Problems encountered in conventional HIV 1/2 Algorithms: lack of necessity for immunoblot assays to confirm repeated ELISA reactive results

Background The use of conventional (serologically based) HIV 1/2 diagnostic algorithms has become controversial in recent years. Objectives Sera from patients who underwent verification tests were evaluated because repeated ELISA-reactive results demonstrated a HIV1+HIV2 positive band pattern. Methods The line immunoassay (LIA) test was used for repeated HIV enzyme immunoassays (EIA)-reactive sera in patients at three centers. The Bio-Rad Geenius™ HIV 1/2 and the HIV-1 RNA tests were used. HIV-1 and RNA HIV-2 were investigated using PCR. Results LIA was used to evaluate 3,224 out of 10,591 samples with repeated ELISA reactivity (30%). We found that 32 (1%) of the sera, along with HIV1 bands and HIV2 gp36 bands, were positive. Only 28 of the 32 verified serum samples with gp36 bands were repeated, and no gp36 band positivity was detected using the Bio-Rad Geenius™ HIV-1/2 confirmatory assay in these serum samples. The HIV-2 proviral DNAs were also negative. Therefore, we excluded the possibility of HIV1+2 co-infection. All samples from the 32 patients were positive for HIV-1 RNA. Conclusion Our findings highlight the need to exclude confirmatory tests like the LIA test from the current diagnostic HIV algorithm and replace it with rapid HIV-1 and HIV-2 confirmatory immunochromotographic tests.


Introduction
Acquired immune deficiency syndrome (AIDS), caused by the human immunodeficiency virus (HIV), has been a globally important health problem for the last 30-35 years 1 . The incidence of HIV cases is constant in developed countries but it is increasing in resource-poor countries. Because of problems (such as indeterminate western blot results, HIV 1/2 cross-reactions) that are experienced when diagnosing HIV-1 infections in vitro, new research and approaches for diagnostic algorithms are needed. Often in West Africa, and to a lesser extent in the United States, problems with diagnostic approaches for HIV-2 infections are more serious 2,3,4,5,6 . From 1989 to the present, the diagnostic algorithm for HIV-1 infections has been primarily based on the repeating reactivity of anti-HIV Ab/Ag tests and the reactivity of Western blot (WB) and line immunoassay (LIA) tests, according to Centers for Disease Control and Prevention (CDC) criteria 7 . However, a reliable diagnostic HIV algorithm that can be used as a reference, has not yet been created despite the absence of an enzyme immunoassays (EIA) with a high sensitivity and specificity for in-vitro diagnoses of HIV-2 infections together with the insufficiency in immunoblot and molecular methods 8 . Because of the above-mentioned problems, the conventional diagnostic HIV algorithm needs to be updated according to the CDC recommendations on problems associated with the in vitro diagnosis of HIV-1 and HIV-2 infections using WB/LIA tests based on immunoblotting. The main characteristics of current molecular tests and immunoassays for the detection of HIV infection are shown in Table 1 8 .  11,12,13 . Here, we re-evaluated the national in-vitro diagnostic HIV algorithm that relies on a conventional immunoblot-based confirmatory test (LIA). We also evaluated the cases of HIV patients with indeterminate and dual reactivity patterns detected by the conventional (serologically based) diagnostic HIV-1/2 algorithm. Our work highlights the necessity of using this new diagnostic HIV algorithm, which has been approved by organizations such as the CDC and CLSI.

Study area and Groups
This cross-sectional multicenter study was conducted between January 2014 and October 2015. The centers involved in this study were as follows: We focused on patients who visited these three centers for clinical diagnoses or blood donor purposes. The patient cases selected for inclusion in this study were chosen according to CDC criteria 7 . Briefly, our study algorithm included the following, initially, serum samples repeatedly positive by EIA (anti-HIV-1 test ) were studied in WB/ LIA confirmatory assays. If both tests (EIA+WB/LIA) were positive, the patient was noted as being truly HIV-1 positive. If only one of the proteins (p24, gp41, and gp120/160) was positive by WB/LIA , the patient was regarded as exhibiting an indeterminate HIV-1 pattern 14 . gp36 and gp105 positivity were used to determine HIV-2 positivity. Only 32 samples showed LIA+HIV1+ HIV-2 with specific gp36 band suggesting a positive HIV-2. Nested PCR was used to detect HIV-2 proviral DNA but none was positive for HIV-2. No gp36 positivity was also detected in these 28 out of 32 samples using the Bio-Rad Geenius™ HIV-1/2 confirmatory assay. All of the 28 study samples were positive for HIV-1 RNA. The baseline characteristics of 28 patients are shown in Table 1.
The sex (male/female) distribution of our study was 21 (75%)/7 (25%). The mean age of the patients was 35.1 years (range, 18-61 years). All participants signed a written informed consent form. We used the LIA method as part of the immunoblotting method to confirm recurrent reactive HIV Ab/Ag (Inno-LIA HIV-1/2 score; Innogenetics, Belgium). The samples were evaluated according to the manufacturers' recommendations.

Immunological (Serological) methods
We considered the CDC criteria as a basis for evaluation 7 . The Immunochromatographic Assay for Differentiating HIV-1 and HIV-2 (the Bio-Rad Geenius™ HIV 1/2 confirmatory assay (Bio-Rad Laboratories, Marnesla-Coquette, France)), which is intended to confirm and differentiate between HIV-1 and HIV-2, is a single-use immunochromatographic test that uses immobilized HIV-1 (p31, gp160, p24, and gp41) and HIV-2 (gp36 and gp140) antigens to detect antibodies to HIV-1 and HIV-2 in serum, plasma, or whole blood. In our investigation, the band patterns were read manually by two experts, and the interpretation criteria were as follows: (a) negative was

Molecular Tests
In We investigated HIV-2 proviral DNA using nested PCR with H2L100 5′-GCTGGCAGATTGAGCCCTG-3′ and H2L200 5′-AAGGGTCCTAACAGACCAGGG-3′ primers for the first round and H2L101 5′-CAGCACTAG-CAGGTAGAGCCTGGG-3′ and H2L201 5′-GGCGG-CGACTAGGAGAGATGG-3′ primers for the second round. All PCR amplifications were performed in a total volume of 50 µL and were carried out on a GeneAmp® 9700 Thermal Cycler (Applied Biosystems Foster City, CA). PCR conditions with first round primers were 2 min at 95°C followed by 35 cycles for 30 sec at 94°C, 30 sec at 45°C, and 1 min at 72°C. The nested PCR with second round primers were performed using 1 µL of the amplified DNA from the first PCR in a fresh 25 µL reaction buffer under the following conditions: 35 cycles of 30 sec at 94°C, 30 sec at 55°C, and 1 min at 72°C 15 .

Statistical methods
This study was not intended to be a comparison of test performance in a particular algorithm. Rather, algorithm strategies were evaluated using current test combinations to assess the relative advantages and magnitude of the differences between algorithm strategies. Simple median calculation data for medians of some baseline characteristics in the study population were analyzed using Microsoft Excel (Microsoft Corporation, USA) and SPSS 20.0 (IBM, SPPS Inc., USA).

Results
Baseline characteristics of the study population with a LIA+HIV1+ HIV-2-specific gp36 band positivity are shown in Table 2.
African Health Sciences Vol 18 Issue 2, June, 2018   Table 3. Twenty-eight samples with HIV-2-specific gp36 band positivity out of 32 samples included in this study were re-analyzed, and no gp36 positivity was detected using the Bio-Rad Geenius™ HIV-1/2 confirmatory assay. Ten samples had a HIV-1 viral load of more than 200,000 copies/ml. All of the 28 samples had detectable p24,gp41, gp120, and gp36 bands but were negative for the gp105 band. While seven of the samples were negative for the p17 band, 10 of them were negative for the gp31 band. The EIA screening, Inno-Lia HIV 1/2, HIV-1 viral load, and CD4 + T cell count from 28 samples that were positive for a gp36 band+HIV-1 are shown in Table 4. However, all of the 28 study samples were positive for HIV-1 RNA, but we only analyzed 28 out of 32 patients in this group because the remaining four patients were foreigners who returned to their native countries. The study algorithm results from the present study are shown in Fig. 1.  Fig. 1. The study algorithm used in the present study We only contacted 28 out of 32 patients in this group because the remaining four patients were foreigners and had returned to their native countries.

Discussion
An early and accurate diagnosis of sexually transmitted infections (e.g., HIV) is very important in countries such as Turkey, where there is much human movement from neighboring countries. Delayed or misdiagnosis in the acute phase of HIV infections may have important consequences for the spread of HIV infections. Currently, using fourth-generation EIAs and confirmatory tests based on immunoblotting (WB/LIA) result in early diagnoses and increased sensitivity/specificity despite the HIV infection window period in Turkey. However, the prevalence of patients with dual HIV-1 and HIV-2 patterns has increased in response to the indeterminate test results because of the lack of differentiation between HIV-1 and HIV-2 infections in the current diagnostic HIV algorithms in Turkey. There were 10,215 HIV carriers and 1,274 officially announced HIV cases in Turkey by the end of 2015; these are only a small portion of the data on the number of patients with HIV. When the last two years' worth of HIV infection data were evaluated from Istanbul, the largest city and the one that receives the most national/international immigrants in Turkey, we found that indeterminate results constituted 271 out of 10,591 cases (2.5%) using a LIA test. Additionally, a dual reactivity pattern (both HIV-1 and HIV-2 depending on gp36) was also noted in 32 (1%) of 3,224 patients with positive test results (30%). In this study, all HIV-2 nucleic acid amplification test and Geen-ius™ HIV-1/2 differentiation test results were negative, and all HIV-1 nucleic acid amplification test (NAAT) results were positive in 28 out of 32 patients. The most important reasons behind the HIV-2 gp36 false reactivity pattern (1% of our patients) may be cross-reactions between envelope (gp36/gp41, gp120 V3 loop) and trans-membrane and surface proteins of HIV-1 and HIV-2. Additional motifs (such as CGC, which can increase the antigen density via polymerization of cystein residues) may affect the sensitivity of these tests. Manocha et al. 16 reported that the gp36 peptide containing the CGG tag detected HIV-2 in serum samples with 100% sensitivity and 98% specificity; the sensitivity and specificity of the gp36 plain peptide were reduced to 98% and 90%. However, there have been several reports of false positive result rates in different populations. For example, Amor et al. 17 reported a positivity rate of 6.2% for the gp36 band in patients infected with HIV-1; McKellar et al. 18 reported higher rates in elite controllers. Elite controllers are reported to be used for individuals who are able to suppress viral replication to undetectable levels for extended periods of time without the use of anti-retroviral therapy 19 . Nucleotide sequences of highly conserved gag and pol genes of HIV-1 and HIV-2 exhibit 60% homology and other viral genes and long terminal repeats (LTRs) exhibit 30-40% homology. Even though these 32 patients with false-positive reactivity according to immunoblotting tests suggest a dual infection pattern, these patients were diagnosed with true HIV-1 infections. These false positive patterns may cause delays in diagnosis of these patients.
However, we were unable to detect a true HIV-2 infection in our study (only two true HIV-2 cases have been reported in Turkey to date), and HIV-2 infections have been detected epidemiologically only in individuals in West Africa and in Western and Asian countries. Our results support diagnostic insufficiency of the conventional diagnostic algorithm. The conventional diagnostic HIV algorithm is insufficient for an HIV-2 diagnosis by immunoblotting tests and often results in a negative impact on surveillance and treatment of patients with indeterminate results 20,21,22 . A 46-85% false positive rate in HIV-1 LIA test results was noted in patients with true HIV-2 infections (in 2010 and 2011 studies). Therefore, diagnosis and treatment were delayed 23,24,25 . The Geenius™ HIV-1/2 confirmatory assay, which was validated and recommended for the proposed new diagnostic HIV algorithm by CDC and CLSI and approved by the FDA in 2014 was used for 28 out of 32 patients who exhibited a dual HIV pattern in our study. This immunochromotographic technique based on the Geenius™ HIV-1/2 test differentiates between HIV-1 and HIV-2 in 100% of patients; these results were confirmed by HIV-1 and HIV-2 NAT tests. Malloch et al. 11 studied 128 HIV-1 and 53 HIV-2 serum samples with this test and found that the sensitivity, specificity, and kappa diagnostic performances of the Geenius™ HIV-1/2 confirmatory assay were 100%, 96.3%, and 96%, respectively; the differentiation rate for true HIV-1 and true HIV-2 serum samples was 99.2% and 98.1%, respectively. Montesinos et al. 25 noted that the Geenius™ HIV1/2 confirmatory assay that was applied to 160 serum samples had a sensitivity and specificity of 100%. Consistent with the studies noted above, Moon et al. 26 demonstrated a 95.3% sensitivity and a 100% specificity for the Geenius™ HIV 1/2 confirmatory assay in their study performed using the ARCHITECT HIV Ag/ Ab Combo assay, the Geenius™ HIV1/2 confirmatory assay, and HIV-nucleic acid amplification tests. Although we studied a small number of serum samples, our results and other international studies suggest that the Geen-ius™ HIV-1/2 confirmatory test is a safe and reliable alternative to conventional immunoblotting tests for HIV 1/2 diagnostics. The Geenius™ HIV-1/2 confirmatory test may also yield important improvements in the quality management of HIV algorithms.
A new diagnostic HIV algorithm was developed based on the CLSI M53-A HIV guidelines, including HIV Ab/ Ag tests and rapid HIV-1 and HIV-2 discrimination tests based on chromatographic techniques and nucleic acid amplification tests and excluding immunoblotting-based tests (WB/LIA), and it accurately differentiated between HIV-1 and HIV-2 infections in patients with a dual HIV pattern. This new algorithm consequently decreased the number of final indeterminate HIV results. Our findings highlight the need to exclude confirmatory tests, such as the LIA test from the former diagnostic HIV algorithm, and replace them with rapid HIV-1 and HIV-2 confirmatory immunochromotographic tests. If this change is not implemented, new diagnostic strategies for HIV-2 infections may be necessary because of immigration, even beyond endemic regions. Thus, peptide-based immunoassays detecting additional bands (e.g., sgp105, sgp140) and optimized HIV-2 nucleic acid amplification tests may be helpful.
Hayderabad, 2016. We also thank Dr. Hayriye Kırkoyun Uysal (Istanbul Medical Faculty, Turkey), Ali Abdelkerem (Cerrahpasa Medical Faculty, Turkey), and Esad Bonabi (Aydin University, Turkey) for their technical assistance related to the poster preparation for the poster presentation in the 17 th International Congress On Infectious Diseases (ICID) Hayderabad, 2016. English language editing was performed by American Manuscript Editors.

Ethics
This study was approved by Clinical Research Ethics Board of Istanbul University, Cerrahpasa Faculty of Medicine (No:83045809/604, Date: 01.07.2014).

Declaration of conflicting interests
The authors declare that there are no conflicts of interest