Performance evaluation of four point-of-care HIV tests using unprocessed specimens

Abstract

Background: The performance of recently approved point-of-care (POC) HIV tests should be assessed using unprocessed specimens.

Objective: To evaluate the sensitivity and specificity of four POC HIV tests using whole blood (WB) and two using oral fluid (OF) among persons recruited from health clinics in Seattle, Washington, during September 2015-September 2017.

Study design: Participants were tested with the POC tests, additional plasma and serum were collected for laboratory testing, and participant- reported use of antiretroviral therapy (ART) or pre-exposure prophylaxis (PrEP) was recorded. Participants testing negative on all tests could reenroll every 90 days. Specimens from persons previously diagnosed with HIV infection as well as from those who were newly diagnosed during the study were included in the sensitivity estimate. Sensitivity and specificity were calculated based on HIV status determined by laboratory testing.

Results: Of 1,256 visits, 179 were from persons with HIV infection; 120 of these were taking ART. Among 1,077 visits from participants not diagnosed with HIV, PrEP use was reported at 155 (14.4%) visits. Sensitivity was similar among POC WB tests (95.53%-97.21%; p>0.05). Among participants on ART, sensitivity was lower for the same test performed on OF compared to WB (p<0.003). Specificity was high for all tests (99.44%- 100.00%); we did not detect specificity differences with PrEP use.

Conclusions: These POC tests displayed relatively high sensitivity and specificity using unprocessed specimens, suggesting their effectiveness in identifying HIV infections whenever laboratory-based testing is not feasible. Nonetheless, clients with recent risk should retest to rule out the possibility of a false-negative result.

Keywords: HIV; Performance; Point-of-care; Rapid test; Unprocessed specimens.

Fig. 1.

Algorithm used to determine HIV infection status^a. Abbreviations: Ag/Ab, antigen/antibody; Ab, antibody; NAT, nucleic acid test; Geenius, Geenius^™ HIV 1/2 Supplemental Assay. Footnotes: ^a There were several exceptions to this algorithm:

1. One participant missing the laboratory screening test and with a negative pooled NAT was classified as not having HIV infection.

2. One participant missing the screening laboratory test and with a negative Geenius result and a positive HIV-1 NAT result was classified as having acute HIV infection.

3. One participant missing the screening laboratory test and with a positive Geenius result was classified as having established HIV infection.

4. Three participants with non-reactive Ag/Ab results and with insufficient blood for NAT were classified as not having HIV infection.

^b For 15 participants recruited early in the study, the Ag/Ab test was not available and the result of an Ab-only test (GS HIV-1/HIV-2 PLUS O EIA, BioRad Laboratories) performed on serum was used as the first step of the algorithm. ^c During this period, there were no HIV-2 positive results from the HIV-1/HIV-2 antibody differentiation test. ^d Individual NAT was performed when there was a record of a previous HIV-positive result or when one of the point-of-care tests conducted during the visit had a positive result or when the Ag/Ab screening test had a reactive result. Pooled NAT was conducted when all point-of-care tests resulted negative during the visit. Pooled NAT was originally performed on 27-member pools and after 10/12/2015 on 10-member pools.