CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Cascades were similar by age and sex, but the lowest wealth quintile had a lower cascade compared to the highest quintile (leading to 43% (95% CI: 37-49) vs. 49% (95% CI: 41-57) virally suppressed). Conclusion: Combining self-reported information with biomarker data reveals the limitations of self-report for HIV cascade construction, and improves cascade depth and accuracy. We find that major gaps in care for older South Africans with HIV persist, in particular at the HIV diagnosis and viral suppression stages. 992 MISDIAGNOSED HIV INFECTION IN PREGNANT WOMEN INITIATING UNIVERSAL ART IN SOUTH AFRICA Allison Zerbe 1 , Nei-yuan Hsiao 2 , Tamsin Phillips 3 , Kirsty Brittain 3 , Elaine J. Abrams 1 , Landon Myer 3 , for the Maternal Child Health - AntiretroviralTherapy (MCH-ART) Study 1 ICAP at Columbia Univ, New York, NY, USA, 2 Div of Med Virology, Cape Town, South Africa, 3 Univ of Cape Town, Cape Town, South Africa Background: Rapid diagnostic tests (RDTs) are used globally to diagnose HIV infection, but with universal ART eligibility there is concern that false positive (FP) RDTs could result in misdiagnosis of HIV infection and inappropriate ART initiation. The use of confirmatory viral load testing before ART initiation (pre-ART VL) is suggested to identify FP RDT results but there are few insights into the potential consequences of universal pre-ART VL testing. Methods: The study took place in a public sector primary care facility in Cape Town, South Africa, where HIV diagnosis employs two RDTs following WHO algorithms. As part of a larger study of ART in pregnancy, we conducted pre-ART VL testing (Abbott RealTime HIV-1) in consecutive HIV+ pregnant women making their 1st antenatal clinic (ANC) visit who were not on ART or ARV prophylaxis according to interview and antenatal record review. We describe the final HIV status of women based on pre-ART VL testing and other investigations, including the proportion of women found to be erroneously started on ART due to FP RDT results, and estimate the costs to identify one erroneous ART initiation using pre-ART VL testing. Results: In 952 pregnant women diagnosed HIV+ based on RDT algorithms and reporting no current ART use, the median CD4 cell count was 352 cells/mm3 (IQR, 236-509). In pre-ART VL testing, 37 women (4%) were aviraemic with no detectable VL and were investigated further as suspected FP from RDTs. Of these, 22% (8/37) were subsequently confirmed to be on ART frommedical records and 13% (5/37) had detectable virus on subsequent VL measurements; the remaining 24 underwent additional testing using ELISA. In the ELISAs conducted, 2 women were found to be HIV-negative, representing 5% of all aviraemic women (2/37) and 0.2% (2/952) of all women identified as HIV-infected by the public sector health services using RDT. Based on this we estimate that approximately $9520 USD (uncertainty interval, $7140-$11,900) would be spent on confirmatory pre-ART VL testing, with further ELISA for aviraemic women, to identify 1 patient erroneously initiated on ART. By contrast, immediate use of an ELISA alone as a confirmatory test would cost $1666 ($1428-$1904) to identify one patient erroneously initiated on ART. Conclusion: False-positive RDT may lead to patients erroneously initiated on ART, but routine pre-ART VL testing appears inefficient compared to ELISA only as an approach to confirm infection before ART initiation. 993 ACCURACY OF HIV AND CD4 FIELD TESTING IN THE BOTSWANA COMBINATION PREVENTION PROJECT Ebi Bile 1 , Pamela J. Bachanas 2 , Fiona Maurice 1 , Sylvia Modise 1 , Liziwe Chebani 3 , Vongai Makovore 1 , Peter Fonjungo 2 , Mary Grace Alwano 1 , William Bapati 4 , Joseph N. Jarvis 1 1 US CDC Botswana, Gaborone, Botswana, 2 CDC, Atlanta, GA, USA, 3 Botswana Ministry of Hlth, Gaborone, Botswana, 4 Tebelopele Voluntary Counseling and Testing Cntr, Gaborone, Botswana Background: Ensuring correct HIV test results are given to participants is critical for both large field-based prevention trials and national HIV-testing programs. During the Botswana Combination Prevention Project (BCPP), a cluster randomized trial designed to evaluate the impact of a combination prevention package on population level HIV incidence in Botswana, robust quality assurance (QA) measures have been implemented to ensure lay counselors deliver accurate test results in the field. Methods: BCPP conducted home-based and mobile HIV testing campaigns in 15 intervention communities from October 2013-February 2016. Lay counselors (130) were trained to conduct rapid HIV testing and point of care PIMA CD4 testing with annual refresher training, weekly internal Quality Control (QC) panel testing, and proficiency testing (PT) 3 times/year. In the first community, 100% of field samples were retested in the reference laboratory using EIA HIV testing. In subsequent communities a random sample of 5-10% of samples were retested. Laboratory based PIMA CD4 counts were repeated on samples from 155 field PIMA CD4 counts. Monthly supervision and monitoring visits were conducted. Results: Weekly QC panels conducted on 1000 positive and 975 negative HIV tests in the field using Determine and First Response assays produced 100% accurate results. Overall, 89% of counselors achieved a pass on PT increasing from 78% at baseline to 96% and 97% in the final two rounds. Repeat EIA testing of 3002 DBS samples demonstrated a 99.6% agreement between field and laboratory HIV-test results, with a kappa score of 0.99, p < 0.0001. Of the 12 discordant results, 4 were false negative and 8 were false positive in field-testing. The overall pass rate for PIMA CD4 PT was 89% (81%, 93%, 95%, and 88% rounds 1-4). Levels of agreement between field and lab PIMA CD4 results are shown in figure 1, with a mean difference of 16.7 cells/µL (95% CI -10 – 43). At the CD4 threshold of 350 there was 86% agreement (21 of 155 misclassified). Conclusion: With a strong training, monitoring, and QA program, BCPP lay counselors conducted HIV-testing in field settings accurately and to a high standard. This indicates that lay counselors can competently and correctly implement HIV testing at scale in field settings, and gives very high confidence in BCPP HIV-test results. Concordance between PIMA CD4 field results and lab results were within expected ranges.

Poster and Themed Discussion Abstracts

994

PATIENTS’ OUT-OF-POCKET EXPENDITURES FOR ART IN SWAZILAND Maria Vaikath 1 , Shaukat Khan 2 , Emma Mafara 2 , Ndumiso Dlamini 2 , Phiwayinkhosi Dlamini 2 , Nothando Ndzinisa 2 , Njabulo Simelane 2 , Velephi Okello 3 , Till Baernighausen 4 , Pascal Geldsetzer 1

CROI 2017 429