To the Editor:
Surveillance data on HIV prevalence among various at-risk populations are essential for the development and evaluation of HIV control initiatives.1 This is particularly true in sub-Saharan Africa, where the HIV-1 epidemic imposes an ever-increasing burden.2 In developing countries with a high HIV prevalence, most HIV testing for surveillance purposes is conducted as part of seroprevalence surveys among specific population groups, such as women attending antenatal clinics (ANCs) and persons attending voluntary counseling and testing (VCT) sites. 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 mucosal transudate, or oral fluid. These are more user-friendly and may be more acceptable, especially to hard-to-reach populations, such as commercial sex workers.3,4 Therefore, oral fluid-based tests have good potential for use in screening and surveillance activities in high-prevalence developing countries.
Reports from Europe and North America have demonstrated good diagnostic accuracy for 2 oral fluid-based tests that have been approved by the US Food and Drug Administration (FDA): OraQuick Rapid HIV-1/2 Antibody Test (OraSure Technologies, Inc., Bethlehem, PA [“OraQuick”])5-7 and Oral Fluid Vironostika HIV Uni-Form II microELISA (bioMérieux Inc., Durham, NC), which is used in combination with an appropriate oral specimen collection device.3,8,9 Test performance data from field settings in developing countries, particularly sub-Saharan Africa, where HIV-1 infections are caused by a high variety of (“non-B”) subtypes,10 are limited, however.4,11,12 This comparative study aimed to evaluate the diagnostic accuracy of the 2 oral fluid-based HIV tests for surveillance purposes in a high-prevalence population in Namibia.
A cross-sectional study was conducted in pregnant women 18 years of age or older with unknown HIV status who attended the 2 ANCs of Katutura State Hospital, Windhoek, Namibia, and Oshakati State Hospital, Oshakati, Namibia, for VCT and/or antenatal care in June and July 2006. Participants provided 2 oral fluid specimens using the OraQuick and OraSure (OraSure Technologies, Inc.) oral devices in accordance with the manufacturer's instructions, which were labeled and processed anonymously. For reference standard testing, paired blood specimens were obtained by venipuncture, which is routinely performed in pregnant women for mandatory syphilis (rapid plasma reagin [RPR]) screening.
The reference standard used was the Namibian national algorithm for HIV testing, which consists of serial dual enzyme-linked immunoassay (ELISA) testing on blood plasma to detect antibodies to HIV-1 and HIV-2. The baseline test was the AxSYM HIV 1/2 gO assay (Abbott Laboratories, Abbott Park, IL). A sample rate by a cutoff rate (S/CO) value of >10.00 was considered clearly reactive. Specimens with an S/CO value of >0.90 to 10.00 were considered indeterminate and were retested in duplicate after recentrifugation. Specimens found to be reactive or repeatedly indeterminate were confirmed with the Access HIV 1/2 New test (Bio-Rad Laboratories, Marnes-la-Coquette, France).
The OraSure oral fluid specimens were tested with the Oral Fluid Vironostika HIV Uni-Form II microELISA system in accordance with the manufacturer's instructions. Specimens that yielded results discrepant with the reference standard were checked for technical errors and were retested. All ELISA testing for this study was performed at the Namibia Institute of Pathology in Windhoek, which is the national medical reference laboratory in Namibia.
OraQuick test results were read after 20 to 40 minutes and classified in accordance with the manufacturer's instructions. OraQuick external quality control procedures (control kits) were performed in accordance with the manufacturer's recommendations.
Only available specimen sets, consisting of at least 1 oral fluid specimen (OraQuick and/or OraSure) and the paired plasma specimen, were included in the analysis. The reference standard results were compared with the paired test results of the OraQuick and/or OraSure devices. The main outcomes for diagnostic accuracy were sensitivity, specificity, and predictive values, with exact 95% confidence intervals (CIs). Concordance between the OraQuick and OraSure devices was evaluated using the κ statistic.
The study was approved by the Research Ethics Committee at the Namibian Ministry of Health and Social Services. Written informed consent was obtained from all participants before enrollment. Participants did not receive study test results (anonymous, unlinked), but they accessed routine VCT as provided at the 2 study sites.
A specimen set was available for 273 participants (175 at Katutura and 98 at Oshakati). Testing of the blood specimens by reference standard identified 70 HIV-1 antibody-positive and 203 HIV-1 antibody-negative specimens, yielding a prevalence of 25.6% (95% CI: 20.4 to 30.8). One participant had an indeterminate test result at reference standard baseline testing, which turned out to be negative at confirmatory testing. No OraQuick and OraSure test results were reported as invalid or indeterminate. OraQuick results were available for all 273 women, showing an accuracy of 100% (Table 1). OraSure results were excluded from analysis for 6 (2.2%) women because of labeling errors. The OraSure test recorded 3 (1.1%) discrepant results among 267 tests, namely, 2 false-negative results and 1 false-positive result, yielding a sensitivity of 97.1% (95% CI: 89.9 to 99.6) and a specificity of 99.5% (95% CI: 97.2 to 100) (see Table 1). Concordance between OraQuick and OraSure test results was high (κ = 0.97; 95% CI: 0.95 to 0.99).
This study, which is the first formal field evaluation of the OraQuick and OraSure devices in a resource-limited setting in southern Africa, demonstrates that the OraQuick test is 100% accurate and that the OraSure test has high specificity (99.5%) but slightly lower sensitivity (97.1%) in a high-prevalence population in Namibia. The high negative predictive value of both tests (≥99%) renders them particularly suitable for surveillance purposes. Our findings contribute to the limited existing data on diagnostic accuracy of oral fluid-based HIV tests among populations in resource-limited settings that are infected with non-B subtypes.4,11,12
OraQuick test performance data from resource-limited countries are limited; however, overall, they show high specificity (100%) and varying sensitivity (75% to 100%).4,11,12 Notably, a recent field study on the OraQuick test device in a high-prevalence population in rural India showed 100% (95% CI: 98 to 100) sensitivity and 100% (95% CI: 99 to 100) specificity,4 which is consistent with our findings.
The 3 discrepant OraSure-Vironostika test results (ie, 2 false-negative results [Katutura and Oshakati] and 1 false-positive result [Katutura]) were checked, and the tests were repeated; however, no additional evaluations were performed to resolve the true serostatus of the specimens. This is a limitation of the study. A possible explanation for the 2 false-negative results is the collection of insufficient oral specimen volume by the collection device; this limitation has been reported previously and warrants the development of safeguards against inadequate sampling.13 Alternatively, the false-negative results may be explained by the presence of low HIV antibody levels during the early phase of infection. Furthermore, it cannot be ruled out completely that the reference standard used produced false-positive test results because it did not include a true confirmatory test, such as a Western blot or p24 antigen test. Finally, it can be speculated that the 1 false-positive result may be associated with cross-reactivity of nonspecific antibodies in the patient specimen as a result of chronic immune system activation in Africans attributable to an increased load of environmental pathogens.14,15
Immunoassay test performance on serum or plasma may be affected by HIV viral variation, including subtypes.16 In Namibia, although no data have been published to date, it is assumed that, like elsewhere in southern Africa, subtype C is the predominant HIV-1 subtype.10 To our knowledge, this study is the first to report on ELISA test performance with oral fluid in a population predominantly infected with subtype C. In view of the dominance of subtype C in high-prevalence areas, such as southern and East Africa and India, the validated oral fluid tests can be of great value to enhance surveillance efforts where these are needed most.
The HIV infection rates we found in this study among participants from Katutura (25.1%) and Oshakati (26.5%) are consistent with results of the 2006 national sentinel serosurvey among pregnant women, which reported estimated HIV-1 prevalences of 21.1% and 27.1%, respectively, for these sites.17 Although this study was not designed to measure prevalence and 95% CIs are wide because of the limited sample size, this finding adds to the assumption that the study population was representative of the general ANC population.
Whereas the OraQuick (rapid) test would be most suitable for areas with limited laboratory resources, the OraSure-Vironostika tests would be preferred when complete confidentiality is required, when real-time diagnosis is not necessary, and in high-volume batch testing, rendering the OraSure-Vironostika test particularly suitable for anonymous surveillance testing. Although the manufacturer already offers differential pricing for the OraQuick and OraSure test kits in sub-Saharan Africa, oral fluid-based tests are still more expensive compared with the conventional HIV tests. If their use expands over time, further reduction of prices might be negotiable. Inclusion in the World Health Organization's HIV Test Kit Bulk Procurement Scheme might be helpful to achieve this.18
In conclusion, the OraQuick and OraSure oral fluid-based HIV tests showed high diagnostic accuracy in a high-prevalence population in Namibia, supporting their public health utility in surveillance initiatives in resource-limited settings in sub-Saharan Africa. The advantages of convenience and user-friendliness may improve patient participation, thus facilitating the collection of reliable prevalence data among various at-risk populations. Additional applications, such as in VCT, should be investigated. Ongoing evaluation of oral fluid testing against the national reference standard as part of a formal quality assurance program is warranted. The development of safeguards against inadequate sampling volume is indicated.
Raph L. Hamers, MD* Ingrid H. de Beer, MA† Harold Kaura, MSc‡ Michèle van Vugt, MD, PhD*§ Lucille Caparos, MSc‡ Tobias F. Rinke de Wit, PhD*
*PharmAccess Foundation Center for Poverty-Related Communicable Diseases, and Academic Medical Center of the University of Amsterdam Amsterdam, The Netherlands †PharmAccess Foundation Namibia Windhoek, Namibia ‡Namibia Institute of Pathology Windhoek, Namibia §Division of Infectious Diseases Tropical Medicine, and AIDS Academic Medical Center of the University of Amsterdam Amsterdam, The Netherlands
1. Mishra V, Vaessen M, Boerma JT, et al. HIV testing in national population-based surveys: experience from the Demographic and Health Surveys. Bull World Health Organ
2. WHO-UNAIDS. AIDS Epidemic Update. December 2007
. Geneva, Switzerland: WHO; 2007.
3. Gallo D, George JR, Fitchen JH, et al. Evaluation of a system using oral mucosal transudate for HIV-1 antibody screening and confirmatory testing. OraSure HIV Clinical Trials Group. JAMA
4. Pant PN, Joshi R, Dogra S, et al. Evaluation of diagnostic accuracy, feasibility and client preference for rapid oral fluid-based diagnosis of HIV infection in rural India. PLoS ONE
5. Delaney KP, Branson BM, Uniyal A, et al. Performance of an oral fluid rapid HIV-1/2 test: experience from four CDC studies. AIDS
6. Jafa K, Patel P, MacKellar DA, et al. Investigation of false positive results with an oral fluid rapid HIV-1/2 antibody test. PLoS ONE
7. Wesolowski LG, MacKellar DA, Facente SN, et al. Post-marketing surveillance of OraQuick whole blood and oral fluid rapid HIV testing. AIDS
8. Granade TC, Phillips SK, Parekh B, et al. Detection of antibodies to human immunodeficiency virus type 1 in oral fluids: a large-scale evaluation of immunoassay performance. Clin Diagn Lab Immunol
9. Soto-Ramirez LE, Hernandez-Gomez L, Sifuentes-Osornio J, et al. Detection of specific antibodies in gingival crevicular transudate by enzyme-linked immunosorbent assay for diagnosis of human immunodeficiency virus type 1 infection. J Clin Microbiol
10. Hemelaar J, Gouws E, Ghys PD, et al. Global and regional distribution of HIV-1 genetic subtypes and recombinants in 2004. AIDS
11. Bhore AV, Sastry J, Patke D, et al. Sensitivity and specificity of rapid HIV testing of pregnant women in India. Int J STD AIDS
12. Reynolds SJ, Ndongala LM, Luo CC, et al. Evaluation of a rapid test for the detection of antibodies to human immunodeficiency virus type 1 and 2 in the setting of multiple transmitted viral subtypes. Int J STD AIDS
13. Mortimer P, Parry J. Non-invasive virological diagnosis: are saliva and urine specimens adequate substitutes for blood? Rev Med Virol
14. Biggar RJ, Gigase PL, Melbye M, et al. ELISA HTLV retrovirus antibody reactivity associated with malaria and immune complexes in healthy Africans. Lancet
15. Messele T, Abdulkadir M, Fontanet al, et al. Reduced naive and increased activated CD4 and CD8 cells in healthy adult Ethiopians compared with their Dutch counterparts. Clin Exp Immunol
16. Koch WH, Sullivan PS, Roberts C, et al. Evaluation of United States-licensed human immunodeficiency virus immunoassays for detection of group M viral variants. J Clin Microbiol
17. Ministry of Health and Social Services. Results of the 2006 national sentinel survey among pregnant women. Press release, April 17, 2007. Windhoek, Namibia: Ministry of Health and Social Services; 2007.