 91 2 0 0 3.01 20.00 204 83 7 23 44 14 33 1 3 5 20 409 370 0 4 33 1 1 1 0 1 Total number 888 483 36 67 90 87 125 4 3 6 test as gold standards, respectively. We determined the diagnostic sensitivity, diagnostic specificity, positive predictive value (PPV), NPV, and their respective exact 95% CI to predict HCV viremia and RIBA status at S/Co ratios of 3.0, 8.0, and 20.0, according to previously published methods [8 11]. Optimal S/Co ratios were identified from the analysis of ROC curves and associated data [12]. We performed ROC analysis using Graphpad Prism 6 statistical software. 100 S/Co = 8.00 80 S/Co = 20.00 S/Co = 3.00 Results A total of 1,017 samples were shipped to the National Center for Clinical Laboratories and retested for anti- HCV antibodies using CIA as the screening assay, and 901 samples demonstrated reactive results (S/Co 1). All reactive samples (S/Co 1) were also tested using quantitative HCV NAT testing and third-generation RIBA. HCV RNA test results that were <15 IU/ml (below the quantifiable linear range) were deemed as HCV RNA reactive since the RIBA results illustrated that 35.8% (63/176) of the samples demonstrated detectable HCV RNA values of <15 IU/ml and were RIBA positive. Of these 901 samples, HCV RNA testing was not performed on 13 samples due to insufficient sample quantity; however, 586 samples (65.0%) and 302 samples (33.5%) demonstrated confirmatory HCV RNA-positive and -negative results, respectively. Furthermore, of the 901 samples, 577 (64.0%), 126 (14.0%), and 198 (22.0%) demonstrated positive, negative, and indeterminate results, respectively, after additional RIBA testing. As shown in table 1, of the 586 samples with positive HCV RNA results, 483 (82.4%) tested positive by RIBA, and of the 302 samples with negative HCV RNA results, 90 (29.8%) tested positive by RIBA. These 90 samples represented true-positive anti-hcv results without viral replication. Sensitivity (%) 60 40 20 Sensitivity (%) Identity (%) 0 0 20 40 60 80 100 100% specificity (%) Fig. 1. RT-PCR test ROC curve based on different CIA cutoff levels for anti-hcv antibody detection. The area under the curve is 0.812 (95% CI: 0.783 0.841). PCR test ROC curves were analyzed to determine cutoff levels based on CIA results for anti-hcv antibody detection. Based on the PCR test ROC curve ( fig. 1 ) and associated diagnostic sensitivity, diagnostic specificity, and PPV, we determined that an S/Co ratio of 20.0 was not optimal, as demonstrated in previous studies [9, 10]. The corresponding diagnostic sensitivity, diagnostic specificity, PPV, and NPV for HCV RNA were 63.82, 88.89, 91.67, and 56.20%, respectively ( table 2 ). We also 312 Zhang/Wang/Lin/Li

Table 2. Diagnostic performance of CIA in the prediction of viremia by RT-PCR S/Co ratio 3.0 8.0 20.0 Diagnostic sensitivity, % 82.94 (79.64 85.89) 75.77 (72.09 79.19) 63.65 (59.61 67.55) Diagnostic specificity, % 57.84 (52.09 63.44) 78.43 (73.39 82.91) 88.89 (84.82 92.18) PPV, % 79.02 (75.81 82.24) 87.06 (84.15 89.97) 91.67 (88.99 94.35) NPV, % 63.90 (58.25 69.56) 62.83 (57.98 67.68) 56.20 (51.78 60.62) Values in parentheses are the limits of the 95% CI. Table 3. Diagnostic performance of CIA in the prediction of the presence of anti-hcv antibodies by RIBA S/Co ratio 3.0 8.0 20.0 Diagnostic sensitivity, % 92.03 (89.51 94.10) 84.23 (80.99 87.11) 69.84 (65.92 73.57) Diagnostic specificity, % 71.91 (66.68 76.74) 90.74 (87.05 93.67) 97.84 (95.60 99.13) PPV, % 81.46 (84.42 88.54) 92.42 (90.18 94.67) 97.75 (96.32 99.17) NPV, % 87.06 (83.60 90.52) 81.10 (77.63 84.57) 70.75 (67.13 74.38) Values in parentheses are the limits of the 95% CI. determined the diagnostic sensitivity, diagnostic specificity, PPV, and NPV and their respective 95% CI for prediction of HCV viremia at S/Co ratios of 3.0 and 8.0 ( table 2 ). We analyzed the RIBA test ROC curve for different cutoff levels based on CIA results for the diagnosis of HCV exposure. The RIBA ROC curve ( fig. 2 ) and associated data demonstrated that an S/Co ratio of 20.0 corresponded to a diagnostic sensitivity, diagnostic specificity, PPV, and NPV for HCV antibody confirmation by RIBA of 69.84, 97.84, 97.75, and 70.75%, respectively ( table 3 ). Tables 2 and 3 illustrate that an S/Co ratio of 20 better discriminates viremia and HCV exposure in screened anti-hcvpositive samples compared to other S/Co ratios such as 8.0. The diagnostic sensitivity, diagnostic specificity, PPV, and NPV, as well as their respective 95% CI in predicting HCV exposure for different cutoff levels at S/Co ratios of 3.0 and 8.0 are shown in table 3. From both the real-time PCR (RT-PCR) ( fig. 1 ) and RIBA ( fig. 2 ) ROC curves, we identified that an S/Co ratio of 3.0 was not the highest value providing a diagnostic sensitivity of 100%. Table 1 shows that 99 out of 275 subjects (36%) with S/Co ratios of <3.0 were viremic subjects, and 15 displayed HCV exposure without viremia. These results illustrate that patients with S/Co ratios <3.0 were not all negative for both HCV viremia and HCV exposure. Sensitivity (%) 100 S/Co = 3.00 S/Co = 8.00 80 60 40 20 0 S/Co = 20.00 Sensitivity (%) Identity (%) 0 20 40 60 80 100 100% specificity (%) Fig. 2. RIBA test ROC curve based on different CIA cutoff levels for anti-hcv antibody detection. The area under the curve is 0.932 (95% CI: 0.916 0.948). Anti-HCV Level as a Supplemental Testing Marker 313

% RIBA IND 100 90 80 70 60 131 RIBA N 61 24 RIBA P 1 6 Color version available online % 100 90 80 70 60 176 Nonviremic 91 Viremic 35 Color version available online 50 40 30 20 10 0 101 45 n = 277 (1.01 3.00) 128 n = 213 (3.01 20.00) 404 n = 411 ( 20) 50 40 30 20 10 0 99 n = 275 (1.01 3.00) 113 n = 204 (3.01 20.00) 374 n = 409 ( 20) Fig. 3. RIBA results relative to S/Co ratios according to antibody level. Samples with an S/Co ratio 20, between 3.0 and 19.99, or between 1.0 and 2.99 were classified as high antibody levels, low antibody levels, or very low antibody levels, respectively. These antibody levels are outlined in parentheses below each bar. The percentages of RIBA-positive samples according to these antibody levels were as follows: high antibody levels, 98.3% (404/411); low antibody levels, 6.0% (128/213), and very low antibody levels, 16.2% (45/277). IND = Indeterminate; N = negative; P = positive. Fig. 4. Distribution of HCV viremic samples according to antibody level. Samples with an S/Co ratio 20, between 3.0 and 19.99, or between 1.0 and 2.99 were classified as high antibody levels, low antibody levels, or very low antibody levels, respectively. The percentages of HCV viremic samples according to these antibody levels were as follows: high antibody levels, 91.4% (374/409); low antibody levels, 55.4% (113/204), and very low antibody levels, 36.0% (99/275). These antibody levels are outlined in parentheses below each bar. Because of insufficient sample volume, 2 samples with high antibody levels, 9 samples with low antibody levels, and 2 samples with very low antibody levels were not tested for HCV RNA. Although an S/Co ratio of 20.0 was not determined to be an optimal cutoff in previous studies [9, 10], results with an S/Co ratio 20.0 were still classified as possessing high antibody levels in the present study. Those samples with an S/Co ratio between 3.0 and 19.99 were designated as low level. Samples with very low anti-hcv levels and S/Co ratios between 1.0 and 2.99 were selected based on a previous study showing that these samples demonstrated no risk of having HCV infection [9]. These three antibody levels were observed in 411 (45.62%), 213 (23.64%), and 277 (30.74%) of the 901 samples, respectively ( fig. 3 ). Except for 13 samples that were not tested for HCV RNA due to insufficient sample volume, HCV viremia was confirmed by positive HCV RNA testing in 91.4% (374/409) of samples with high antibody levels, 55.4% (113/204) of samples with low antibody levels, and 36.0% (99/275) of samples with very low antibody levels ( fig. 4 ). A significant difference was observed in the viral replication frequency between samples with high anti-hcv antibody levels (S/Co ratios 20.0; 91.4%) and those in the low and very low antibody level groups (S/Co ratios between 1.0 and 19.99; 44.3%; p < 0.001, χ 2 test). In our study, 586 viremic individuals demonstrated higher antibody levels (mean S/Co ratio: 19.23, 95% CI: 18.4 20.1) than individuals (90 samples) with confirmed serological HCV without viremia (mean S/Co ratio: 14.33, 95% CI: 12.2 16.4, p < 0.05). A mean S/Co ratio of 2.94 (95% CI: 2.51 3.37) was observed in 212 samples defined as false positive for hepatitis C without viremia, and with negative or indeterminate RIBA results. The results of Monolisa Plus testing in 888 CIA-positive sera (S/Co 1) that underwent previous HCV RNA testing are shown in table 4. The sensitivity and specificity of the Monolisa Plus test were 81.8 and 85.8%, respectively, compared to confirmation with RT-PCR and RIBA. The PPV and NPV were 94.85 and 59.67%, respectively. In 90 samples with confirmed serological HCV without viremia, 16 samples (17.8%) were missed using Monolisa Plus as a supplemental test. In the 212 individuals defined as false positive for hepatitis C without viremia with negative or indeterminate RIBA, 31 samples (14.6%) were detected as false positive for anti-hcv antibody using the Monolisa Plus test. 314 Zhang/Wang/Lin/Li

Table 4. Results of the Monolisa Plus test in 888 CIA-positive sera (1 S/Co <20) Monolisa Plus True-positive anti-hcv False-positive anti-hcv HCV RNA positive HCV RNA negative/ RIBA positive H CV RNA negative/ RIBA negative or indeterminate Positive (S/Co 1) 94 53 29 Negative (S/Co <1) 106 16 181 The 888 CIA-positive sera were confirmed by HCV RNA test. Table 5. Interpretation for anti-hcv results utilizing S/Co ratios and type of recommended supplemental testing Antibody level (S/Co ratio) Recommended supplemental testing Result Interpretation Recommendation S/Co <1.0 None No HCV antibody detected Notify 1.0 S/Co HCV RNA Positive Current HCV infection Notify and link to care <20.0 Negative Nonviremic hepatitis C or nonhepatitis C Additional testing as appropriate 1 S/Co 20.0 HCV RNA Positive Current HCV infection Notify and link to care Negative Nonviremic hepatitis C Notify 1 Repeat HCV RNA testing or follow-up testing for HCV antibody is recommended if the individual tested might have been exposed to HCV within the past 6 months or has clinical evidence of HCV disease. At this time, an immunoblot assay as supplemental testing is still necessary. Discussion Our study shows that very low anti-hcv antibody levels with S/Co ratios <3.0, as determined by the VITROS anti-hcv assay, did not identify anti-hcv false positives; therefore, this group is not an accurate marker that can prevent the need for supplemental testing. Confirmatory anti-hcv testing by RIBA or secondary HCV antibody assay is not necessary when the S/Co ratio is >20 because of the high rate of true positives detected (PPV: 97.75%). Lai et al. [9] reported that an S/Co ratio of 3.0 determined by the VITROS anti-hcv assay was the highest value associated with a diagnostic sensitivity of 100% and NPV of 100%, using either PCR or RIBA as gold standards. No positive RIBA or PCR test results were found in samples with an S/Co ratio <3.0 in their analyses. Therefore, it was suggested that supplemental testing was not necessary for patient samples with S/Co ratios <3.0. Similarly, Contreras et al. [10] and Oethinger et al. [4] also demonstrated that very low hepatitis C antibody levels were false positives and thus avoided supplemental testing. Samples with very low hepatitis C antibody levels in the aforementioned studies were designated as having S/Co ratios of 4.5 and 5.0, respectively [4, 10]. In the present study, the mean S/Co ratio for false-positive hepatitis C individuals without viremia, and with negative or indeterminate RIBA results (n = 212) was 2.94 (95% CI: 2.51 3.37). However, the CIA versus RT-PCR ( fig. 2 ) and CIA versus RIBA ( fig. 3 ) ROC curves illustrated that an S/Co ratio of 3.0 was not associated with a diagnostic sensitivity or NPV of 100%, using either PCR or RIBA as gold standards. In our population, a cutoff S/Co ratio of 3.0 would prevent detection of 41.5% (114/275) of CIA-positive samples with S/Co 3.0 (table 1 ), with either positive RNA (n = 99) or positive RIBA (n = 15). Therefore, we recommend that supplemental testing should still be required for patients with very low anti-hcv antibody levels and S/Co ratios 3, as determined by CIA. Moreover, although a previous study showed that only 1.8% of subjects with an S/Co ratios <20.0 were viremic [4], we demonstrated that 23.9% (212/888) of samples with the same S/Co ratios were viremic. Therefore, performing supplementary testing using HCV RNA tests on all samples, in- Anti-HCV Level as a Supplemental Testing Marker 315

cluding those with low antibody levels, is not cost-effective ( table 5 ). Interestingly, the ROC curve analysis in our study demonstrated that an S/Co ratio of 20.0 was not an optimal cutoff, as has been suggested in previous studies [9, 10]. The differences in the findings could be attributed to the following three reasons. First, the previous study [9] assumed that all samples with positive results, determined by RT-PCR testing, would also be positive by RIBA. The present study demonstrated that 6.14% (6/586) and 11.4% (67/586) of samples with positive RT-PCR results displayed negative and indeterminate results by RIBA, respectively. Second, the study populations were different. Lai et al. [9] proposed an algorithm for HCV testing based on the results in a population of veterans. Oethinger et al. [4] conducted the study using blood donor samples. However, the population in our study came from three groups: patients from a liver disease clinic (14%, 127/901), patients from other disease clinics (83%, 748/901), and blood donors (3%, 26/901). The difference in the prevalence of anti-hcv antibodies in the various study populations might account for the differences in optimal S/Co ratio cutoffs. Third, the distribution of the predominant HCV subtype varies regionally; for example, HCV-1b and 2a are the most common subtypes in China [13]. This might have caused the varying results between the studies for the VITROS anti-hcv assay. Our values for diagnostic specificity (91.67%) and PPV (88.89%) in predicting HCV viremia using the VITROS anti-hcv assay at an S/Co ratio of 20.0 were higher than the values (58.8 and 81%, respectively) reported by Lai et al. [9], but were lower than those (96.6 and 93.7%) reported by Contreras et al. [10]. The results demonstrated that the proportion of our population with an S/Co ratio of 20.0 but with HCV viremia was between the proportions of populations studied by Lai et al. [9] and Contreras et al. [10]. Diagnostic sensitivity (75.77%) and NPV (62.83%) at an S/Co ratio of 8.0 were much lower than comparable results in the two previous studies, which suggest that we identified a greater proportion of our population with S/Co ratios 8.0 but who also have HCV viremia. The present study showed that 586 viremic individuals had higher antibody levels (mean S/Co ratio: 19.23, 95% CI: 18.4 20.1). We also showed that of the 409 samples tested for HCV RNA with S/Co ratios 20.0, based on CIA, 374 were positive for HCV RNA (91.4%). The RNA positivity rate was different from that reported in previous studies: 81% [9], 90% [14], 93% [10], 81% [4], and >60% [15]. Of the 34 samples with S/Co ratios 20.0 and with negative HCV RNA, 33 samples were RIBA positive and 1 sample was RIBA indeterminate. Of the two samples not tested for HCV RNA due to insufficient volume, one was RIBA positive and the other was RIBA indeterminate. The true-positive rate was at least 99.5% (408/410). The results also showed that 99.8% (409/410) of samples with S/Co ratios 20.0 were reactive, as determined by the Monolisa Plus test. In our study, using ROC curve analysis, the high diagnostic specificity and PPV values for the prediction of either viremia by RT-PCR or presence of anti-hcv antibodies by RIBA showed that an S/Co ratio 20.0 strongly indicates HCV exposure. Therefore, we recommend, at least for our study population, that samples with S/Co ratios 20.0 should not undergo supplemental RIBA testing or secondary immunoassay testing. This would be unnecessary as these samples with such high S/Co ratios are confirmed by positive anti-hcv RIBA results 98%. Upon evaluation for antiviral therapy, these samples should directly proceed to NAT to assess HCV viremic status ( table 5 ). A strategy for HCV antibody testing using two enzyme immunoassays in a routine clinical laboratory has been validated, and the sensitivity and specificity of confirmation of the second enzyme immunoassay were 98.15 and 98.33%, respectively [7]. The CDC recently recommended that testing be done with a second HCV antibody assay that is different from the initial antibody assay used for diagnosis of HCV infection when the result of HCV RNA is negative, as a result of the discontinuation of HCV RIBA [6]. In the present study, samples were also tested by Monolisa Plus. The results showed that 17.8% of samples (16/90) with confirmed serological HCV without viremia were missed when Monolisa Plus was used as the supplemental testing method. Immunodeficiency might be the common cause of false-negative anti-hcv results in chronic HCV-infected patients [16]. Our previous study of blood donors also showed that even with two screening assays in addition to NAT, some anti-hcvpositive samples were still missed, suggesting that there may be no suitable combination to achieve a 100% sensitivity rate and avoid viral transmission [17]. In this study, 14.6% of samples (31/212) were detected as false positive for anti-hcv by Monolisa Plus. The results show that a strategy employing secondary HCV antibody assay testing with different HCV antigens from the first assay as a supplemental screening method was not an excellent strategy for accurate HCV screening in our population. Therefore, although there are many disadvantages, such as high cost, requirement of specialized equipment 316 Zhang/Wang/Lin/Li

and qualified personnel, extended execution time, and indeterminate results, the immunoblot assay as supplemental testing is still necessary, especially for the nonviremic individual with false-negative anti-hcv results. A previous study [18] evaluated the sensitivity of five anti-hcv immunoblot assays licensed in France and found that the results were less divergent across assays with more uniform criteria for interpretation. The RIBA HCV 3.0 assay is no longer available; therefore, other immunoblot assays should be selected for supplemental testing. When the S/ Co ratio based on the VITROS anti-hcv assay is between 1.0 and 20.0, additional testing should be performed as appropriate, i.e. an immunoblot assay for samples without viremic hepatitis ( table 5 ). Our interpretation of anti-hcv results utilizing S/Co ratios and the type of recommended supplemental testing is summarized in table 5. If the S/Co ratio is <20.0, based on CIA, HCV RNA testing should also be performed as a confirmatory test for discrimination of nonviremic hepatitis C or nonhepatitis C. Repeat HCV RNA testing or follow-up testing for HCV antibodies is recommended if the person tested might have been exposed to HCV within the past 6 months or has clinical evidence of HCV disease. For positive samples without viremia, immunoblot assays as supplemental testing is still necessary. If the S/ Co ratio is 20.0, based on CIA, then performing RT- PCR could further assess the presence of HCV viremia. Acknowledgments We gratefully acknowledge all of the listed institutions in Material and Methods for the provision of samples. References 1 World Health Organization. Hepatitis C fact sheet No. 164. Updated July 2013. http:// www.who.int/mediacentre/factsheets/fs164/ en/. 2 Di Bisceglie AM: Hepatitis C. Lancet 1998; 69: 213 216. 3 Lauer GM, Walker BD: Hepatitis virus infection. N Engl J Med 2001; 345: 41 52. 4 Oethinger M, Mayo DR, Falcone J, Barua PK, Griffith BP: Efficiency of the Ortho VITROS assay for detection of hepatitis C virus-specific antibodies increased by elimination of supplemental testing of samples with very low sample-to-cutoff ratios. J Clin Microbiol 2005; 43: 2477 2480. 5 Chapko MK, Sloan KL, Davison JW, Dufour DR, Bankson DD, Rigsby M, et al: Cost effectiveness of testing strategies for chronic hepatitis C. Am J Gastroenterol 2005; 100: 607 615. 6 Centers for Disease Control and Prevention (CDC): Testing for HCV infection: an update of guidance for clinicians and laboratorians. MMWR Morb Mortal Wkly Rep 2013; 62: 362 365. 7 Vermeersch P, Van Ranst M, Lagrou K: Validation of a strategy for HCV antibody testing with two enzyme immunoassays in a routine clinical laboratory. J Clin Virol 2008; 42: 394 398. 8 Alter MJ, Kuhnert WL, Finelli L; Centers for Disease Control and Prevention: Guidelines for laboratory testing and result reporting of antibody to hepatitis C virus. Centers for Disease Control and Prevention. MMWR Recomm Rep 2003; 52: 1 13. 9 Lai KK, Jin M, Yuan S, Larson MF, Dominitz JA, Bankson DD: Improved reflexive testing algorithm for hepatitis C infection using signal-to-cutoff ratios of a hepatitis C virus antibody assay. Clin Chem 2011; 57: 1050 1056. 10 Contreras AM, Ochoa-Jiménez RJ, Celis A, Méndez C, Olivares L, Rebolledo CE, Hernandez-Lugo I, Aguirre-Zavala AI, Jiménez- Méndez R, Chung RT: High antibody level: an accurate serologic marker of viremia in asymptomatic people with hepatitis C infection. Transfusion 2010; 50: 1335 1343. 11 Contreras AM, Tornero-Romo CM, Toribio JG, Celis A, Orozco-Hernández A, Rivera PK, Méndez C, Hernández-Lugo MI, Olivares L, Alvarado MA: Very low hepatitis C antibody levels predict false-positive results and avoid supplemental testing. Transfusion 2008; 48: 2540 2548. 12 Akobeng AK: Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr 2007; 96: 644 647. 13 Zhuang H, Tracy L, Cui Y: Study on hepatitis C virus genotyping in some parts of China (in Chinese). Zhonghua Liu Xing Bing Xue Za Zhi 2001; 22: 99 101. 14 Dufour DR, Talastas M, Fernandez MD, Harris B: Chemiluminescence assay improves specificity of hepatitis C antibody detection. Clin Chem 2003; 49: 940 944. 15 Dufour DR: Lot-to-lot variation in anti-hepatitis C signal-to-cutoff ratio. Clin Chem 2004; 50: 958 960. 16 Fabrizi F, Poordad FF, Martin P: Hepatitis C infection and the patient with end-stage renal disease. Hepatology 2002; 36: 3 10. 17 Zhang K, Wang L, Sun Y, Zhang R, Lin G, Xie J, Li J: Improving the safety of blood transfusion by using a combination of two screening assays for hepatitis C virus. Transfus Med 2014; 24: 297 304. 18 Couroucé AM, Noel L, Barin F, Elghouzzi MH, Lunel F, North ML, Smilovici W: A comparative evaluation of the sensitivity of five anti-hepatitis C virus immunoblot assays. Vox Sang 1998; 74: 217 224. Anti-HCV Level as a Supplemental Testing Marker 317