|
|
Abstract
Laboratory evaluation of a rapid HIV testing algorithm
D. Kissin1, Z. Lisitsina2, L. Krutitskaya2, S. Hillis1, D. Jamieson1, B. Branson3, N. Akatova4, J. Robinson5, W. Miller6
Background: In St. Petersburg, Russia, HIV-infected patients often do not return for confirmatory results. Rapid testing algorithm (RTA) may facilitate linkage to care by allowing immediate confirmation of HIV status. We compared performance of RTA and conventional HIV diagnostic algorithm (CDA).
Methods: For the CDA, serum specimens from women presenting for delivery with unknown HIV status in 2005-2006 were prospectively screened with rapid test (RT, Determine) and IV-generation EIA (Genscreen) and, if either was positive, confirmed with Western Blot (WB). All stored, frozen WB-positive (n=74) and WB-indeterminate (n=5) specimens and a random sample of 181 Determine- and EIA-negative specimens, were thawed and tested with 2 additional RTs: Oraquick and UniGold. Sequential RTA with Determine as screening, and Oraquick and UniGold as confirmatory tests (RTA classified as positive if all 3 RTs were positive) was evaluated. HIV status was considered negative if EIA and all 3 RT were negative, and positive if WB was positive. If EIA and the 3 RTs or WB were inconsistent, HIV status was determined by PCR (Amplicor).
Results: Sensitivities of each RT were 97.5% and of EIA, 100%. Specificities of each RT and of EIA were 100%. RTA would have triggered immediate referral for follow-up for 97.5% (77/79) HIV-positive patients by identifying 74 as positive and 3 as indeterminate. CDA would have triggered follow-up for all (79/79) HIV-positive patients by identifying 74 as positive and 5 as indeterminate (EIA detected HIV in 2 specimens that were negative on the initial RT). Both algorithms correctly identified 181 HIV-negative specimens.
Conclusions: Determine, OraQuick and Uni-Gold rapid tests had equivalent performances. Although CDA and RTA had perfect specificities, sensitivity of the RTA was lower. Nevertheless, prospective assessment of both algorithms to compare net effect on linkage into care is warranted.
AIDS 2008 - XVII International AIDS Conference
Abstract no.
CDB0050
Suggested Citation
"D.Kissin, et al.
Laboratory evaluation of a rapid HIV testing algorithm.
:
AIDS 2008 - XVII International AIDS Conference:
Abstract no.
CDB0050"
|
|
|