CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: Blinded panels can assess the relative performance of HIV blood screening and diagnostic assays. We confirmed similar performance across subtypes between FDA-approved 4th/5th gen and p24 assays, enhanced sensitivity of next generation p24 platforms, and poor performance of POC assays. This diversity panel allows the evaluation of diagnostic assays to enable unbiased assessment of performance and is essential to validation of new diagnostic tools for identifying future recombinant strains of HIV.

569 EVALUATION OF THE CDC/APHL HIV DIAGNOSTIC ALGORITHMWITH RECENTLY FDA-APPROVED ASSAYS Wei Luo, William Fowler, Krystin A. Price, Vickie Sullivan, Timothy Granade, Steven Ethridge, Laura Wesolowski, Kevin P. Delaney, WilliamM. Switzer, S. Michele Owen, Silvina Masciotra CDC, Atlanta, GA, USA Background: Two serologic assays, the BioPlex 2200 HIV Ag-Ab screening assay (BPC) that detects and differentiates p24 antigen (Ag) and HIV-1 and HIV-2 antibodies (Ab) and the Geenius TM HIV-1/2 supplemental assay (Geenius) that differentiates HIV-1 from HIV-2 Ab, were approved after the 2014 CDC/ APHL laboratory HIV diagnostic algorithmwas released. We evaluated the performance of these assays in HIV-1 seroconverters (SC) and HIV-2 infections in the context of the diagnostic algorithm. Methods: 501 sequential plasma specimens (134 collected during antiretroviral therapy (ART)) from 49 U.S. HIV-1 SC and plasma from 100 HIV-2 Ab-positive persons (16 from the U.S. and 84 from Ivory Coast (IVC)) were tested with BPC and Geenius. Nucleic acid testing (NAT) was done with APTIMA HIV-1 RNA qualitative assay in 497 SC samples and with the Roche COBAS AmpliPrep/ TaqMan HIV-1 v2.0 test in four. The algorithmwas evaluated individually for each SC collection time. Results: In SC, BPC was non-reactive (NR) in 91 samples, 67 of which were collected before the first NAT-reactive (R) and 24 were NAT-only R. After the first NAT-R, 407/434 (93.8%) were BPC-R and 27 were BPC-NR including four in a secondary negative diagnostic phase in three SC. Of 407 BPC-R, 315 (77.4%) were Geenius HIV-1-positive, 31 (7.6%) HIV-1-indeterminate, and 61 (15%) HIV-negative. All 41 BPC Ag-only R samples were Geenius HIV-negative. APTIMA resolved 88/92 (95.7%) Geenius HIV-negative or HIV-1-indeterminate samples. Overall, the algorithm detected 92.9% of early HIV-1 infections including 99.3% of samples during ART. Among 100 HIV-2 samples, 95 were BPC HIV-2 Ab-R (16 U.S., 79 IVC) and five from IVC were HIV-1/HIV-2 Ab-undifferentiated. Of 95 BPC HIV-2 Ab-R samples, 81 were Geenius HIV-2 positive (including 46 with HIV-1 cross-reactivity), 2 repeatedly HIV-2 indeterminate (gp36 reactivity only), 9 HIV untypable, 1 HIV indeterminate and 2 repeatedly HIV negative. Of 5 BPC HIV-2 undifferentiated from IVC, all were Geenius HIV-2 positive. Conclusion: BPC performed well and detected Ag in early HIV-1 infections, but rarely a second negative diagnostic phase was observed. In the algorithm, BPC Ag-only R samples could be confirmed directly with NAT. BPC differentiated HIV-1 from HIV-2 Ab in most samples, but in combination with Geenius results, the possible final interpretations increased. These results provide important data for potential updates to the CDC/APHL recommendations for HIV diagnostics.

Poster Abstracts

568 INCREASING POSITIVITY FOR NON-PLASMA SAMPLES USING THE ABBOTT REALTIME HIV-1 ASSAY Ming Chang , Junhui Wang, Jared Baeten, AndrewMujugira, Robert Coombs University of Washington, Seattle, WA, USA Background: Measuring HIV-1 RNA levels in various biological compartments of infected patients provides insight to viral shedding, transmission, reservoirs, and pathological manifestations. Commercial HIV-1 RNA quantitative assays approved by the FDA for plasma samples are often used to quantify HIV-1 RNA in non-plasma specimens. Endocervical swabs, for example, contain bacterial and human byproducts which may affect PCR amplification of HIV-1 and subsequently generate false negative results. In this study, we retrospectively investigated endocervical swab results tested by the Abbott RealTime HIV-1 assay (Abbott). Methods: Abbott assay-derived HIV RNA results, cycle numbers (CN), MaxRatios (MR), and amplification graphs for 762 endocervical swab samples collected from HIV-1 infected female individuals within the first year of ART in Kenya and Uganda were analyzed. Plasma assay reported a numerical HIV RNA copy, based on the CN at which the amplification curve crossed the defined HIV threshold (ΔRn=0.010), for samples with MR≥ 0.070; samples with MR< 0.070 yielded a “not detected” result. Amplification profiles of default “not detected” samples were visually evaluated one-by-one; if their curves crossing the threshold, CNs were estimated, otherwise samples’ results were determined be HIV-1 negative. Results: One hundred and fifty six endocervical swabs were detected to be HIV positive using the default cutoff and confirmed upon observation of the amplification curve crossing the threshold (Table 1). Among 606 default “not detected” endocervical swab samples, 188 (31%) swab samples were found to contain amplification curves crossing the threshold with MR ranging 0.011 to 0.068; 64 swab samples had HIV RNA ranging from 2.49 to 4.29 log 10 copies/ swab, and 124 swab samples had < 2.39 log 10 copies/swab. PCR interfering agents may cause more flat amplification profiles for swab samples, resulting in lower MR values and failing the default criterion. Conclusion: In addition to HIV RNA results, the Abbott HIV-1 assay provided MR values and amplification graphs useful for examining non-plasma sample results. Thus, in this study an additional 188 HIV positive endocervical swab were identified; HIV RNA positivity significantly increased from 20% (156/762) to 45% (344/762) in swabs collected from infected female individuals after initiation of ART, implying a greater rate of genital shedding than previously reported.

CROI 2018 209