CLINICAL TRIAL OF THE HEMA-STRIP HIV RAPID TEST USING FINGERSTICK BLOOD, WHOLE BLOOD, PLASMA, AND SERUM Niel T. Constantine 1, Dan Bigg 2, Daniel Cohen 3, Joanne Court 4, Alison Jones 5, Fassil Ketema 1, Thomas Koppes 6, Vincent Marsiglia 7, Harry Prince 8 1 University of Maryland, Baltimore, MD, 2 Chicago Recovery Alliance, Chicago, IL, 3 Fenway Community Health, Boston, MA, 4 MDC Associates, Boston, MA, 5 Tapestry Health Services, Northhampton, MA, 6 Northstar Medical Center Chicago, IL, 7 BDC Laboratory, Baltimore, MD, 8 Focus Technologies, Cypress, CA Insert your information here RESULTS Objectives To evaluate the performance characteristics of a novel rapid HIV test in a large number of persons from several populations To determine the concordance of results when using several different blood matrices. Methods Sites: Seven sites from 4 states (Maryland, Illinois, Massachusetts, and Texas) were selected to enroll study participants for the evaluation of the rapid test. Prior to their participation in the study, participants signed consent forms approved by the Institutional Review Board at the respective study sites. Samples: A total of 1,515 participants were enrolled and tested in this evaluation and included 504 at low risk for HIV, 511 at high risk, and 500 persons known to be infected with HIV. Fingerstick whole blood, venous whole blood, plasma, and serum sample sets were collected from each participant and tested by the Hema-Strip HIV test, and results were compared to FDA approved reference tests. Rapid and Reference Tests: The Hema-Strip HIV (SDS, Beverly, MA) is a rapid, one- step, lateral flow immunoassay that incorporates a protein A colloidal gold conjugate, includes a procedural control, and provides results within 15 minutes. The reference assays included the Abbott HIVAB HIV-1/HIV-2 (rDNA) EIA Kit (Abbott Laboratories, Abbott Park, IL), the Genetic Systems HIV Types 1 and 2 (recombinant and synthetic peptide) EIA, (Bio-Rad Laboratories, Redmond, WA), Genetic Systems HIV-1 Western blot (Bio-Rad Laboratories, Redmond, WA), and the Roche Amplicor HIV-1 Monitor Test (version 1.5) (Roche, Indianapolis, IN). Algorithm: Serum samples from each participant were tested by the Abbott HIV 1/2 EIA and the Hema-Strip HIV test. Samples initially reactive by EIA or the Hema-Strip HIV test were repeated in duplicate by the same assay. Western Blots were performed on specimens exhibiting a repeatedly reactive Abbott EIA and/or a reactive result on the Hema-Strip HIV test. Western Blots were not performed on samples from known HIV Positive participants as long as the EIA was reactive. Bio-Rad Genetic Systems HIV-1/2 EIA testing (in triplicate) and the Roche Amplicor NAT were performed on all discordant samples for resolution. Interpretation of Results: Sensitivity, specificity, and positive and negative predictive values of the Hema-Strip HIV test were determined for each of the 4 sample matrices. Overall, the sensitivity was based on 568 (37%) true HIV positive samples and the specificity was based on 947 (63%) true negative samples. All statistical tests were performed in StatXact, and 95% confidence intervals were also calculated for the performance characteristics of the Hema-Strip HIV test. Conclusion Background HIV tests are indispensable for providing results in a clinically-relevant time frame for occupational exposures and women in labor without known HIV status, and in public health clinics to maximize the reporting of results to subjects in a timely manner to minimize the effects of loss to follow up. Moreover, it is increasingly becoming important, especially in developing nations where resources are scarce, that alternate sample matrices/media (such as fingerstick whole blood and saliva samples) be available for testing by HIV assays. Of the 1515 samples tested, 947 were classified by the reference test(s) as being from non-infected subjects, and 568 were from infected persons. The Hema-Strip HIV test correctly identified 1513 of the 1515 samples producing 566 positive results and 949 negative results. Overall, there were 3 false negatives and 1 false positive. This resulted in an overall sensitivity of 99.5% (95% CI: 99.4, 100) and an overall specificity of 99.9% (95% CI: 99.4, 100), thereby exceeding the 98% lower bounds of acceptable performance. Overall Performance Characteristics (n=1,515) - Sensitivity and Specificity Site summary with participant populations (n=1,515) Performance Characteristics by Population – Low Risk - Sensitivity and Specificity Performance Characteristics by Population – High Risk - Sensitivity and Specificity Performance Characteristics by Population – Known HIV Positive - Sensitivity and Specificity RESULTS Overall Performance Characteristics (n=1,515) - Positive and Negative Predictive Value The positive and negative predictive values among all populations were estimated at 99.8% and 99.7%, respectively. The sensitivities for the different blood matrices were between 99.5 and 99.7%, and the specificities were all 99.9%. Only one site produced sensitivity values below 99%, and all sites produced specificity values above 99.1%. Performance Characteristics by Population – Low Risk - Positive and Negative Predictive Value Performance Characteristics by Population – High Risk - Positive and Negative Predictive Value Performance Characteristics by Population – Known HIV Positive - Positive and Negative Predictive Value LR = Low Risk, HR = High Risk, Pos = Positive The Hema-strip HIV test has exhibited excellent performance characteristics when evaluated at multiple sites and using four blood matrices. The test meets expectations for use in low risk, high risk, and HIV positive populations. The Hema-Strip HIV test is simple to perform, requires minimal procedural steps, and is suitable for point of care testing applications. It is currently pending FDA PMA approval. Poster #2 HIV Diagnostics Meeting March 2005 Orlando, Florida