CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

blood spots (DBS) at the CDC or locally. At CDC, those consenting to storage of DBS were confirmed using Abbott RealTime HIV-1 assay (VL), Bio-Rad GS HIV Combo Ag/Ab EIA and Geenius HIV-1/2 assays. Self-reported data and valid test results were analyzed using SAS. Results: The false-negative rate was 2.3% (45/1936). Of 1936 participants, 42.4%were tested with INSTI (21/820, 2.6% RT-NR), 31.1%with Determine (10/603, 1.7% RT-NR), 12.3%with Sure Check (5/239, 2.1% RT-NR), 11.6%with OraQuick (5/224, 2.2% RT-NR) and 2.6%with Uni-Gold (4/50, 8.0% RT-NR). The table shows reactivity of RTs from participants by VL results and self-reported ART use. Of 1655 RT-R participants, 1311 (79.2%) had undetectable VL or detected <2.92 log 10 (cop/mL) of whom 1263 (96.3%) reported being on ART. Of 18 RT-NR participants, 17 (94.4%) had undetectable VL of whom 16 (94.1%) reported being on ART. The laboratory-based serology testing algorithm did not confirm HIV-positive status in 5 of 18 RT-NR persons self-reported to be living with HIV and on ART (2 Determine, 1 OraQuick, 1 Uni-Gold, 1 Sure Check). Conclusion: False non-reactivity of rapid HIV tests occurred but was low and consistent across most RTs. In a small number of samples, VS was associated with non-reactivity possibly due to seroreversion; however, the percent of participants virally suppressed on ART was similar among those who were RT-R and RT-NR. Given the sensitivity limitation of RTs, our results highlight challenges with relying on rapid HIV testing alone, particularly in circumstances of VS in which non-reactivity could lead to misinterpretation of HIV status. This could have implications for monitoring for virologic breakthroughs with PrEP and surveillance systems that use RTs to gauge HIV prevalence.

1 University of California San Diego, San Diego, CA, USA, 2 East Bay AIDS Center, Oakland, California, 3 East Bay AIDS Center, Oakland, CA, USA Background: Universal opt-out HIV screening in low prevalence settings such as emergency departments (EDs) has increased identification of persons with HIV infection. However, false positive (FP) 4th generation HIV test results may impact the positive predictive value (PPV) and lead to a delay of disclosure of HIV diagnosis. The objective of this analysis was to assess factors associated with false-positive test results. Methods: Opt-out HIV screening was conducted among adults at four California locations (two EDs at UC San Diego from July 2017 - March 2019 and two EDs at Alta Bates Summit Medical Center in Oakland fromMay 2017 - March 2019) using a 4th generation HIV Ag/Ab combination assay. We identified all individuals with FP HIV Ag/Ab results. Demographics, clinical data (ED chief complaints, discharge diagnoses, and medical conditions), and HIV risk factors were extracted from electronic medical records and compared with data from individuals with true positive (TP) HIV test results using non-parametric statistical tests. Results: A total of 32,450 HIV tests were performed across four EDs using a 4th generation Ag/Ab assay (Architect® and Roche Elecsys®) resulting in 104 TP cases and 34 FP cases (PPV: 75.4%; FP rate: 0.1%). Among FP cases, the median age was 42 (IQR: 32-55), more than half (64.7%) were female, and more than half (58.8%) were White (Table). In univariate analyses, FP cases were significantly more likely than TP cases to be female (64.7% vs 28.9%, p < 0.05), White (63.6% vs 35.6%, p < 0.05), and pregnant (9.7% vs 0%, p < 0.05). None of the false-positive cases were in men who have sex with men and none were persons who inject drugs. Several factors were common (>20%) but not statistically significant: history of flu vaccination (lifetime) (65.5%), history of multiparity (30.0%), and obesity (24.2%). Additionally, 3 cases had a history of FP HIV tests and 1 case had autoimmune hepatitis. Conclusion: The PPV of 4th generation HIV tests was suboptimal during universal opt-out HIV screening in EDs at two medical centers in California. Individuals who were female, White, and pregnant were more likely to have FP tests. Understanding these factors associated with FP test results in a population with low pretest probability may be important for early HIV disclosure as universal HIV testing in low-prevalence settings becomes more commonplace.

Poster Abstracts

968 PERFORMANCE OF ORAQUICK RAPID TEST ON HIV DIAGNOSIS AMONG CADAVERS IN KISUMU, KENYA Frankline O. Mboya 1 , Anthony Waruru 2 , Alex Sila 3 , Dickens Onyango 4 , Solomon Sava 4 , Lily Nyaga 4 , Mary Mwangome 5 , Nyakeriga Nyakeriga 5 , Muthoni Junghae 2 , Paul K. Musingila 1 , Frank L. Basiye 1 , Kimberly McCarthy 1 , Boaz Oyaro 6 , Peter W. Young 7 1 US CDC Kisumu, Kisumu, Kenya, 2 US CDC Nairobi, Nairobi, Kenya, 3 Galicia Sur Health Research Institute, Vigo, Spain, 4 Ministry of Health, Nairobi, Kenya, 5 Global Solutions for Infectious Diseases, San Francisco, CA, USA, 6 KEMRI–Centre for Global Health Research, Kisumu, Kenya, 7 US Department of State Nairobi, Nairobi, Kenya Background: Ascertaining HIV status at time of death can be useful for identifying missed opportunities to diagnose and treat HIV infection. Routinely testing HIV status among deaths, though not commonly practiced, may complement vital statistics in Kenya, where cause-specific attribution to death is often not documented. Currently available and validated rapid test kits (RTKs) in Kenya require a blood sample, which is hard to draw from cadavers. To avoid logistical difficulties of blood-based HIV testing, a minimally invasive assay using oral fluid such as OraQuick®, may be an alternative. OraQuick® has previously been used in Kenya for HIV self-testing, showing a higher specificity than sensitivity. We verified the feasibility and diagnostic accuracy of OraQuick® for HIV screening among cadavers. Methods: Trained morticians collected pre- and post-embalming oral fluids from 132 cadavers >18 months old at Jaramogi Oginga Odinga Teaching and Referral Hospital (JOOTRH) mortuary in Kisumu, Kenya. These were tested for HIV using OraQuick®. Test results from OraQuick® were compared with those obtained using national RTK algorithm on matched pre-embalming whole blood specimens as a gold standard (Determine® HIV and First Response® HIV 1-2-O). We calculated positive predictive value (PPV), negative predictive value (NPV), false detection rate (FDR), false omission rate (FOR), sensitivity and specificity of OraQuick® compared to the gold standard.

967 EVALUATION OF VIRAL SUPPRESSION ON RAPID HIV TEST REACTIVITY AMONG MSM, NHBS, 2017 Shamaya Whitby 1 , Amanda Smith 2 , Johanna Chapin-Bardales 2 , Rebecca Rossetti 2 , Cyprian Wejnert 2 , Silvina Masciotra 2 , for the for the NHBS Study Group 1 Oak Ridge Institute for Science and Education, Atlanta, GA, USA, 2 CDC, Atlanta, GA, USA Background: Antiretroviral therapy (ART) leads to viral suppression (VS) and potentially seroreversion. However, the impact of sustained VS on the ability of rapid tests (RTs) to identify HIV infection has not been extensively reported. To assess RT performance in populations with likely exposure to antiretrovirals, we evaluated RT results among self-reported HIV-positive (SRP) men who have sex with men (MSM) in 23 U.S. cities participating in 2017 National HIV Behavioral Surveillance (NHBS). Methods: Sites performed at least one point-of-care RT on all consenting SRP MSM. Participants with RT-nonreactive (RT-NR) results were considered discrepant and resolved with further laboratory testing using plasma or dried

CROI 2020 363