To the Editors:
In the Letter to-the-Editor, Scott et al1 and Hamers et al2 reported the results of 2 studies that evaluated the diagnostic accuracy of the oral fluid (OF)-based OraQuick (OraSure Technologies, Inc, Bethlehem, PA) rapid test for the detection of HIV antibodies compared with a blood-based reference standard in a cohort of 150 randomly selected outpatients in Johannesburg Hospital, South Africa (with HIV prevalence of 54%). The study was conducted as a second investigation after Hamers et al2 evaluation study of OF OraQuick in a population of 273 pregnant women attending antenatal clinics in Namibia (with HIV prevalence of 25.6%), which showed a sensitivity of 100% [95% confidence interval (CI): 94.9% to 100%] and a specificity of 100% (95% CI: 98.2% to 100%), yielding positive predictive value (PPV) and negative predictive value (NPV) of 100% in a predominantly subtype C population.2
Although these articles present interesting data on differences between OF-based versus blood-based devices, enthusiasm is markedly diminished by methodological and technical concerns about the data and the significance in the context of high level genetic diversity and relatively low prevalence in Cameroon (with HIV prevalence of 5.5%).
HIV-1 has been classified into 4 groups, M (major) that has caused the worldwide pandemic, group N (non-M, non-O) originated from simian immunodeficiency virus strains found in chimpanzees, whereas the simian immunodeficiency virus origin of group O (outlier) has not been identified and a new group P reported in a Cameroonian woman living in France.3
Reports of inability of some HIV serologic assays to detect all variants of HIV, including HIV-1 group O and N, have raised concerns about the accuracy of rapid tests available in West Africa and the challenge for ongoing universal access to antiretroviral treatment and HIV prevention in Cameroon.4,5
This comparative study aimed to evaluate the accuracy of the 2 Aware rapid HIV-1/2 test kits for the routine diagnosis of highly divergent strains such as HIV-1 group O and N for surveillance purposes in Cameroon. These devices were evaluated for use with blood serum plasma (BSP) or oral mucosal Transudate (OMT). An additional goal was to look at the performance of these assays on a set of genotyped group M specimens with the subtypes A to H and HIV-2 infection.6
This cross-sectional study was conducted from June 2006 through September 2006, using plasma and oral mucosal transudate in a randomly selected population of HIV-1-infected participants in CHUY (Yaounde University Teaching Hospital). Only participants who gave their informed consent form were enrolled into the study, approved by the National Science and Ethics Committee. We have identified and characterized a large collection of group O-, N-, and P-infected specimens with more than 100 group O infections evaluated to date. To determine the sensitivity of the Aware HIV1/2 BSP devices on variants common in Cameroon, we further tested these kits using a well characterized panel of 30 isolates: HIV-1 group O (n = 18); HIV-1 group N (n = 1); HIV-2 B/A recombinant (n = 1); and several HIV-1 group M clades such as subtype A (n = 3), D (n = 1), unclassified (n = 4) HIV-1 strains based on env-gp41 and negative human plasma sample (n = 2). Each of these panels had been characterized using a variety of serologic (EIAs: HIV1/2 Ab/Ag Combo, Abbott, IL) and genotypic assays using the ViroSeq HIV-1 genotyping system (Celera Diagnostics, Alameda, CA) approved by the US Food and Drug Administration.
Matched fresh whole blood (n = 132) and OMT specimens were collected from each participant. Approximately 5 ml of blood were collected by well-trained personel into EDTA-anticoagulant tubes bearing the corresponding code number of the participant. The OMT specimens were collected using the Aware HIV-1/2 OMT collection device provided within the test kit, following manufacturer's instructions. A portion of the plasma and all of the OF samples were tested on site using the Aware HIV-1/2 BSP and OMT Rapid test kits. The remaining portion of the blood was transported in insulated boxes containing ice packs to the hematology laboratory, centrifuged at 6000g (5000-8000 rpm) and the EDTA plasma stored at -20°C, and later tested in parallel using Determine HIV1/2 OM2 rapid assay (Abbott Laboratories, Tokyo, Japan) and the Murex EIA HIV-1/2 Ab/Ag Combo (Abbott Laboratories, IL) as reference standard test. The Laboratory technicians were blinded for the rapid test results generated in the field. The reference standard results were compared with the paired test results of the Poisson distribution. The main outcomes for diagnostic accuracy were sensitivity, specificity, and predictive values, with 95% CIs.
The cases and results obtained are summarized in Table 1. Sensitivity and specificity for plasma were found to be 92% (95% CI: 75.9% to 97.9%) and 100% (95% CI: 51.0% to 100%) for the Poisson distribution, whereas oral fluid demonstrated a sensitivity of 81.5% (95% CI: 63.3% to 91.8%) and specificity 100% (95% CI: 43.8% to 100%). The lower sensitivity observed with OF was due to 5 false-negative results, from 4 HIV-1 group O with 1 HIV-1 group N samples, as a result of high level diversity. Other possible explanations include the collection of insufficient OF specimen volume possibly related to low titer in OMT sample buffer (<1:1000 dilution for specimens HIV-1 subtype A 01CMR278 or HIV-1 group O T41/2003 in Table 1). Oral fluid-based device yielded PPV and NPV 100% and 37.5%, respectively. All fresh samples HIV-1 group M (CRF02_AG) accounting for more than 80% of HIV strain in Cameroon yielded concordant result with Aware HIV-1/2 BSP and OMT. PPV and NPV on HIV-1 group M were found to be 100% against the gold standard.
Our findings confirm that immunoassay and primers/probes assays test performance on plasma or OF may be affected by HIV viral variation, including subtypes. Because the majority of HIV serologic assays rely on antibody responses to the structural proteins of the virus, especially the envelope glycoprotein gp41, mutations observed in this domain, especially in the immunodominant region composed of the cytotoxic T Lymphocyte epitope (LAVERYLKDQQLL) and the cysteine loop (CSGKLIC), could affect the sensitivity of serologic assays.7 Despite the large genetic diversity (up to 30% in env), 4 of the unclassified HIV-1 group M samples genotypted on gp41 were reactive on Aware devices (Table 1). Aware HIV-1/2 OMT RT kit seems to be the cheapest, safest and easiest to use compared with other available blood test kits. In addition to invasive HIV diagnostic tests using whole blood, plasma, or serum, noninvasive tests are available that are based on detecting antibodies to HIV present in oral fluid. These are more user friendly and may be more acceptable, especially to hard-to-reach populations, such as commercial sex workers and community-based research settings.8
In conclusion, the evaluated rapid tests, plasma and oral fluid based rapid tests do not perform well on HIV-1 group N and O, but have a good performance on HIV-1 group M limiting use in West Central Africa. Oral fluid-based devices have good potential for use as rapid point-of-care and very important component of HIV control initiatives and programs.
The authors acknowledge the support of Calypte grant #V010305R.
Nicaise Ndembi, MSc, MPhil, PhD*,†
Charlotte Ngansop, BSc*
Sylvie Moudourou, MSc‡
Claude Tagny Tayou, MD*,‡
Alash'le Abimiku, PhD†
Dora Shu Mbanya, MD, PhD*,‡
Lazare Noche Kaptue, MD, PhD*,§
*Université de Yaoundé-I, Yaoundé, Cameroon
†Institute of Human Virology, University of Maryland School of Medicine, Baltimore; Institute of Human Virology-Nigeria, Abuja, Nigeria
‡Centre Hospitalier Universitaire (CHUY), Service d'hématologie et transfusion sanguine, Yaoundé, Cameroon
§Université des Montagnes (UdM), Bangangté, Cameroon
1. Scott L, Noble L, Langeveldt M, et al. Can oral fluid testing be used to replace blood based HIV rapid testing to improve access to diagnosis in South Africa? J Acquir Immune Defic Syndr
. 2009; 51:646-647.
2. Hamers R, de Beer I, Kaura H, et al. Diagnostic accuracy of two oral fluid based tests for HIV surveillance in Namibia. J Acquir Immune Defic Syndr
3. Plantier JC, Leoz M, Dickerson JE, et al. A new Human immunodeficiency virus derived from Gorillas. Nat Med
4. Aghokeng AF, Mpoudi-Ngole E, Dimodi H, et al. Inaccurate diagnosis of HIV-1 group M and O is a key challenge for ongoing universal access to antiretroviral treatment and HIV prevention in Cameroon. PLoS One
5. Granich RM, Gilks CF, Dye C, et al. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet
6. Ndembi N, Abraha A, Pilch H, et al. Molecular characterization of human immunodeficiency virus type 1 (HIV-1) and HIV-2 in Yaounde, Cameroon: evidence of major drug resistance mutations in newly diagnosed patients infected with subtypes other than subtype B. J Clin Microbiol
7. Oral P, Svennerholm B, Jeansson S, et al. Continuous epitopes of the human immunodeficiency virus type 1 (HIV-1) transmembrane glycoprotein and reactivity of human sera to synthetic peptides representing various HIV-1 isolates. J Virol
8. Kagulire SC, Stamper PD, Opendi P, et al. Performance of two commercial immunochromatographic assays for rapid detection of antibodies specific to human immunodeficiency virus types 1 and 2 in serum and urine samples in a rural community-based research setting (Rakai, Uganda). Clin Vaccine Immunol