In Guinea-Bissau, West Africa, the 2 major strains of HIV (HIV-1 and HIV-2) are cocirculating. Even though the highest prevalence of HIV-2 has been found in Guinea-Bissau,1 the strain has also been found in many other countries due to globalization and widespread immigration.2
As HIV-2 is intrinsically resistant to nonnucleoside reverse transcriptase inhibitors (NNRTIs),3 which is part of first-line treatment of HIV-1 in sub-Saharan Africa, it is mandatory to discriminate between HIV types before initiating antiretroviral treatment (ART).4 The availability of rapid tests has facilitated the implementation of HIV testing in areas without advanced laboratories and highly skilled personnel.
In Guinea-Bissau, a serial algorithm consisting of 2 distinct rapid HIV assays, Determine HIV-1/2 (Abbott Laboratories, Tokyo, Japan) and SD Bioline HIV 1/2 3.0 (Standard Diagnostics Inc, Kyonggi-do, South Korea) was implemented in 2006 for use in local testing centers and hospitals on a nationwide scale. However, in a recent study, from neighboring Guinea-Conakry where 4 rapid HIV tests with the ability to discriminate between HIV types were validated and compared5 serious doubts about the capability of SD Bioline HIV 1/2 3.0 were cast. SD Bioline was the worst overall performer due to misclassification of HIV-1 or HIV-2 seropositive samples as HIV-positive untypable. In November 2011, SD Bioline HIV 1/2 3.0 was delisted from the WHO e-catalog.6 The Genie II HIV-1/HIV-2 test was found to be the most accurate rapid test for discriminating between HIV-1 and HIV-25 in the Conakry study. However, the test requires serum/plasma instead of whole blood and cold chain storage. A newer version, the Genie III HIV-1/HIV-2, can however, be used with serum/plasma or whole blood without the need for cold chain storage.
The aim of this study was to compare 2 new discriminatory assays, using SD Bioline HIV 1/2 3.0 as an internal control, with the gold standard test in a relevant clinical context.
MATERIALS AND METHODS
Study Setting and Serum Sample Collection
The study was conducted at the outpatient ART center at the Hospital National Simão Mendes (HNSM) in Bissau, the capital of Guinea-Bissau, in collaboration with Bandim Health Project, the National HIV Programme, and the National Public Health Laboratory, between July 2011 and December 2011. Included in this study, were 219 consenting HIV-positive adults, previously identified through screening with Determine HIV-1/2 assay (Abbott Laboratories, Abbott Park, IL, USA) and HIV-1 and HIV-2 discrimination with SD Bioline HIV 1/2 3.0, who reported for routine blood analysis at the HIV clinic at HNSM in Bissau.
Selection of Rapid HIV Assays for Evaluation and Comparison With the Gold Standard
SD Bioline HIV 1/2 3.0 was the discriminatory test used in Guinea-Bissau during the study period. Two other rapid HIV tests [Immunoflow HIV1-HIV2 (WB); Core Diagnostics (Birmingham, United Kingdom) and Genie III HIV-1/HIV-2; Bio-Rad (Steenvorde, France)] were selected for evaluation on the basis of their local availability, procedural simplicity, storage requirements, and their previously reported ability to discriminate between HIV-1 and HIV-2.5
All samples were analyzed in Bissau immediately after collection with SD Bioline HIV 1/2 3.0 and Immunoflow HIV1-HIV2 (WB) using heparinized whole blood. Plasma was subsequently separated and frozen at −20°C. Plasma was then tested by the Genie III HIV-1/HIV-2 and the results were confirmed by the INNO-LIA HIV I/II at Aarhus University Hospital, Denmark. All rapid test assays were performed and interpreted as recommended by the manufacturer by 2 well-trained operators in Guinea-Bissau and at Aarhus University Hospital, Denmark. INNO-LIA HIV I/II Score testing was interpreted by a software algorithm after digital scanning of the tests results [Liras for infectious Diseases, v2.01; Innogenetics N.V., Zwijnaarde (G'hent), Belgium].
Data Collection and Statistical Analysis
Information about patient demography and CD4 cell count was collected as previously described.7 CD4 cell counts were included in the study if measured within 90 days of the HIV testing. Kappa statistic was used to assess the performance of rapid assays in the discrimination between HIV types with INNO-LIA as gold standard. For each rapid test, the risk of discrepancy with the gold standard was compared between patients with CD4 cell count below and above 200 cells per microliter with χ2 test. The level of significance was set at P = 0.05. Data were analyzed using Stata IC 11.0 (Stata Corp, College Station, TX).
From a total of 219 serum samples available for this study, 141 were HIV-1 seropositive, 29 were HIV-2 seropositive, and 49 were HIV-1 and HIV-2 seropositive untypable as confirmed by INNO-LIA HIV-I/II (Table 1).
Genie III HIV-1/HIV-2 was the best overall performer with regard to discriminatory capacity (Kappa statistic: 0.84, agreement 91.8%), closely followed by Immunoflow HIV1-HIV2 (WB) (Kappa statistic: 0.82, agreement 90.9%) (Table 2). SD Bioline HIV-1/2 3.0 was the worst performer with regard to discriminatory capacity (Kappa statistic: 0.71, agreement 84.5%). SD Bioline HIV 1/2 3.0 and Immunoflow HIV1-HIV2 (WB) had a sensitivity of 100% for the confirmation of HIV infection, whereas Genie III HIV-1/HIV-2 had a sensitivity of 99.3%, as 1 HIV-1–positive sample was identified as HIV negative. The sample was tested with identical results 3 times using different Genie III kits. In total, Immunoflow HIV1-HIV2 (WB) and Genie III HIV-1/HIV-2 disagreed with the gold standard in 23 samples (2 HIV-1, 7 HIV-2, and 14 HIV-1/2), of which Immunoflow HIV1-HIV2 (WB) disagreed in 20 samples and Genie III HIV-1/HIV-2 disagreed in 18 samples. Interestingly, we found that the 2 rapid tests agreed in 15 of 23 (65.2%) of these cases.
CD4 cell count had been measured in 178 of the patients within 90 days of the discriminatory test performed in Bissau and ranged from 16 to 955 cells per microliter. CD4 cell count was not associated with rapid test failure to discriminate between HIV types for any of the rapid tests: SD Bioline HIV 1/2 3.0 (test failure, below/above 200 cells/μL: 11.4% vs. 18.8%, P = 0.18), Immunoflow HIV1-HIV2 (WB) (7.0% vs. 12.5%, P = 0.22), and Genie III HIV-1/HIV-2 (6.1% vs. 9.4%, P = 0.43). Stratifying for ART (yes/no) did not alter results; SD Bioline HIV 1/2 3.0 (15.4% vs. 15.8%, P = 0.95), Immunoflow HIV1-HIV2 (WB) (8.6% vs. 10.5%, P = 0.67), and Genie III HIV-1/HIV-2 (8.6% vs. 7.0%, P = 0.70). In 17 Genie III HIV-1/HIV-2 tests, the HIV-1– and HIV-2–specific bands were noted as being weak/uncertainly positive. Among these samples, the Genie III HIV-1/HIV-2 had a higher number of disagreements with INNO-LIA HIV- I/II (35.3% vs. 5.9%, P < 0.01).
During the study period, SD Bioline HIV-1/2 3.0 was used as a discriminatory test in Guinea-Bissau. Among the 219 included patients, 162 patients had initiated ART. Six patients had been misdiagnosed as HIV-1 positive but were testing HIV-1/2 dually seropositive untypable in the INNO-LIA HIV I/II test and 4 of these patients were on a NNRTI-based regimen. Eighteen patients were misdiagnosed as HIV-2 or dually seropositive by SD Bioline but were HIV-1 seropositive according to the INNO-LIA HIV I/II test. Eight of these had initiated ART with a protease inhibitor–based regimen.
In this study, we evaluated the performance of 3 HIV-type discriminatory rapid tests by comparing them with INNO-LIA HIV I/II. Though misclassifying 1 patient as HIV negative, Genie III HIV-1/HIV-2 was the best overall performer with regard to discriminatory capacity and SD Bioline HIV 1/2 3.0 was the worst. Several of the included patients were in a suboptimal ART regimen due to HIV-type misclassification using SD Bioline HIV 1/2 3.0.
Immunocompromised patients may have impaired HIV-1–specific antibody levels.8,9 The current study included patients attending an HIV clinic for routine testing. The study thus reflected real life and included patients with very low to normal CD4 cell counts receiving ART or not. Also, we did not exclude samples that were inconclusive or difficult to interpret as presented elsewhere.10
In Guinea-Bissau, HIV-type discrimination is performed each time a CD4 cell count is measured among monoinfected patients due to the possibility of superinfection with HIV-1 or HIV-2. Thus, 74% of the patients included in this study were receiving ART. A serological test, the INNO-LIA HIV-I/II assay, was thus chosen as gold standard in our study based on availability, experience in the laboratory at Aarhus University Hospital, and previous studies using INNO-LIA HIV-I/II.5,11,12 HIV-1 and HIV-2 RNA measurements were not performed as levels could have been undetectable due to ART, and moreover, HIV-2 RNA levels may be undetectable even in untreated individuals.13 Furthermore, plasma was scarce, and there was insufficient material for the detection of HIV-1 and 2 RNA among most of the 26% untreated patients. CD4 cell count measurements were not available for 41 patients, and the power to detect an effect of low CD4 cell count was consequently weakened.
According to the package inserts of all 3 rapid test, HIV-type sensitivity is 100% and specificity is >99%. For SD Bioline HIV 1/2 3.0, a similarly high sensitivity and specificity was reported in a study from India.14 The test's performance was, however, lower in the current study similar to the findings reported by Chaillet et al,5 which also found that the Genie rapid test (previous version) had the best overall performance. In another study from Ghana, the authors found the Genie II test to be useful for rapid and accurate differentiation of HIV-1 and HIV-2 antibodies.15 However, we evaluated a new version of the Genie rapid test, which does not require cold storage, making it easier to implement in tropical resource-limited settings.
Cross-reactivity between HIV-2–specific antibodies and HIV-1 antigens and HIV-1–specific antibodies and HIV-2 antigens has previously been described.16–18 However, interpretation of test results may also rely on subjective measures: According to the SD Bioline HIV-1/2 3.0 manufacturer specifications, the test should be interpreted as HIV-1 or HIV-2 only if “the color intensity of the test line 1 is darker than one of the test line 2” and “The presence of three lines as control line (C), test line 1 (1) and test line 2 (2) within the result window indicates a positive result for HIV-1 and/or HIV-2.” In the Conakry study, the investigators chose to call all dually reactive tests for HIV-positive untypable. In Bissau, the procedure was to let the trained operator interpret the test as HIV-1, HIV-2, or HIV-1/2 dually positive. Interobserver variation and differences in test result definitions may have influenced the outcome of the studies.
When this study was performed, SD Bioline HIV 1/2 3.0 was used as the discriminatory HIV test in Guinea-Bissau. Several potentially HIV-1/2–infected patients in this study were misdiagnosed with HIV-1 causing them to be initiated with a NNRTI-based ART regimen not effective against HIV-2.3 These patients may experience treatment failure and progressive decline in CD4 cell count despite good adherence. An even higher number of patients were misclassified with either HIV-2 or HIV-1/2 infection instead of HIV-1, depriving them of the standard HIV-1 treatment regimen. According to national guidelines, HIV-2 and HIV-1/2 infections should be treated with a protease inhibitor–based ART regimen. As PIs are used for second-line treatment of HIV-1, the misdiagnosis compromises the treatment options in case of treatment failure. Although Genie III HIV-1/HIV-2 had the overall best performance, it misclassified a substantial number of potentially HIV-1/2–infected patients as being HIV-1 only positive.
We could not find any preceding published test evaluations of SD Bioline HIV 1/2 3.0 from a similar setting that dated back from before the implementation in 2006. According to the Joint United Nations Programme on HIV and AIDS, test combinations should always be evaluated in the context in which they will be used before widespread implementation.19 This study demonstrates the need for that statement to be taken into action.
We evaluated the discriminatory capacity of 3 rapid HIV tests and found the overall best performing test to be Genie III HIV-1/HIV-2. Nevertheless, Genie III HIV-1/HIV-2 misdiagnosed a substantial number (12.2%) of HIV-1/2 seropositive untypable patients (according to the gold standard) as being HIV-1–only positive, and hence, room for improvement is present. Correct discrimination between the HIV-1 and HIV-2 is important to choose an effective ART regimen. With globalization, HIV-2 is not confined to Western Africa and affordable reliable tests are needed. Proper evaluation of these tests before implementation is imperative.
The authors are grateful to the health care personal at the HIV clinic at HNSM for providing medical care and data acquisition of the HIV-infected patients in this study.
1. da Silva ZJ, Oliveira I, Andersen A, et al.. Changes in prevalence and incidence of HIV-1, HIV-2 and dual infections in urban areas of Bissau, Guinea-Bissau: is HIV-2 disappearing? AIDS. 2008;22:1195–1202.
2. Campbell-Yesufu OT, Gandhi RT. Update on human immunodeficiency virus (HIV)-2 infection. Clin Infect Dis. 2011;52:780–787.
3. Witvrouw M, Pannecouque C, Switzer WM, et al.. Susceptibility of HIV-2, SIV and SHIV to various anti-HIV-1 compounds: implications for treatment and postexposure prophylaxis. Antivir Ther. 2004;9:57–65.
4. Peterson K, Jallow S, Rowland-Jones SL, et al.. Antiretroviral therapy for HIV-2 infection: recommendations for management in low-resource settings. AIDS Res Treat. 2011;2011:463704.
5. Chaillet P, Tayler-Smith K, Zachariah R, et al.. Evaluation of four rapid tests for diagnosis and differentiation of HIV-1 and HIV-2 infections in Guinea-Conakry, West Africa. Trans R Soc Trop Med Hyg. 2010;104:571–576.
7. Oliveira I, Andersen A, Furtado A, et al.. Assessment of simple risk markers for early mortality among HIV-infected patients in Guinea-Bissau: a cohort study. BMJ Open. 2012;2:e001587.
8. Ortiz de Lejarazu R, Soriano V, Eiros JM, et al.. HIV-1 infection in persistently HIV-1-seronegative individuals: more reasons for HIV RNA screening. Clin Infect Dis. 2008;46:785.
9. Toro C. Absence of HIV-1 antibody response in HIV patients: what is the foe, the virus or the host? AIDS Rev. 2007;9:188–189.
10. Vijayakumar TS, David S, Selvaraj K, et al.. Performance of a rapid immunochromatographic screening test for detection of antibodies to a human immunodeficiency virus type 1 (HIV-1) and HIV-2: experience at a tertiary care hospital in South India. J Clin Microbiol. 2005;43:4194–4196.
11. Maueia C, Costa D, Meggi B, et al.. Frequency of human immunodeficiency virus type-2 in HIV infected patients in Maputo city, Mozambique. Virol J. 2011;8:408.
12. Nkengasong JN, Maurice C, Koblavi S, et al.. Field evaluation of a combination of monospecific enzyme-linked immunosorbent assays for type-specific diagnosis of human immunodeficiency virus type 1 (HIV-1) and HIV-2 infections in HIV-seropositive persons in Abidjan, Ivory Coast. J Clin Microbiol. 1998;36:123–127.
13. Popper SJ, Sarr AD, Travers KU, et al.. Lower human immunodeficiency virus (HIV) type 2 viral load reflects the difference in pathogenecity of HIV-1 and HIV-2. J Infect Dis. 1999;180:1116–1121.
14. Kannangai R, Prabu K, Vincent AA, et al.. Performance evaluation of four different kits available in the Indian market, for the rapid detection of HIV antibody. Indian J Med Microbiol. 2003;21:193–195.
15. Sagoe KW, Agyei AA, Lartey M, et al.. Diagnosis of dual human immunodeficiency virus types 1 and 2 infections in a resource-limited setting. East Afr Med J. 2009;86:417–421.
16. Weiss RA, Clapham PR, Weber JN. HIV-2 antisera cross-neutralize HIV-1. AIDS. 1988;2:95–100.
17. Ozkaya Sahim G, Holmgren B, da Silva Z, et al.. Potent intratype neutralizing activity distinguishes human immunodeficiency virus type 2 (HIV-2) from HIV-1. J Virol. 2012;86:961–971.
18. Sagoe K, Mingle J, Affram R, et al.. Virological characterization of dual HIV-1/HIV-2 seropositivity and infections in Southern Ghana. Ghana Med J. 2008;42:16–21.