INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In the 1970s and 1980s, posttransfusional non-A, non-B hepatitis was the most frequent infection transmitted by blood and blood products, representing 80 to 90% of all cases of posttransfusional hepatitis (23).

In the first half of the 1990s, diagnosis of hepatitis C virus (HCV) infection was mainly based on detecting antibodies against recombinant and/or synthetic viral antigens, and with the introduction of tests to detect these antibodies in the screening of blood units, there was a sharp drop in posttransfusional hepatitis (5, 9, 12, 24).

Antibody tests are unable to identify subjects in the early stage of infection, in what is known as the diagnostic window period, during which specific antibodies have not yet been produced, but the virus is present in the plasma, sometimes in large quantities. This stage prior to seroconversion may last up to 2 months in immunocompetent subjects and as long as 6 to 12 months in immunodeficient patients (22). The risk of a blood donation occurring during the window period for HCV has been estimated at 1/103,000 with a 95% confidence interval of 1/28,000 to 1/280,000 (21).

With the introduction of nucleic acid amplification technology (NAT)-based tests based on nucleic acid amplification (PCR and transcription-mediated amplification) or signal amplification (branched DNA [bDNA]), it is possible to identify viremic samples in which antibodies are not yet present and therefore reduce the window period to 15 to 20 days (3, 15). There are, however, problems concerning their possible use in the screening of blood units. Any viral RNA present in the sample is fairly unstable, which means that the sera must be tested immediately or frozen as quickly as possible and stored at 20 or 70°C. NAT-based tests require considerable skill on the part of the operators, although the recent availability of semiautomatic PCR methods could limit problems connected with this aspect (18). The high sensitivity of these methods and the high viral load of samples in the early stage of infection may enable these techniques to be applied to pools of sera (13, 14). The European Medicinal Evaluation Agency (EMEA) planned to introduce testing of plasma pools used in the manufacture of blood products for HCV RNA from July 1999 (10). These recommendations were followed by numerous European countries, which are starting to assess whether to introduce these techniques into the screening of noninactivated products. Pool analysis methods are complex with regard to organization and require high quality standards (19). Furthermore, the risk of false positives (for nucleic acid amplification methods) should be considered, and some studies reported a significant percentage of false-positive results connected with environmental contamination and carryover (2, 7).

On the other hand, bDNA-based methods, although of simple execution, were designed mainly as quantitative tests and are characterized by low sensitivity and long incubation times (20).

Furthermore, both methods involve expensive reagents and long execution times, aspects that have a significant effect on the final cost of each test, which is around $50 (U.S.)/test when the test is performed in diagnostic laboratories. The cost of NAT-based screening of blood donations decreases when tests are performed in pools. Therefore, a quick, inexpensive, sensitive, and specific test is clearly needed to identify potentially infective blood units that have not been identified by specific antibody tests.

A new test has recently been developed to detect the HCV core protein (HCV antigen [Ag]), which is coded for by one of the most conserved regions of the virus genome and which in anti-HCV-positive patients appears to be correlated with HCV-RNA levels (4, 11, 16). This protein may be an ideal target for the development of methods to detect an HCV Ag and more importantly to identify samples from individuals in the early stage of infection (17).

A preliminary study with a small study population was performed, and more definitive and complete data are needed (6).

The aim of this study was to assess the performance of this test for detecting HCV Ag in terms of its specificity and sensitivity and its suitability for use in the screening of blood units to identify infecting samples that do not contain specific antibodies.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Serological assay. The newly developed Ortho Antibody to HCV Core Antigen Elisa Test System (Ortho Diagnostics, Raritan, N.J.) is a sandwich immunoenzymatic test designed to detect the HCV core protein. Monoclonal antibodies reactive with HCV core Ag were used to coat each well in the assay microplate. Other monoclonal antibodies capable of recognizing the N-terminal region of the core protein were conjugated with horseradish peroxidase.

Molecular biological assay. Qualitative and quantitative detection of HCV RNA was performed with the Cobas Amplicor Hepatitis C Virus Test, version 2.0, and the Cobas Amplicor HCV Monitor Test, version 2.0 (Roche Diagnostics, Branchburg, N.J.), respectively. The Cobas Amplicor Hepatitis C Virus Test, version 2.0, is able to detect HCV RNA at a concentration of 50 IU/ml, with a positive rate of 95% or greater, whereas the Cobas Amplicor HCV Monitor Test, version 2.0, detects viral genomes at a concentration of 600 IU/ml, with a positive rate of 95% or greater.

Design of the evaluation and procedure. (i) Specificity. In order to evaluate the specificity of the assay, we tested 2,586 fresh or frozen sera taken from blood donors, 500 samples from the general population, and 58 difficult sera. The last group was taken from patients with dysproteinemia, systemic lupus erythematosus, or autoimmune hepatitis. It also included samples positive for rheumatoid factor, antinucleus antibodies (anti-ANA), anti-hepatitis A virus (HAV) immunoglobulin M (IgM), HbsAg, and anti-HBc IgM. All of the samples were screened for anti-HCV by two immunoenzymatic assays, and none was found to be reactive. They were all tested with the HCV Ag test. The initially reactive samples were retested with the HCV Ag test and screened for HCV RNA by qualitative PCR.

(ii) Sensitivity. In order to assess sensitivity, we applied the HCV Ag test to 103 qualitative PCR-positive and anti-HCV-negative sera, 6 natural seroconversion panels, and 9 commercial seroconversion panels (30994B by Serologicals; 6222, 6225, 6227, and 9041 by Bioclinical Partners, Inc.; and PHV 901, PHV 905, PHV 908, and PHV 917 by Boston Biomedica, Inc.).

(iii) Intra-assay precision. In order to assess intra-assay precision, we selected an HCV RNA-positive, anti-HCV-negative serum (serum A) with a cutoff index (expressed as the ratio of OD to cutoff value [OD/CO]) in a critical area, namely 1.5 to 2. The sample was tested 10 times in a single run. The results, expressed as the cutoff index, were assessed in terms of central trend index (mean) and absolute and relative dispersion (range and percentage of the coefficient of variation [CV]).

(iv) Interassay precision. In order to assess interassay precision, we selected two HCV RNA-positive, anti-HCV-negative sera (sera A and B) with different concentrations of HCV Ag. The first one, just used for the intra-assay precision evaluation, was tested in duplicate for five runs. The second one, with more marked reactivity (approximate cutoff index of 3), was tested in duplicate for six runs. The results, expressed as cutoff index, were assessed in terms of central trend index (mean of individual runs) and relative and absolute dispersion (range and percentage of CV).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Specificity. Tables 1 and 2 show that 17 (0.66%) of the 2,586 sera collected from blood donors presented initial reactivity, 4 with a borderline cutoff index. The distribution of cutoff indices showed that 2520 (97.44%) of the 2,569 nonreactive samples had a value of less than 0.5, and 42 (1.62%) fell within the range 0.5 to 0.8. Seven (0.27%) were within the grey zone. In the repeat test carried out with the 17 initially reactive samples, only 2 (0.08%) samples were reactive, and another had a cutoff index in the grey zone. Qualitative HCV RT-PCR in the 17 initially reactive sera revealed no viremic samples.

Sensitivity. Of the 103 HCV-RNA-positive and anti-HCV-negative sera, 97 (94.2%) were positive with the HCV Ag test, all having a cutoff index greater than 1.2. Two of the six nonreactive samples in the HCV Ag test had a cutoff index in the grey zone (Table 3). The repeat HCV Ag test of the six initially nonreactive sera confirmed this result.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Regression plot between HCV Ag OD/CO and Log HCV-RNA level in 45 HCV RNA-positive anti-HCV-negative sera.

Intra-assay and interassay precision. Data on intra-assay and interassay precision are shown in Table 5. The dispersion indices observed in both assays had values of less than 10%.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ELISA for detecting HCV core Ag assessed in our study was the first to be standardized and ready to be marketed. This assay meets the need for a method capable of detecting a viral replication marker before routinely used tests can detect specific HCV antibodies.

With regard to laboratory performance, there is no recognized "gold standard" for determining the presence or absence of infection in the window period. Therefore, it was not possible to determine the absolute specificity and sensitivity of the test under study. Considering the possible applications of the HCV Ag test, we took as a reference the most standardized, most commonly used NAT-based testPCR.

With regard to specificity, the algorithm adopted entailed performance of the PCR assay with samples that were reactive in the HCV Ag test. Only 21 (0.67%) of the 3,144 sera tested overall were initially HCV Ag reactive, and being HCV RNA negative, they were considered as false positives. However, considering the sera were repeatedly reactive and 3,141 out of 3,144 sera (99.9%) were identified as HCV Ag negative, they could be false-negative PCRs (i.e., below the detection limits of the PCR test used in this study). Therefore, the HCV Ag test had a specificity comparable to that of ELISA kits routinely used in screening for blood-transmitted infections such as hepatitis B and human immunodeficiency virus infection (8, 25). Inter- and intra-assay reproducibility data also showed identical performances to ELISA tests currently on the market. Furthermore, the test for detecting HCV Ag has been shown to have excellent discriminatory ability. Indeed, virtually all (97.78%) of the 3,144 samples tested for specificity gave absorbance values of less than 50% of the cutoff value. Only 7 of the 3,144 samples (0.22%) had a cutoff index in the range 0.8 to 1, the grey zone.

With regard to sensitivity, this study showed that the HCV Ag test's ability to detect the virus is comparable to that of PCR. The most significant finding is undoubtedly the reduction in window period as calculated with the seroconversion panels used. The first positive sample in the HCV Ag test and PCR was on average 33 days (33 versus 33.4 days) prior to antibody detectability. Furthermore, when we considered the 170 anti-HCV-negative HCV RNA-positive samples analyzed (103 individual samplings and 67 samples from seroconversion panels), 162 (95.29%) were HCV Ag positive and 3 (1.76%) were in the grey zone.

The finding of some HCV RNA-positive and HCV Ag-negative samples was not entirely unexpected considering the high sensitivity of NAT-based tests, which have taken about 10 years to reach the present degree of standardization. Also, an analysis of the regression curve for the HCV Ag test cutoff index and viral load expressed in international units per milliliter shows that the curve meets the cutoff value at an HCV RNA concentration of about 4 log IU/ml. This shows that the analytical sensitivity of the test, which was not the subject of the analysis, was less marked than that of NAT-based tests. However, viral kinetics in the early stage of HCV infection, characterized by rapid growth in viral replication indices in only a few days, as shown in the natural and commercial panels, caused PCR and HCV Ag positivity to be virtually simultaneous.

In conclusion, the Ortho Antibody to HCV Core Antigen ELISA Test System has been shown to have high sensitivity and specificity, allowing it to be used in the screening of blood donations.

Our experience has also shown that the test is easy to use, which, together with the low cost of the kit, means it can be used to analyze individual blood donations and solve organizational problems connected with pool analysis (1, 13).

It is therefore a valid alternative to NAT-based tests for identification of HCV infection in the diagnostic window period, specially in situations in which organizational and technological problems connected with pool analysis are difficult to solve (e.g., in countries not able to set up nucleic acid tests). The HCV Ag test could become a useful means for improving the safety of blood and blood products and for detecting an early HCV infection in high risk individuals and late seroconverters.