Evaluation of HIV-1 rapid diagnostic tests in the context of viral genetic diversity in Libreville (Gabon)

¹Laboratoire de Biologie Moléculaire et Cellulaire (LABMC), Université des Sciences et Techniques de Masuku, BP 943, Franceville, Gabon, ²Centre Hospitalier Régional Georges RAWIRI (CHRGR), Lambaréné, Gabon, ³Unité Mixte de Recherches sur le VIH et les Maladies Infectieuses Associées (UMR, MIA), Hôpital d´Instruction des Armées Omar BONGO ONDIMBA, Libreville, Gabon, ⁴Centre Interdisciplinaire de Recherches Médicales de Franceville (CIRMF), Gabon

^&Corresponding author
Cyrille Bisseye, Laboratoire de Biologie Moléculaire et Cellulaire (LABMC), Université des Sciences et Techniques de Masuku, Franceville, Gabon

Introduction    

Human immunodeficiency virus (HIV) infection remains a public health problem in Gabon. Its seroprevalence was estimated at 4.1% in 2012 during the last demographic and health survey carried out in the country [1]. In Gabon, in order to promote the rapid treatment of HIV-seropositive patients and to reduce the transmission of HIV, the diagnosis of this virus is based essentially on Rapid Diagnostic Tests (RDTs) in the absence of 4th generation immunoenzymatic tests and nucleic acid tests (NAT) by PCR in health facilities because of their prohibitive cost. However, in order to be eligible for an HIV screening algorithm according to the recommendations of the World Health Organization (WHO), the RDTs must have a sensitivity ≥ 99% and a specificity ≥ 98% [2].

In Gabon, the national HIV screening algorithm offers three RDTs: Alere Combo Ag/Ab 4th generation HIV-1/2, Alere Determine HIV-1/2 and SD HIV-1/2 3.0 (Multi). HIV infection in this country is characterized by a high viral genetic diversity which could lead to the absence of detection of certain HIV-1 subtypes for RDTs. Indeed, HIV-1 is divided into four groups: M (Major), O (Outlier), N (non-M, non-O), and P [3]. Among the HIV-1 groups, group M is responsible for the pandemic of HIV infection [4]. This M group includes ten subtypes (A-D, F-H, and J-L) with several Circulating Recombinant Forms (Yamaguchi et al., 2020) (CRFs) [5]. In Gabon, the HIV group M is the main circulating strain [6] with several subtypes and recombinant forms identified [7-9] while group O is poorly represented with a prevalence of 0.9% [10]. The available RDTs are often introduced on the local market without prior evaluation and could have the disadvantage of not being adapted to the viral subtypes circulating in the country. An evaluation before their use on the local market and the regular evaluation of their effectiveness is therefore necessary. Some HIV RDTs have already been evaluated in previous studies. The reported sensitivities and specificities were 99.60% and 100% for INSTI HIV1/HIV2 test [11] and 90.9% and 100% for Alere Determine HIV1/HIV2 test [12]. The diversity of HIV-1 in Gabon could be constantly changing due to the geographical proximity with Cameroon, a country known to harbor the majority of HIV-1 subtypes [13,14]. This study was carried out to evaluate three RDTs used in HIV screening in Libreville, Gabon in a context of viral genetic diversity.

Methods     

Study design and settings: this was a prospective cross-sectional study that took place from August to November 2020 at the Unité Mixte de Recherches sur le VIH et les Maladies Infectieuses Associées (UMR, MIA) and the Centre Interdisciplinaire de Recherches Médicales de Franceville (CIRMF), located at the Hôpital des Instructions des Armées Omar Bongo Ondimba (HIAOBO), based in Libreville, the capital of Gabon.

Participants: the study concerned samples from HIV-positive patients in consultation for a viral load and the identification of antiretroviral resistance mutations; and on sera from patients with negative or indeterminate HIV serology. A total of 60 group-M HIV-1+ plasmas were tested. The genotypic proportions of HIV-1 were respectively 46.7% (28/60) for CRF02-AG; 18.3% (11/60) for A; 11.7% (7/60) for G; 5% (3/60) for A/G; 5% (3/60) for CRF06-Cpx; 3.3% (2/60) for CRF04-Cpx; 3.3% (2/60) for H; 1.7% (1/60) for B; 1.7% (1/60) for D; 1.7% (1/60) for A/J and 1.7% (1/60) for CRF18-Cpx. In addition, 51 negative sera and 5 with undetermined HIV serology were also tested.

HIV diagnostic and confirmation: the Alere Combo Ag/Ab 4th generation HIV1/2 (Alere Medical., Ltd Japan), Alere Determine HIV1/2 (Alere Medical., Ltd Japan) and Vikia HIV1/2 (Bio-Mérieux, France) RDTs were evaluated in this study. These tests based on the principle of immunochromatography for the qualitative detection of anti-HIV-1/2 antibodies were used according to the manufacturers´ recommendations. The performances stated by the manufacturers varied from 99.4% to 100% for the sensitivities and from 99.4% to 99.9% for the specificities. According to the manufacturer's instructions, the 5 undetermined patient sera were tested for HIV-1 group O using the Mini Vidas (Biomérieux, France) and not detectable by the GENERIC HIV Charge Viral kit (Biocentric, Bandol-France) and all the samples were tested by qRT-PCR, which served as a reference test for the validity of the RDTs.

Viral RNA was extracted with the Arrow Viral RNA semi-automated extractor (DiaSorin, Ireland Ltd). Plasma viral load was quantified by real-time quantitative PCR. Briefly, 250 μL of serum was added to 5 microliters of proteinase K and 2 microliters of internal control for a final elution volume of 50 μL. The reverse transcription reaction was assessed using the GENERIC HIV Charge Viral Kit (Biocentric, Bandol, France) in a reaction volume of 20 μL (10 μL RNA+10 μl mix) with the Fluorocycler 96 Real-time PCR machine (Hain Lifescience) according to the following amplification program: 50°C for 10 minutes for reverse transcription; 95°C for 5 minutes of denaturation; followed by 50 cycles including denaturation at 95°C for 15 seconds and annealing at 60°C for 1 min. The HIV detection algorithm used during the study is shown in Figure 1.

Statistical analysis: data were entered using Microsoft Excel version 2016. R software (version R x 64.3.6.2) was used for the sensitivities, specificities, positive predictive values (PPV) and negative predictive values (NPV ) determination, with 95% confidence intervals. Cohen's Kappa coefficient (http://graphpad.com/quickcalcs/kappa/) was calculated to assess the level of agreement between the different RDTs for the detection of HIV.

Ethical consideration: this study was approved by the HIAOBO Ethical Committee. All adults ≥18 years of age gave informed consent before participating in the study.

Results     

Sociodemographic characteristics of patients: a total of 116 patients were included in this study. Of these samples, 51.7% (60/116) were HIV-positive while 44% (51/116) were HIV-negative. Five (5) patients with indeterminate HIV serology were also included. Women were mostly more represented than men (66.4% vs 33.6%) both in HIV-positive patients (37.9% vs 13.8%) and HIV-negative patients (26.7% vs. 17.3%). The patients´ age varied from 18 to 72 years with a mean of 40 years ± 12.

Performance of RDTs in the diagnosis of HIV-1: the 60 HIV-1 positive plasmas belonged to group M with 11 distinct subtypes. All viral subtypes were detected by the three RDTs (Table 1). Of the 60 HIV+ and 56 HIV- samples detected by qRT-PCR, the Determine HIV-1/2 and Combo HIV1/2 RDTs found 60 HIV+ and 54 HIV- samples, respectively. Only 2 negative samples by qRT-PCR were positive for both RDTs. The Vikia HIV1/2 test detected all 60 HIV+ samples while only 53/56 HIV- samples were determined (Table 2).

The sensitivity of the 3 RDTs compared to quantitative RT-PCR for the detection of HIV-1 M subtypes was 100% (95% CI = 92.5-99.9). The specificity of the Combo Ag/Ab 4^e G and Determine tests was 96.4% (95% CI = 86.6-99.3) while their positive and negative predictive values were 96.7% (95% CI = 87.8-99.4) and 100% (95% CI = 91.7-100), respectively. The Vikia test showed a specificity of 94.6% (95% CI = 84.1-98.6), its positive and negative predictive values were 95.2% (95% CI = 85.8-98.7) and 100% (95% CI = 91.6-100), respectively (Table 2).

We also compared the levels of agreement between the different RDTs used for HIV detection. These agreements were excellent for HIV detection between Combo and Determine on one hand and between Combo/Vikia and Vikia/Determine on the other hand with kappa values of 0.931 (95% CI = 0.865-0.977) and 0.948 (95% CI = 0.891-1.00), respectively (Table 3).

Discussion     

The wide use of RDTs in HIV screening in Gabon and the high viral genetic diversity observed implies regular evaluation of these tests performance for their rational use in the national screening algorithm. In this study, three HIV RDTs used in Gabon were evaluated on a panel of confirmed HIV-positive samples, negative and undetermined samples.

The sensitivities of the three RDTs for HIV-1 M subtypes detection were 100%. These sensitivities are similar to those reported in an earlier study carried out in Cameroon [15]. Our results are also similar to those reported in France for the Vikia and Alere Combo tests [16]. However, they are different from the results obtained for the Alere Determine test in Congo [17] and Gabon [12] with the sensitivities of 57.1% and 90.9%, respectively. The difference between the sensitivities observed could be partly explained by the type of samples tested. Indeed, in the present study, we tested samples from confirmed HIV-positive patients, while in the two studies carried out in Congo and Gabon, they were blood donors who may have been diagnosed during the seroconversion phase. With regard to the study performed in Thailand, the evaluation of the Alere Combo test has been done with samples from patients in the acute phase of the infection, a period during which RDTs show reduced sensitivities [18]. This difference could also be attributed to the genetic variability of HIV [15]. RDTs could present a risk both in routine diagnosis and for transfusion safety in geographical areas where the qualification of blood bags is done using them, such as sub-Saharan Africa, including Gabon. Indeed, a previous study had reported a high HIV prevalence of 3.1% among blood donors in rural areas [19]. This could constitute a threat to the transfusion safety of recipients in these regions. Indeed, financial and technical resources are necessary to set up a strict algorithm including tests capable of circumventing the HIV genetic variability. Furthermore, molecular amplification tests by PCR which could minimize the transmission of this virus by blood transfusion in Gabon are needed.

The RDTs´ specificities ranged from 94.6% for Vikia test to 96.4% for the Alere Determine and Combo tests. These specificities were closer to those found in Cameroon, which ranged from 95.8% to 97.1% for the Alere Determine and Alere Combo tests [15]. However, they were lower than those announced by the manufacturers which ranged from 99.4% to 99.9%. RDTs specificities obtained in this study were also lower than those reported in France for Alere Combo and Vikia tests [16]; but higher than the 94.0% specificity found in Brazil for Alere Determine test [20]. The false-positive results observed with RDTs could be explained by serological cross-reactions due to exposure to various endemic infections (malaria, schistosomiasis, human African trypanosomiasis, etc.), technical problems such as contamination, poor storage of RDT kits, non-compliance with the manufacturer´s protocol and misinterpretation of results which can lead to misdiagnosis [21].

This study suggests that a small proportion of patients diagnosed according to the national HIV testing algorithm could have been inappropriately admitted to ARV treatment with consequences such as psychological distress, and long-term liver and kidney damage due to ARVs toxicity in the latter. To ensure the reliability of the results obtained with the RDTs, it would be important not only to test a second time the patients at least three weeks after the first serology but also to confirm their results either by western blot or PCR before putting them under ARVs, contrary to the WHO recommendation "test-and-treat" consisting of putting all HIV-positive patients under treatment.

Limitations of the study: our study has some limitations such as the small size of the sample studied. It concerned only one specific region of Gabon. However, HIV subtypes may vary across the country. We were only able to evaluate three RDTs among many others used in HIV screening in Gabon.

Conclusion     

Despite the high viral genetic diversity of HIV, the sensitivities of the 3 RDTs (100%) obtained in this study complied with WHO recommendations, while their specificities ranged from 94.6% to 96.4%. These specificities would not allow the 3 RDTs to be included in an HIV screening algorithm according to WHO recommendations.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Study design: Cyrille Bisseye and Soulemane Parkouda. Sampling: Yann Vital Sima Biyang, Berthold Bivigou Mboumba. Samples processing: Yann Vital Sima Biyang, Berthe Amélie Iroungou and Berthold Bivigou-Mboumba. Analysis and data interpretation: Cyrille Bisseye, Yann Vital Sima Biyang and Augustin Mouinga Ondeme. Manuscript drafting and corrections: Yann Vital Sima Biyang, Cyrille Bisseye, Berthe Amélie Iroungou, Soulemane Parkouda, Augustin Mouinga Ondeme and Berthold Bivigou-Mboumba. All authors read and approved the manuscript.

Acknowledgments     

The authors wish to thank the many collaborators involved in the collection of the samples analyzed in this report, in particular Mrs Aurore Bouassa, Captains Bouanga and Hugues Ossindji are thanked for their technical support. The staff of all medical centers participating in this study is thanked for their participation.

Tables and figure     

Table 1: RDTs´ reactivities according to HIV-1/M subtypes

Table 2: comparison of RDTs and qRT-PCR for the detection of HIV

Table 3: comparison of RDTs for the detection of HIV

Figure 1: algorithm of HIV testing in the study

References