UBI'S UBI HCV EIA IS APPROVABLE AS BLOOD SCREEN ASSAY, PANEL RECOMMENDS; DISCREPANCY BETWEEN UBI AND ABBOTT, ORTHO/CHIRON TESTS NOT A "SAFETY ISSUE"

United Biomedical, Inc.'s UBI HCV EIA is approvable for hepatitis C virus antibody blood screening, FDA's blood products advisory committee recommended at a March 25 meeting in Bethesda, Maryland. The panel voted eight to zero, with one abstention, that UBI's product license application for the enzyme immunoassay provided sufficient data "to validate the UBI assay for screening donor blood" for HCV. The second-generation synthetic peptide-based UBI assay detects antibodies to three HCV antigens: one from the HCV core region, one from the "NS3/NS4 interface," and one from the NS5 non-structural HCV region. UBI Director of Diagnostic Research and Development Barbara Hosein, PhD, noted that "the NS5 antigen is unique to the UBI test as no NS5 antigen is present in the current licensed tests." The company chose to use the NS5 antigen because of its "contribution to detection of early seroconversion, but more importantly for its strong correlation with the chronically infected [HCV] state," Hosein explained. UBI presented data comparing its assay to the two currently licensed second-generation HCV assays: Ortho/Chiron's Ortho HCV 2.0 recombinant antigen ELISA test, which was licensed in March ("The Gray Sheet" March 16, p. 3), and Abbott Laboratories' Abbott HCV 2.0 recombinant antigen EIA test, which was licensed in May ("The Gray Sheet" May 11, p. 8). The head-to-head trials were conducted at the request of FDA. In a comparison with Abbott's second-generation assay in 1,054 plasma donors, the two tests "were concordant on more than 99% of the samples," Hosein reported. Similar correlation was seen in a comparison with the Ortho 2.0 assay involving 3,000 blood donors. In addition, the UBI test demonstrated a specificity rate of 99.83% in a study of 13,146 blood donors, Hosein said. Discussing discrepancies that appeared in the comparison with the Abbott test, Hosein noted that four samples identified by UBI's test as negative were indicated as positive by the Abbott test. However, none of these samples were shown to be positive using polymerase chain reaction confirmatory testing. Hosein also noted that two of five samples "that were UBI positive and Abbott negative were PCR positive." Regarding differences with the Ortho/Chiron assay, Hosein said that UBI's assay "detected five samples that were PCR positive, but were negative on the Ortho test." Three of the PCR-positive samples "were detected by the UBI test solely because of the NS5 antibody," Hosein asserted. The Ortho test detected one positive sample missed by the UBI product. Hosein concluded that "head-to-head comparison of the UBI HCV EIA with second-generation recombinant tests shows that introduction of the UBI test will provide an additional increment of [increased] sensitivity of .15% by detecting antibody in blood donors who are infectious for HCV but are not detected by currently licensed tests." She noted that introduction of the second-generation Abbott and Ortho HCV tests "provided an increment of sensitivity...of .1%" over first-generation assays. Following UBI's discussion of trial results, Ortho and Abbott made presentations to the committee. The firms highlighted the need to identify NS3 antibodies in HCV testing and downplayed the importance of detecting NS5 antibodies. Unlike the UBI test, both companies' assays include antigens for the NS3 region. Mitchell Nelles, PhD, manager of Ortho Diagnostics' hepatitis research and development section, argued that "assays lacking an HCV antigen derived from NS3 are unable to detect cases in which NS3 antibody is the only marker of HCV." Nelles cited study results in which assays employing NS3 antigens detected 25 of 35 seroconversion cases (71%). In nine of these cases (26%), NS3 was the only indication of HCV infection, Nelles said. In addition, a study of 231 acute non-A, non-B hepatitis samples revealed that 20 samples "showed reactivity only to NS3 recombinant antigen," Nelles stated. Luann Penday, PhD, hepatitis technical product development manager at Abbott, noted that "in a population of over 30,000 volunteer blood donors" evaluated by Abbott, "we have not found one NS5-antibody-only PCR-confirmed positive case" of HCV. "The issues of the necessity for detection of antibodies to NS5 and the scientific basis for detectability of NS5 remain open," Penday claimed. Discussing data in the UBI application, Leonard Wilson, a staffer at FDA's Center for Biologics Evaluation and Research, commented that the UBI test "appears to detect some possible true positives that the licensed tests fail to detect." However, he pointed out that "on the other hand, UBI's assay failed to detect other possible true positives that the licensed tests did detect." FDA asked the panel to make recommendations on whether "currently licensed tests and the UBI test" are "statistically equivalent" and whether the "observed discrepancies" in test results are "a safety issue." A majority of panel members apparently did not share industry concerns with the differing test results; the panel voted four to two that the discrepancies were not a safety issue. Three panel members abstained from the vote. The committee, feeling that the second question was adequately answered by their response to the safety question, did not address the issue of statistical equivalence.