Issues in Biological Sciences and Pharmaceutical Research Vol.9(3),pp.

86-92, October 2021

Available online at https://www.journalissues.org/IBSPR/
https://doi.org/10.15739/ibspr.21.009
Copyright © 2021 Author(s) retain the copyright of this article ISSN 2350-1588

Original Research Article

Comparison of 2 HIV1 RNA VL measurement commercial

kits in Cote d'Ivoire
Received 26 July, 2021 Revised 5 September, 2021 Accepted 15 September, 2021 Published 10 October, 2021

Kone-Dakouri Yekayo1*, The goal of this study was to compare the two operational platforms
Kone-Kone Fatoumata1,2, (Biocentric and Roche CAP/CTM) in the country to assess their degrees of
Aka Tano Cyrielle2, concordance in order to avoid interruptions in the virological follow-up of
people living with HIV (PLHIV) in Côte d'Ivoire. HIV1 RNA viral load (VL) of
Yapo Vincent De Paul2, Toni
182 plasmas were measured by means of the 2 platforms in the CeDReS
Thomas D. Aquin2, laboratory. The comparison was made by concordance, Bland-Altman (BA)
Ahiboh Hugues1, 2, and the appreciation of clinical differences between the measured VL. The
Blidieu Juste2, proportions of suppressed viral loads (76% and 75%) were similar on the 2
Yayo Sagou Eric1, platforms. The mean VLs were 3.43 and 3.66 Log10 /mL respectively on
Edjeme Ake Angele1, CAP/CTM and Biocentric. According to BA, the bias was 0.23 Log 10 in favour
Attoungbre-Hauhouot Marie of Biocentric and the correlation coefficient of agreement (0.922) was poor.
Laure1 , The Kappa coefficients of 0.846 at the detection thresholds and 0.878 at the
Menan Hervé2 , viral suppression threshold indicated a near perfect level of agreement.
and Clinically significant differences (> |0.6|) were observed between clinically
equivalent VL, with no influence on management. Biocentric and CAP/CTM
Monnet Dagui1
had good agreement but the power of agreement was low. Clinical
1Department of Biochemistry,
differences were not significant, however by interchanging these platforms,
Molecular Biology, Reproduction communication between biologists and clinicians is useful.
Biology and Medical Pathologies,
Unit of Education and Research of
Keywords: HIV1 RNA VL, biocentric, Roche CAP/CTM, concordance, clinical
Pharmaceutical and Biological
Sciences, Felix Houphouet Boigny
difference
University, Abidjan, Ivory Coast
2Diagnostic and Research Center on

AIDS and Other Infectious Diseases

(CeDReS), Treichville, University
Hospital, Abidjan, Ivory Coast.

*Corresponding Author
Email: [email protected]
Tel.:+225 0707220143

INTRODUCTION

The Human Immunodeficiency Virus (HIV) pandemic is a prevalence of 2.9% in the adult population (CIPHIA, 2017).
major public health issue worldwide, particularly in Africa. The major pathophysiological event during this infection
Globally, 37.7 million [30.2- 45.1 million] people were is viral replication, and the resulting virions destroy the
living with HIV in 2020. (UNAIDS, 2021). West and Central cells of the immune system, mainly TCD4 lymphocytes.
Africa is the third most affected region in the world Currently, molecular biology quantification of plasma viral
(UNAIDS, 2021), with Côte d'Ivoire, the second most RNA, also known as viral load (VL) has emerged as the gold
affected West African country (after Nigeria), having a standard test for assessing the level of viral replication
 Issues Biol. Sci. Pharma. Res. 87

Table 1. Main characteristics of the 2 platforms (Generic HIV Viral Load, 2020; Roche Molecular Diagnostics, 2021)

Generic Viral Load (Biocentric) CAP/CTM v2.0 ®

(Roche Molecular Systems)
Target region LTRs (Long Terminal Repeat) LTRs (Long Terminal Repeat) + GAG gene (gene
coding for HIV surface proteins)
Detection threshold 390* copies/mL (2.59 Log10 copies/mL) 20 copies/mL (1.3 Log10 copies/mL)
Testing 250 µL 1000 µL
Specificity HIV-1 Group M, subtypes A to H HIV-1 Group M subtypes A to H
Linearity 165 to 5,106 copies/mL (2.21 to 6.69 Log10 copies/mL) 48 to 10,106 copies/mL (1.68 to 7 Log10
copies/mL)
Accuracy Standard deviation at 5.3 Log10 copies/ml = 0.22 Log10 ; Standard deviation at 6.333 Log10 copies/ml =
Standard deviation at 2.9 Log10 copies/ml = 0.17 Log10 0.10 Log10 ;
Standard deviation at 2 Log10 copies/ml =
0.10Log10
Duration 3 hours (extraction + amplification) for 72 samples 8 hours (extraction + amplification) for 63
samples
Type of sample Plasma collected on anticoagulants (EDTA or sodium citrate) Plasma collected on anticoagulants (EDTA)
*The detection limit applied in our laboratory is 100 copies/mL (2 Log10 copies/mL)

(WHO, 2016). Therapy involves controlling viral replication Diagnostics and Research on AIDS and other Infectious
with antiretroviral (ART) molecules to promote immune Diseases (CeDReS). These were the COBAS®
restoration. Thus, the main objective of treatment is to AmpliPrep/COBAS® TaqMan 48 (CAP/CTM) HIV-1 Test,
make plasma VL undetectable. However, VL quantification Version 2.0; Roche Molecular Diagnostics) and the
is still not widely available in Africa, mainly because of its Biocentric® (Generic HIV Viral Load) platforms.
cost and the lack of laboratory equipment, especially in
rural areas. In order to make the VL accessible routinely, Technique and reagents
Côte d'Ivoire has benefited from the support of various
donors who have enabled the acquisition of 4 types of VL The main characteristics of these 2 platforms given by the
measurement platforms, namely the Biocentric open manufacturers are presented in Table 1.
platform and the closed platforms ROCHE COBAS In order to improve the monitoring of PLHIV, we
Taqman/COBAS AmpliPrep (CAP/CTM), NucliSens and evaluated and validated a lower HIV1 RNA viral load
Abbott m2000. Among them, only 2 types are functional: detection threshold for the Biocentric platform, which is 2
Biocentric (Generic HIV Viral Load) and Roche (Molecular Log10 copies/mL. For this purpose, we tested successive
Systems: COBAS® TaqMan® HIV-1 Test). Since VL dilutions of the last point of the Biocentric Viral Load (E3)
measurement is essential for verifying the achievement of kit standard range and retained the viral load of the
the 3rd objective 90 of the global objective of HIV dilution (E/5) as the threshold value of the assay. As the E3
elimination (VL suppressed in 90% of people living with point has an average VL of (2.7 Log10 copies/mL), its 1/5
HIV (PLHIV) under ART), the supply of VL should not be dilution has an average VL of (2 Log10 copies/mL).
interrupted. Thus, in case of unavailability of one of the
platforms, particularly due to stockout of reagents or METHODS
equipment maintenance parts, common in the country
during this period of the COVID-19 pandemic, these Biological specimens and sample selection
platforms should be interchangeable in order to guarantee
the continuity of virological monitoring of PLHIV. The aim We worked on plasma from venous blood samples collected
of this study was to compare these two platforms to verify from EDTA tubes. Over a period of 3 months, 2 aliquots of
the concordance of the VLs they measure and to assess the plasma were taken and stored at -80°C for each sample
clinical relevance of the differences between these VLs. received at CeDReS for the measurement of HIV1 RNA VL as
part of the routine virological monitoring of PLHIV under
ARV treatment.
MATERIALS AND METHODS The method comparison was performed on 184 samples
from the 419 collected.
MATERIALS The selection was made in such a way that the VL
quantities cover the entire measurement interval. In the
Viral load measurement platforms routine practice of our laboratory, the VL is measured
either on Biocentric or on CAP/CTM v2.0 depending on the
This study was performed on the 2 platforms used for the circuit of origin of the sample. This routine measurement
measurement of HIV VL in our laboratory, the Centre for was the 1st VL of the sample. Based on this 1st result, the
 Yekayo et al. 88

Table 2. Frequency of VL levels on the two platforms

Viral load (Log10 copies/mL) CAP/CTM (F) Biocentric (F)

Undetectable 93 (51.10%) 103 (56.59%)
[1.3 - 2[ 17 (9.34%) NA
[2- 3[ 27 (14.84%) 36 (19.78%)
[3 - 4[ 11 (6.04%) 11 (6.04%)
≥4 34 (18.68%) 32 (17.58%)

study samples (n=184/419) were selected on both sides virological success or viral suppression has been defined as
and retested on the other platform. Thus, the VL of each a VL < 3 Log10 copies/mL. The scale established by Landis
study sample was measured on the 2 platforms from the 2 and Koch from the coefficients Kappa values (1977) was
aliquots made. Samples that did not have VL results on both used to describe the quality of agreement between the
platforms were excluded from the study. These were results obtained on the 2 platforms.
invalid results on one of the 2 platforms. For VL above detection limits, Bland-Altman (BA)
analysis (Altman and Bland, 1983) and Lin's concordance
Methods of biological analysis correlation coefficient (Lawrence and Lin, 1989) were used
to study the agreement between the 2 platforms. The BA
The CAP/CTM v2.0 platform is a fully automated closed analysis evaluates the mean difference between the results
system that quantifies HIV-1 RNA by reverse transcription of 2 methods and estimates the limits of agreement within
followed by real-time polymerase chain reaction (qRT-PCR) which 95% of the differences lie (interval within which
in human plasma. RNA extraction from samples is 95% of the differences between the results of the 2
automated using the COBAS®AmpliPrep (CAP) instrument. methods lie). Since the maximum acceptable difference in
The COBAS® TaqMan Analyzer (CTM) allows reverse clinical practice is ± 0.5 Log10 copies/mL, we used the
transcription of the extracted RNA into complementary threshold of ± 0.6 Log10 copies/mL to avoid critical
DNA (cDNA), followed by amplification and detection of interpretation on the limits values. The applicability of the
this cDNA. BA was verified by the Shapiro-Wilk test. Lin's correlation
The Biocentric® platform is an open system quantifying coefficient of concordance (CCC) (pc) evaluates the strength
HIV-1 RNA by qRT-PCR in human plasma. In our laboratory, of the agreement between 2 measurements of the same
RNA extraction from samples is semi-automated using variable. It quantifies the differences between the abscissa
nucleic acid extractors (ARROW Nordiag® and Hain®). The points (method 1 results) and the ordinate points (method
retro transcription of the extracted RNA into 2 results) and the 45° line representing perfect agreement.
complementary DNA (cDNA) and the amplification and To assess the degree of agreement, we compared the lower
detection of this cDNA is done on open analyzers (Abi bound of the 95% confidence interval of Lin's CCC
Prism 7500 Fast and Quant Studio 5 Dx). calculated with the scale proposed by McBride (McBrid,
In our laboratory, the 2 platforms are part of the same 2005).
external quality assessment (EQA) program and give 100%
satisfactory results for each of the panels we receive.
Moreover, the open platform was set up in 2002 and has RESULTS
been participating in this EQA program since 2009. The
Roche platform was acquired in 2013 and has been Among the 184 samples in the study, 2 gave an invalid
participating in this program since then. result on one of the 2 assays, one on CAP/CTM v2.0 and the
other on Biocentric.
Statistical analysis The samples were drawn from subjects of average age
46,09±8,95 years old and the gender distribution was
The results were analyzed using NCSS statistical software 29,83% (n=54) male and 70,16% (n=128) female.
(Utah, USA). Statistical significance levels were set at 0.05 The distribution of VL quantification by the 2 platforms is
and clinical differences were set at ±0.6 Log10 copies /mL. presented in Table 2. The proportion of patients with
VL results were expressed as mean ±standard deviation. suppressed viral load 76.37% versus 75.27% and those
The Kappa coefficient obtained with the NCSS software with unsuppressed viral load 23.62% versus 24.72%
was used to evaluate the diagnostic concordance between respectively on Biocentric and CAP/CTMv2.0 were similar
the 2 platforms, on the one hand, at the detection threshold on both platforms.
of the kits (detection threshold = 1.30 Log10 copies/mL for
the CAP/CTM v2.0 platform and 2 Log10 copies/mL for the Evaluation of the strength of agreement between the 2
Biocentric platform) and, on the other hand, at the viral platforms for detectable VL values
suppression threshold (3 Log10 copies/mL). In accordance
with World Health Organization (WHO) guidelines, This step of the concordance evaluation was performed on
Issues Biol. Sci. Pharma. Res. 89

Figure 1: Distribution diagram of the differences between the CAP/CTM v2.0 and Biocentric platforms.

77 samples that had detectable VL on both platforms. The coefficient was 0.846 indicating a near perfect level of
means VL for the 2 platforms were 3.43±1.25 Log10 /mL on agreement according to the criteria proposed by Landis and
CAP/CTM and 3.66 ±1.22 Log10 /mL on Biocentric. Koch (1977) (Table 3). Regarding to the WHO viral
The normality of the distribution of differences, a condition suppression threshold, the Kappa coefficient was 0.878
for the applicability of BA, was verified and accepted (p = indicating a near perfect level of agreement according to
0.975) by the Shapiro-Wilk test (Figure 1). The graph of the the criteria proposed by Landis and Koch (1977) (Table 4).
differences according to BA is shown in Figure 2. BA
analysis revealed a bias of 0.23 (± 0.44) Log10 copies/mL in Clinical relevance of VL differences
favour of Biocentric and the 95% [-0.63 to 1.08] Log10
copies/mL range of limits of agreement was wide compared Efficiency of the 2 Log10 RNA copy threshold of the
to the clinically acceptable threshold (± 0.6 Log10 Biocentric kit
copies/mL). Lin's correlation coefficient of agreement ρ
was 0.922 [0.88 to 0.94]. According to the criteria proposed Of 23 VL between 100 and 390 copies/mL on Biocentric, 3
by (McBride. 2005), this level of agreement was poor. (13.04%) were undetectable, 17 (73.91%) had a detectable
CV between 100 and 390 copies/mL and 3 had a VL > 390
Evaluation of the diagnostic agreement at the threshold copies/mL on CAP/CTM v2.0. Thus, the reduction of the
of detection and viral suppression between the 2 threshold from 390 to 100 copies made it possible to give a
platforms numerical result for 73.91% of presumed undetectable VL
at the threshold of 390 copies/mL.
We evaluated the distribution of VL according to the
detection threshold of the 2 platforms on one hand (Table Clinically significant differences between VLs
3) and according to the WHO viral suppression threshold
on the other hand (Table 4). So, taking into account the Four (2.19%) patients had a VL suppressed on CAP/CTM
detection threshold of the 2 platforms, the Kappa but not suppressed on Biocentric. Of the 103 patients who
 Yekayo et al. 90

Figure 2: Bland Altman mean difference analysis between CAP/CTM v2.0 and Biocentric
platforms (n=77)
Mean Diff (red line) = bias or average of measurement differences between Biocentric and CAP/CTM v2.0 (value: 0.23
Log10 copies/mL).
Dotted line = zero deviation.
Blue lines = 95% limits of agreement. They have ordinates (-0.63 Log10 copies/mL and 1.08Log10 copies/mL).

Table 3. Distribution of undetectable VL diagnoses made by the 2 platforms

Biocentric VL (Log10 copies/ml)

threshold: ˂ 2 Total
Detectable Undetectable
CAM/CTM VL (Log10 Detectable 77 12 89
copies/ml) threshold: ˂ 1.3 Undetectable 2 91 93
Total 79 103 182

Table 4. Distribution of suppressed VL diagnoses performed by the two platforms

Biocentric VL (Log10 copies/ml) Total

Virological success Virological failure
CAM/CTM VL (Log10 Virological success 135 4 139
copies/ml) Virological failure 4 39 43
Total 139 43 182

had undetectable VL (<2 Log10 copies/mL) on Biocentric, patients (6.79%).

91 (88.35%) were also undetectable (<1,3 Log10 Among the 36 who had VL between 2 and 3 Log 10
copies/mL), 5 (4.85%) had detectable VL but < 2 Log 10 copies/mL on Biocentric, 4 (11,11%) had VL < 1,3 Log10
copies/mL, 5 (4.85%) had VL between 2 Log10 and 3 Log10 copies/mL, 12 (33.33%) had VL between 1,3 and 1,99 Log10
copies/mL and 2 (1.94%) had VL ≥ 3 Log10 copies/mL on copies/mL (99 copies/mL), 18 (50%) had VL between 2
CAP/CTM. The difference was clinically significant for 7 and 3 Log10 copies/mL and 2 (5.55%) had VL ≥ 3 Log10
Issues Biol. Sci. Pharma. Res. 91

copies/mL on CAP/CTM v2.0. The difference was therefore 2007; Gueudin et al., 2007; Scott et al., 2009; Swanson et al.,
clinically significant in 5.55% of cases. 2005). Indeed, probes and primers are designed from
Among the 43 patients with unsuppressed VL (≥ 3 Log10 subtype B HIV-1 genome (majorly in developped
copies/mL) on Biocentric, 4 (9.30%) had suppressed but countries), but HIV-1 diversity is increasing worldwide
detectable VL and 39 (90.7%) had unsuppressed VL on (Rouet et al., 2011; Luft et al., 2011; Peeters et al., 2010).
CAP/CTM. The difference was clinically significant in 9.30% However, a discrepancy between 2 analytical methods in
of cases. clinical biology is acceptable if it does not alter the
Overall, there were 18/182 subjects for whom the diagnosis. Thus, we evaluated the agreement between the 2
difference between the 2 VL measurements exceeded 0.6 platforms at the threshold of viral suppression (3 Log10
Log10 copies/mL in absolute value. Also, the VL were higher copies/mL) and the Kappa coefficient obtained was 0.878
on CAP/CTM v2.0 for 3/182 (1.64%) subjects and for the indicating an almost perfect level of agreement. For all VL
others (15/182) they were higher on Biocentric (8.24%). levels, the differences were clinically significant for less
Moreover, these VL concerned 9 suppressed VL on the 2 than 10% of the measurements. Overall, the difference
platforms, 8 unsuppressed VL on the 2 platforms and 1 VL between the 2 VL measurements was clinically significant
suppressed on Biocentric but unsuppressed on CAP/CTM (> 0.6 Log copies/mL absolute value) for 18 patients with
v2.0. These differences, although clinically significant, only higher VL on CAP/CTMv2.0 for 3/182 (1.64%) subjects and
affected the diagnosis and therefore the management of higher on Biocentric for the others (15/182; 8.24%) This
only 1 patient. distribution was similar to that reported by Avetand et al.
(2019). These 18 cases involved 9 suppressed VL on both
platforms, 8 unsuppressed VL on both platforms and 1
DISCUSSION suppressed VL on Biocentric but unsuppred VL on
CAP/CTM v2.0. These differences, although clinically
This study compared the Biocentric HIV viral load kits and significant, only influenced the diagnosis of virological
the Roche CAP/CTM48 HIV-1 Test, Version 2.0 to assess the failure, hence the caring of only 1 patient. The study by
equivalence of their results. Lack of viral laod measurement Kerschberger et al. (2018) showed more significant
equipement is an issue wich interest particularly resources discordance, but also the impact of handling variability
limited settings. So data from the literature are relatively factors by comparing the results of 2 laboratories. The
few. training and experience of the handlers reduced the
The proportion of patients with suppressed viral load discrepancies between the VL measurements on these 2
(76.37% versus 75.27%) on Biocentric and CAP/CTM 48 platforms.
was similar. This result is similar to that (73.7%) reported
by the CIPHIA survey in 2020 (CIPHIA, 2017) among adults
under ART in Côte d'Ivoire. Conclusion
BA analysis revealed a systematic difference of 0.23 Log 10
copies/mL in favor of VL measured on Biocentric. This For the measurement of VL, the Biocentric and CAP/CTM
result was similar to those of Avettand-Fénoël et al. (2019) platforms presented good concordance despite of a small
who reported a systematic difference of 0.4 Log10 bias. So, they could be exchanged without altering the
copies/mL in favour of Biocentric. The range of 95% limits diagnostic capacity for the patient. However, the clinical
of agreement of [-0.63 to 1.08] Log10 copies/mL was wide and biology teams must be in contact for exchanges and
compared to the clinically acceptable threshold (+/-0.6 decision-making for an optimal caring of PLHIV.
Log10 copies/mL); similar result to those of Kerschberger
et al. (2018) in Swaziland who reported 95% limits of Conflict of Interests
agreement of [-1.15 to 1.08] Log10 copies/mL despite a
systematic difference of -0.03 Log10 copies/mL. The level of The authors declare that there is no conflict of interests
agreement of the Lin ρ concordance correlation coefficient regarding the publication of the paper.
was poor. Thus, the 2 platforms were not in satisfactory
agreement. This result could be explained by the difference
of the 2 platforms in the ability to quantify the different REFERENCES
strains of HI V-1. Indeed, Avettand-Fénoël et al. (2019)
showed that the CAP/CTM v2.0 platform underestimated Altman DG, Bland JM (1983). Measurement in medicine: the
the CRF02_AG strain, the preponderant HIV-1 strain (˃ analysis of method comparison studies. J. R. Stat. Soc.:
80%) in Côte d'Ivoire (Toni et al., 2002; 2003). The study Series D (The Statistician), 32(3), 307-317.
by Bourlet et al. (2011) indicated a variable risk of VL Avettand-Fénoël V, Mélard A, Gueudin, M, Maillard A, Dina J,
estimation error or discordance between 4 platforms Gousset M, Chaix ML, Lerolle N, Viard JP, Meyer L,
depending on the HIV-1 strain and the overall disparity was Plantier JC for ANRS-AC11 HIV Quantification Working
higher for CRF02_AG than for B strains. The genetic Group (2019). Comparative performance of the
diversity of HIV-1 subtypes may present a challenge in Biocentric Generic Viral Load, Roche CAP/CTM v1. 5,
monitoring PLHIV with non-B subtypes (Damond et al., Roche CAP/CTM v2. 0 and m2000 Abbott assays for
 Yekayo et al. 92

quantifying HIV-1 B and non-B strains: Underestimation https://diagnostics.roche.com/global/en/products/para

of some CRF02 strains. J. Clin. Virol., 110:36-41. ms/cobas-ampliprep-cobas-taqman-hiv-1-test-v2-0;
Bourlet T, Signori-Schmuck A, Roche L, Icard V, Saoudin H, [page consulted on June 05th, 2021].
Trabaud MA, Tardy JC, Morand P, Pozzetto B, Ecochard R, Rouet F, Liégeois F, Mouinga-Ondémé A, Kania D, Viljoen J,
André P (2011). HIV-1 load comparison using four Wambua S, Ngo-Giang-Huong N, Ménan H, Peeters M,
commercial real-time assays. J. Clin. Microbiol., Nerrienet E (2011). Current challenges to viral load
49(1) :292-297. testing in the context of emerging genetic diversity of
Côte d’Ivoire Population-based HIV Impact Assessment HIV-1. Expert opinion on medical diagnostics, 5(3):183-
(CIPHIA) (2017). Fiche récapitulative Novembre 2020. 202.
URL : https://phia.icap.columbia.edu/wp- Scott LE, Noble LD, Moloi J, Erasmus L, Venter WD, Stevens
content/CIPHIA-French-Summary-Sheet_final.pdf; [page W (2009). Evaluation of the Abbott m 2000 RealTi me
consultée le 25 Juin, 2021]. Human Immunodeficiency Virus Type 1 (HIV-1) Assay for
Damond F, Roquebert B, Benard A, Collin G, Miceli M, Yeni HIV Load Monitoring in South Africa Compared to the
P, Brun-Vezinet F, Descamps D (2007). Human Roche Cobas AmpliPrep-Cobas Amplicor, Roche Cobas
immunodeficiency virus type 1 (HIV-1) plasma load AmpliPrep-Cobas TaqMan HIV-1, and BioMerieux
discrepancies between the Roche COBAS AMPLICOR HIV- NucliSENS EasyQ HIV-1 Assays. J. Clin. Microbiol,
1 MONITOR Version 1.5 and the Roche COBAS 47(7):2209-2217.
AmpliPrep/COBAS TaqMan HIV-1 assays. J. Clin. Swanson P, de Mendoza C, Joshi Y, Golden A, Hodinka R L,
Microbiol, 45(10):3436-3438. Soriano V, Devare S G, Hackett Jr J (2005). Impact of
Generic HIV Charge Virale (2020). URL: human immunodeficiency virus type 1 (HIV-1) genetic
https://www.biocentric.com/generic-hiv-charge-virale; diversity on performance of four commercial viral load
[page consulted on June 05, 2021]. assays: LCx HIV RNA Quantitative, AMPLICOR HIV-1
Gueudin M, Plantier JC, Lemée V, Schmitt MP, Chartier L, MONITOR v1. 5, VERSANT HIV-1 RNA 3.0, and NucliSens
Bourlet T, Ruffault A, Damond F, Vray M, Simon F (2007). HIV-1 QT. J. Clin. Microbiol, 43(8): 3860-3868.
Evaluation of the Roche Cobas TaqMan and Abbott Toni TDA, Masquelier B, Bonard D, Faure M, Huët C,
RealTime extraction-quantification systems for HIV-1 Caumont A, Roques P, Dabis F, Salomon R, Fleury H and
subtypes. JAIDS J. Acquir. Immune Defic. Syndr., the Primo-CI/PAC-CI Study Group (2002). Primary HIV-1
44(5):500-505. drug resistance in Abidjan (Cote d'Ivoire): a genotypic
Kerschberger B, Mpala Q, Uribe PAD, Maphalala G, de la and phenotypic study. Aids, 16(3):488-491.
Tour R, Kalombola S, Bekele A, Chawinga T, Mliba M, Toni TDA, Recordon-Pinson P, Minga A, Ekouevi D, Bonard
Ntshalintshali N, Phugwayo N, Kabore SM, Goiri J, Dlamini D, Bequet L, Huet C, Chenal H, Rouet F, Dabis F, Lafon ME,
S, Ciglenecki I, Fajardo E (2018). Field suitability and Salomon R, Masquelier R, Fleury HJ (2003). Presence of
diagnostic accuracy of the Biocentric® open real-time key drug resistance mutations in isolates from untreated
PCR platform for plasma-based HIV viral load patients of Abidjan, Cote d'Ivoire: ANRS 1257 study. AIDS
quantification in Swaziland. BMC Infect. Dis., 18(1):1-10. research and human retroviruses, 19(8) :713-717.
Landis JR., Koch GG (1977). An application of hierarchical UNAIDS (2021). FACT SHEET 2021 - Latest global and
kappa-type statistics in the assessment of majority regional statistics on the status of the AIDS. URL:
agreement among multiple observers. Biometrics, 363- https://www.unaids.org/sites/default/files/media_asset
374. /UNAIDS_FactSheet_fr.pdf; [consulted on June 25th
Lawrence I, Lin, K. (1989). A concordance correlation 2021].
coefficient to evaluate reproducibility. Biometrics, 255- World Health Organization. (2016). Guidance notes:
268. Consolidated guidelines for the use of antiretroviral
Luft LM, Gill MJ, Church DL (2011). HIV-1 viral diversity and drugs for the treatment and prevention of HIV infection:
its implications for viral load testing: review of current latest information. World Health Organization. URL:
platforms. Int. J. Infect. Dis., 15(10), e661-e670. https://apps.who.int/iris/handle/10665/206448;
McBride GB (2005). A proposal for strength-of-agreement [consulted on June 20th 2021].
criteria for Lin’s concordance correlation coefficient.
NIWA client report: HAM2005-062, 62.
Peeters M, Aghokeng AF, Delaporte E (2010). Genetic
diversity among human immunodeficiency virus-1 non-B
subtypes in viral load and drug resistance assays. Clin.
Microbiol. Infect., 16(10):1525-1531.
Roche Molecular Diagnostics (2021). COBAS®
AmpliPrep/COBAS® TaqMan® HIV-1 Test, v2.0. URL: