Introduction
HIV drug resistance (HIVDR) to antiretroviral therapy (ART) has increased globally, with a prevalence of pretreatment HIVDR (PDR) of 5–20% and 50–90% in people failing ART [1]. In women and children with previous exposure to antiretroviral drugs for prevention of mother-to-child transmission (MTCT) of HIV, PDR to nucleoside reverse transcriptase inhibitors (NRTIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) can be up to 30 and 75%, respectively [2]. In addition, a study in Uganda reported 98% of children failing an NNRTI first-line regimen had dual-class HIVDR, with thymidine analog mutations in 50% [3], which can undermine the effectiveness of second-line and third-line regimens.
In response to the high prevalence of PDR to NNRTI, the WHO recommends countries with more than 10% PDR to start all people living with HIV on a preferred first-line regimen containing dolutegravir/NRTIs [4], which has demonstrated a high barrier to resistance in multiple clinical trials [5,6]. However, NNRTI-based regimens are still the preferred first-line option in many communities that recognize drug patents, due to the high cost of dolutegravir-based regimens [7], and in populations in which fixed-dose combinations containing dolutegravir are not recommended such as persons co-infected with tuberculosis or young infants. In countries with more than 10% PDR where using an integrase inhibitor-based regimen is not feasible, the WHO recommends testing for PDR to guide the selection of an effective regimen [4]. In addition, in populations switching to dolutegravir-based regimens where HIVDR to NRTIs is prevalent, HIVDR testing could aid in selecting effective NRTI combinations to prevent the possibility of functional dolutegravir monotherapy, as increased virologic failure has been associated with dolutegravir monotherapy [8].
At present, in resource-limited settings (RLS), HIVDR testing using Sanger sequencing is limited to centralized laboratories. In these settings, especially where NNRTI-based first-line regimens are still used, HIV management could be more effective with the use of simple and economical HIVDR tests yielding fast turnaround time for test results, opening the possibility of making informed clinical decisions within a single visit. To expedite care in RLS, our group has developed ‘OLA-Simple’ – a near point-of-care test based on the oligonucleotide ligation assay (OLA) [9,10], which detects a set of mutations associated with virological failure of first-line NNRTI/NRTI regimens [11]. OLA-Simple detects the presence of specific drug-resistance mutations in a PCR amplified HIV template with nearly 100% specificity conveyed by enzyme-mediated ligation of labeled genotype-specific oligonucleotide probes to a common capture probe at the site of the mutation. The OLA-Simple kit contains lyophilized reagents for easy assay setup, lateral flow strips for visual detection of HIVDR, and software guidance to enable first-time users to perform the assay with minimal training [11]. The OLA-Simple software guidance was developed based on ‘Aquarium’ – a human in-the-loop application that provides users step-by-step interactive instruction [12], and can be run on an inexpensive tablet. Moreover, an OLA-Simple kit that detects the NRTI mutation M184V and NNRTI mutations K103N, Y181C, and G190A was clinically validated across multiple HIV-1 subtypes, showing 99.6% sensitivity and 98.2% specificity compared to Sanger sequencing [11]. OLA-Simple testing performed at the Coptic Hope Center in Kenya by local laboratory technicians showed that 12 of 12 technicians successfully completed the entire workflow, and correctly genotyped 92.3% of codons tested (111/120) (unpublished). In this study, we have further improved the OLA-Simple assay in two areas: validated probes for detection of K65R – a major resistance mutation to tenofovir, which is an NRTI widely used with NNRTIs or dolutegravir, and exploited an in-house software program to automatically quantify signal and classify results from lateral flow tests. Additionally, we have adapted the OLA-Simple kit probe sequences to detect these NNRTI/NRTI mutations in a Mexican population and validated the improved kits by genotyping PDR in a cohort of adults initiating first-line NNRTI/NRTI regimens in Mexico City.
Methods
Study design for clinical validation
A database containing HIVDR profiles from 2412 individuals initiating first-line ART in Mexico City genotyped by MiSeq was used to select 60 plasma specimens enriched for mutations K65R, K103N/S, Y181C, M184V, and G190A. This set of mutations was selected due to their frequency in the Mexican cohort and their relevance for the preferred first-line regimens in use at the time of study design. All participants provided written informed consent before blood sample donation for HIV genotyping. The study was approved by the Institutional Review Board of the National Institute of Respiratory Diseases – institution where HIV sequencing was performed (Project code: E02–17). De-identified specimens from treatment-naive adults were blindly tested by OLA-Simple, and the results compared to results from MiSeq sequencing. GenBank accession numbers are MT044195-MT044237; MT070765-MT070781.
Design of OLA-Simple probes for a Mexican population
OLA-Simple probes designed to detect K65R, K103N/S, Y181C, M184V, and G190A across multiple subtypes were modified to include polymorphisms observed frequently in pol sequences from individuals initiating an NNRTI/NRTI regimen in Mexico City between September 2017 and March 2018 [13]. Supplementary Table 1 (https://links.lww.com/QAD/B693) lists the probe sequences used for the Mexican cohort.
Preparation of OLA-Simple kits and sample testing
The OLA-Simple kit was prepared as previously reported [11]. All kit components were manufactured and packaged in the Lutz lab at the University of Washington, except the lyophilized reverse-transcription reagents (Takara Bio Inc., Shiga, Prefecture, Japan) and QIAamp viral RNA mini kit (Qiagen, Hilden, Germany). Briefly, PCR, ligation, and gold detection mixtures were lyophilized in the presence of additives to protect reaction components during lyophilization and drying process. In this study, ligation mixtures included the newly designed oligonucleotide probes. Lateral flow strips were striped with antidigoxigenin antibody, antifluorescein antibody, and biotin-bovine serum albumin. All kit components were stored at 4°C until use. Assay optimization was performed using HIV subtype B plasmids mixtures containing 0, 10, 25, and 100% mutant for all codons tested as previously described [11]. Plasma specimens were extracted using the Qiagen QIAamp Viral RNA kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Extracted RNA was reverse-transcribed in a rehydrated lyophilized reverse-transcription reaction. cDNA from plasma was then amplified in rehydrated lyophilized PCR reactions. Aliquots of amplicon were added to rehydrated lyophilized ligation mixtures for each mutation tested and subjected to ligation. Ligation products were then visualized by lateral flow detection using gold particles. Lateral flow detection strips were scanned after 15–20 min and analyzed using an in-house software. To determine the threshold for indeterminate (IND) results [negative for both mutant (MUT) and wild-type bands], 12 mock IND strips were generated using an HIV template-free ligation reaction.
Analysis of OLA-Simple results
The in-house python software was designed to automatically crop the region of interest (ROI) from the scanned images of lateral flow tests. The ROI was then converted to grayscale by linearly combining the Red-Green-Blue channels using coefficients selected by linear discriminant analysis to differentiate pixels in the band and paper background regions. The locations of the control, MUT, and wild-type bands were identified by a combination of peak detection and heuristic peak selection. To compute the intensity, the software calculated a t-statistic between rectangular pixel regions on and immediately adjacent to each band. This normalized t-statistic is referred to as the ‘signal’ of the strip. The threshold for IND results was selected as six standard deviations (SDs) above the mean of the wild-type signal from mock indeterminate strips. A receiver-operating characteristic (ROC) curve was generated using SPSS Data Analysis Software (IBM, Armonk, New York, USA) setting the MUT signal from all OLA-Simple as the ‘test variable’ and the diagnostic wild-type or MUT classifications as the ‘state variable.’ The state variables were determined by MiSeq using 2, 5, 10, 15, 20, and 25% mutant frequency thresholds. At 10% mutant diagnostic threshold, a MUT signal threshold was selected to maximize the concordance of OLA-Simple results and MiSeq data.
Analysis of HIVDR using MiSeq
Deep sequencing by MiSeq (Illumina, San Diego, California, USA) on near-full genome HIV amplicons was performed at the CIENI/INER laboratory in Mexico City as previously described [13,14]. Briefly, DNA libraries were prepared using Nextera XT DNA Sample Preparation kit and Nextera XT Index Kits (Illumina), according to the manufacturer's protocol, and were run with 500-cycle MiSeq Reagent Kits v.2. Fastq files from the sequencing runs were aligned and assessed for HIVDR frequencies using HyDRA (National Microbiology Laboratory, Public Health Agency of Canada, Winnipeg, MB, Canada) [15]. Amino acid mutations were queried against the Stanford HIV Drug Resistance Database [16]. Note that MiSeq has reported analytical sensitivities of less than 2% variants [17], but we chose a more conservative 2% threshold [18] as the lowest diagnostic threshold in this study.
Statistical analysis
The number of technical replicates in each experiment was reported at the figure legends and provided in the ‘Methods’ section. Multiple comparisons of means across groups were calculated using t test, with alphas adjusted using the Holm-Bonferroni method. Concordance, sensitivity, specificity, and rates are reported with standard errors (SEs) of binomial proportion. The areas under the curve of ROC curves are reported with 95% confidence intervals (95% CI).
Data availability statement
All data relevant to this study were included in the main text and supplementary information.
Results
The workflow of the OLA-Simple kit includes four main steps: nucleic acid extraction (not shown), RT-PCR amplification, ligation, and detection (Fig. 1a). These steps have been significantly simplified by premeasuring and lyophilizing reaction components to enable fast and accurate setup, and to increase stability without freezer storage. The lyophilized reaction mixtures include necessary enzymes, primers, and/or probes for HIVDR testing, and can be activated by rehydration with water. OLA-Simple detection generates visual results on lateral flow strips, which can be read by the unaided eye, as previously reported [11,19]. Here, we employed an in-house software to quantify the signal and classify HIVDR results to enable consistent analysis and interpretation across users.
;)
Fig. 1: Workflow and analytical sensitivity of the OLA-Simple kit.
We adapted the ligation module of the OLA-Simple using probes optimized for the Mexican population. The concentrations of ligation probes and assay conditions were initially optimized and tested on HIV subtype B plasmid DNA standards (wild-type, mutant, or wild-type/mutant mixtures at the five codons tested). The updated OLA-Simple assay clearly differentiated 0% (wild-type HIV) from 10% mutant across all codons (Fig. 1b).
Clinical validation of this version of the OLA-Simple was conducted on 60 plasma specimens from a cohort in Mexico selected by collaborators at CIENI/INER to include a high proportion of specimens with PDR. Across these specimens, 98.3% (59/60) amplified by reverse transcription polymerase chain reaction (RT-PCR) (median viral load: 103 586, range 176–10 000 000 copies/ml). The one specimen that failed amplification had a plasma HIV RNA load of 176 copies/ml, with an estimated 29 copies submitted to the RT-PCR reaction. The 59 specimens with successful amplification underwent ligation and detection for HIVDR testing on a total of 295 codons (five mutations each). OLA-Simple genotyping was indeterminate (IND) for 2.7% (8/295) of all codons tested (Table 1), with 8.5% (5/59) at K65R, 3.4% (2/59) at Y181C, and 1.7% (1/59) at G190A. There were no IND results at K103N/S and M184V.
Table 1: Genotyping results by OLA-Simple.
Compared to MiSeq using 10% mutant frequency as the diagnostic threshold, OLA-Simple had five false-positive, four false-negative, and eight IND results. False-positive occurred at M184V (n = 2/59) and G190A (n = 3/59); all these false-positives by OLA-Simple were 100% wild-type by MiSeq. Of the four false-negative results, three were K103N/S (at 10.5, 11.9, and 22.4% mutant frequency) and one at G190A (at 20.59% mutant frequency). Based on these results, OLA-Simple had 93.1% sensitivity and 97.8% specificity across all codons. Genotyping data by OLA-Simple and MiSeq are available in Supplementary Table 2 (https://links.lww.com/QAD/B693). Percentage sensitivity and specificity at other diagnostic thresholds are shown in Supplementary Table 3 (https://links.lww.com/QAD/B693).
The accuracy of OLA-Simple genotyping results was estimated by comparison to MiSeq genotypes, excluding IND results (n = 287). ROC curves were generated using different diagnostic mutant frequency thresholds (Fig. 2). The areas under the curve reflect the accuracy of the MUT signal from OLA-Simple for classification of HIVDR results compared to those determined by MiSeq. At 2, 5, 10, 15 (or 20%), and 25% mutant frequency thresholds, the accuracy of OLA-Simple was 93.2% (95% CI 89.0–97.5%), 96.3% (95% CI 93.0–99.6%), 97.2% (95% CI 94.1–100%), 97.4% (95% CI 94.4–100%), and 98.8% (95% CI 97.1–100%), respectively.
Fig. 2: Receiver-operating characteristic (ROC) curves.
Discussion
We have adapted a low-cost, easy-to-use, rapid OLA-Simple HIVDR test for a Mexican population. Specifically, we have designed probes to detect major drug resistance mutations to NNRTIs (EFV/NVP) and NRTIs (3TC/TDF). The analytical sensitivity of OLA-Simple is 10% mutant among the HIV quasi-species, which is better than the reported sensitivities of 20–50% by Sanger sequencing [20,21]. The OLA-Simple results compared favorably to MiSeq results performed at CIENI/INER in Mexico City, which has met the WHO requirements for good laboratory practices and quality assurance in HIV genotyping.
In this study, OLA-Simple primers amplified all except one specimen with a viral load of 176 copies/ml (29 RNA copies/RT-PCR reaction), which is near the detection limit of the OLA-Simple PCR (10 copies of DNA/reaction) [11]. Amplification failure in this sample could be due to variations in the RNA extraction efficiency between the plasma aliquots used for viral load measurement and OLA-Simple, or lower efficiency during reverse transcription.
The IND results in OLA-Simple can occur due to mismatches of the probes to the target HIV sequence. The alignment of 2412 Mexican HIV pol sequences showed considerable variability in the area of several OLA-Simple kit probes, which were designed as universal probes to test HIV subtypes A, B, C, D, and AE. To improve ligation efficiency in the Mexican cohort, where HIV subtype B accounts for nearly 99% of infections [14,22–24], we designed subtype B-specific probes that included mixtures of nucleotides at sites of frequent polymorphisms, or mixtures of probes to accommodate variations at multiple sites close to the probes’ ligation site. Probe design for K103N proved particularly challenging due to relatively frequent changes at this or adjacent codons 102 (K102Q/R), 103 (K103S, K103R), and 104 (K104R). K102Q/R, K103R, and K104R have not been associated with reduced susceptibility to NNRTIs [15], but can interfere with detection of K103N by OLA-Simple, as their presence could interfere with the DNA ligase requirement that the four nucleotides surrounding the ligation site have perfect complementarity with the target [25], causing an IND result. In contrast, K103S confers high-level resistance to NVP and intermediate to EFV [26]. Therefore, probes detecting both K103N and K103S were tested simultaneously in this updated OLA-Simple kit by mixing four genotype-specific probes and one common probe, resulting in no IND results at K103N/S in this cohort. Universal probes for K65R in the OLA-Simple kit were slightly modified for the subtype B Mexican variants, and these had not been previously validated on clinical specimens. Despite modifications, these K65R probes had the highest rate of IND results (8.5%) mainly due to closely spaced nucleotide variations at codons 67, 68, and 69, including the polymorphic mutation S68G (Supplementary Table 3, https://links.lww.com/QAD/B693). To reduce the IND rate for K65R, a mixture of common probes complementary to these polymorphisms/mutations may be necessary at this site. The subtype B-specific probes tested in this study reduced the combined IND rate at codons K103N/S, Y181C, M184V, and G190A to 1.3% (3/236) from a 2.7% IND rate observed when using probes designed to work ‘universally’ across multiple HIV-1 group M subtypes, and tested on specimens from Kenya, South Africa, Peru, and Thailand [11]. Further reduction of the overall IND rate could be achieved by including probes for alternative variants at the codons tested, such as M184I and G190S, depending on the frequency of these mutations in the target population.
Quantitative comparison of OLA-Simple results and MiSeq was presented in ROC curves at different mutant frequency thresholds, because a clinically relevant threshold has not yet been established [27]. The use of unique molecular indices (UMI) to quantify the actual number of HIV templates sequenced would likely allow a more accurate assessment of these thresholds. At 10% mutant threshold based on MiSeq, OLA-Simple's accuracy was higher than at 5% (97.2 vs. 96.3%), but gave 1.4% (4/295) false-negative and 1.7% (5/295) false-positive. Because the number of templates tested by each assay was unknown, the false-positive results could be true positives not detected by MiSeq. On the contrary, false-positives could be due to faint background on the MUT band, which was classified as MUT by our software. In this study, we used a single signal threshold for MUT classification across all codons tested, but the OLA-Simple signals from different codons are not equal in their intensity. To reduce the potential for false-positives, the signal threshold for software analysis can be further optimized for each codon once a larger dataset of OLA-Simple results is available. Additionally, the present algorithm used to automatically select regions of interest is sensitive to image artifacts (e.g. visible stains, dirt, or defects on the strip) that can skew the means and variances of all pixels within the cropped region. Future versions of our software will be designed to be more robust to these artifacts by dynamically determining the shape of the band region based on the width of the peak, and raising unresolvable errors detected in the image to users for manual verification.
Setting up HIVDR testing using the approach presented here requires an RNA extraction step, which can be performed with commercially available kits in 20 min, and approximately 10 min of hands-on time for setting up RT/PCR amplification, ligation, and lateral flow detection. Lyophilized reagents drastically reduce assay setup time, eliminating the need to thaw reagents and precisely measure and mix different volumes of multiple reagents to prepare master mixes for each step (RT, PCR, and ligation). The total wait time includes 3-h RT/PCR amplification, 1-h ligation, and 20-min detection (see a detailed comparison of workflow, assay time and cost between OLA-Simple and MiSeq sequencing in Supplementary Tables 4 and 5, https://links.lww.com/QAD/B693). Recognizing that most of the wait-time comes from the amplification step, we have recently revamped this step to combine RT-PCR in a single tube and reduce the amplification time to 40 min [28]. With this modification, the turnaround time can be reduced from approximately 4.5 h to less than 2.5 h.
The OLA-Simple is a platform technology that can be adapted to detect other point mutations. In the context of HIV, we have designed ligation probes to detect mutations associated with resistance to other drug classes [e.g. integrase strand transfer inhibitors (unpublished) and protease inhibitors] [10]. As new recommendations for first-line ART regimens are implemented globally, new formats of the OLA-Simple kit can be manufactured to detect relevant mutations associated with virologic failure of these ART.
The OLA-Simple kit has the potential to improve clinical management of HIV treatment in resource-limited settings by providing a less costly and complex workflow, simple data analysis, and faster turnaround time from sample-to-results compared to drug resistance testing by Sanger sequencing. Sequencing usually requires centralized testing due to the expensive and sophisticated equipment and technical expertise needed. In its present form, the OLA-Simple could be implemented in local clinics with laboratories of limited infrastructure as it only requires basic laboratory equipment such as micropipettes, a thermal cycler (∼$500 miniPCR), a computer or tablet (∼$50 Amazon Fire) to operate Aquarium software for guidance and analysis, and an office scanner (∼$50 Cannon CanoScan LiDE300) for acquiring the image of lateral flow tests. The lateral flow strips can be read visually or captured with a cell phone; however, we have chosen to use an office scanner to ensure even light exposure across images. In addition, digital images allow automated analysis using an in-house Python script, which decreases variation in interpretation of results when low-frequency mutations (minority variants) produce mutant bands of lower signal intensity. Settings in which access to electricity is intermittent, applications of OLA-Simple could use a utility power supply (e.g. ∼$50 APC UPS BN450M). These simple equipment requirements and relatively short turnaround time of OLA-Simple could potentially enable clinicians in low-resource settings prescribe same-day guided antiretroviral treatment.
Although the validation of the kit reagents and software on Mexican clinical specimens conducted in the Seattle laboratory showed high sensitivity and specificity to detect key HIV drug-resistance mutations, demonstration of the feasibility of on-site adoption and use of the current OLA-Simple kit is warranted. At the time of writing of the present manuscript, Mexican ART guidelines were modified to include second-generation integrase inhibitor-based first-line ART regimens as the preferred option at the national level [29], making baseline HIVDR testing less relevant. Nevertheless, the need for rapid and economical HIVDR testing persists in several scenarios, including in infants born to ART-exposed mothers, persons failing first-line ART, persons switching to optimized ART regimens, and persons that for medical conditions or drug availability issues need to start with efavirenz-based ART regimens. This OLA-Simple validation experience will be highly valuable for other countries with restricted access to HIV sequencing, which have shown high pretreatment HIVDR level and that remain constrained in their ART regimen options [2].
Conclusions
Compared to MiSeq, the OLA-Simple detected HIVDR with high sensitivity and accuracy. Moreover, minor changes in probe design and refinement of the software analysis could further reduce indeterminate results and improved accuracy of genotype classification. Use of OLA-Simple in RLS could improve access to affordable and rapid HIVDR testing, thus facilitating the appropriate choice of ART in populations using NNRTI-based regimens.
Acknowledgements
We thank our collaborators at the University of Washington: Dr James Lai, Dr Eric Klavins, and Justin Vrana for their helpful technical discussion. We also thank Justin Vrana for his participation in the initial visit at the CIENI/INER along with N.P., B.R.L., and L.M.F. We thank other laboratory members: Colin Eckhoff for his generous technical support; Enos Kline, David McIntyre, Nikki Higa, Annie Wong-On-Wing, Jonathan Lim, Daisy Ko, and Ian Andrews for their technical contribution in the development of OLA-Simple that has led to this work. We thank our clinical collaborators who provided clinical insights in other related work: Drs Theresa Rossouw and Ute Feucht at the University of Pretoria, South Africa; Dr Michael Chung at Coptic Hope Center, Kenya; Drs Gonzague Jourdain and Nicole Ngo-Giang-Huong at PHPT Thailand; and Dr Jaime Soria at Hospital Nacional Dos de Mayo, Peru. Last, we thank Rahil Jain, Blythe Adamson, Joshua Buser, Qin Wang, Justin Vrana, and Richard Lee for collaborating with N.P. to acquire awards from UW Health Innovation Challenge and Business Plan Competition.
Authors’ contributions: N.P., L.M.F., B.R.L. forged the collaboration between the University of Washington and CIENI/INER that has led to this work. L.M.F. and B.R.L. oversaw and provided feedback for the study design and blinding protocol. I.A.B. designed the probe sequences and optimized the assay condition for Mexico's population using plate OLA. N.P. optimized the assay conditions for the OLA-Simple kit and blindly analyzed the specimens using the OLA-Simple kit. N.P. and P.S.R. manufactured the OLA-Simple kits. N.P., P.S.R., and I.A.B. analyzed the OLA-Simple data. S.A., C.G., M.S., D.T., M.M., H.E., S.D., and G.R. collected clinical specimens, developed the clinical specimen panels, and analyzed specimens using MiSeq. All authors provided input into writing this manuscript and approved the final version of this manuscript.
Funding: UW Health Innovation Challenge Award funded the transportation for the team at the University of Washington to visit CIENI/INER in Mexico City. National Institutes of Health (AI110375) and the Clinical and Retrovirology Research Core and the Molecular Profiling and Computational Biology Core of the UW CFAR (P30 AI027757) funded OLA-Simple development and validation; the Mexican Government (Comisión de Equidad y Género de las Legislaturas LX-LXI y Comisión de Igualdad de Género de la Legislatura LXII de la H. Cámara de Diputados de la República Mexicana) and Consejo Nacional de Ciencia y Tecnología (CONACyT SALUD-2017-01-289725), funded HIV sequencing in Mexico City using MiSeq. The funders have no roles in study design, data collection, analysis or interpretation.
Conflicts of interest
A patent application related methods for OLA-Simple was filed under the University of Washington and Seattle Children's Research Institute.
Data were partially presented at the 2019 International AIDs Society Conference and published in the conference proceedings (Abstract# A-1077-0086-01711) in Mexico City, Mexico.
References
1. World Health Organization.
HIV Drug Resistance Report 2017. Geneva, Switzerland: WHO: 2017.
2. World Health Organization.
HIV Drug Resistance Report 2019. Geneva, Switzerland: WHO: 2019.
3. Sigaloff KCE, Kayiwa J, Musiime V, Calis JCJ, Kaudha E, Mukuye A, et al.
Short communication: high rates of thymidine analogue mutations and dual-class resistance among HIV-infected Ugandan children failing first-line antiretroviral therapy.
AIDS Res Hum Retroviruses 2013; 29:
4. World Health Organization.
Guildelines on the public health response to pretreatment HIV drug resistance; July 2017. Geneva, Switzerland: WHO: 2017.
5. Dow DE, Bartlett JA.
Dolutegravir, the second-generation of Integrase Strand Transfer Inhibitors (INSTIs) for the treatment of HIV.
Infect Dis Ther 2014; 3:83–102.
6. Christopher E, Kandel, Walmsley SL.
Dolutegravir: a review of the pharmacology, efficacy, and safety in the treatment of HIV.
Drug Des Devel Ther 2015; 9:3547–3555.
7. Sim J, Hill A.
Is pricing of dolutegravir equitable? A comparative analysis of price and country income level in 52 countries.
J Virus Eradicat 2018; 4:230–237.
8. Hocqueloux L, Raffi F, Prazuck T, Bernard L, Sunder S, Esnault JL, et al.
Dolutegravir monotherapy versus Dolutegravir/Abacavir/Lamivudine for virologically suppressed people living with chronic human immunodeficiency virus infection: the Randomized Noninferiority MONotherapy of TiviCAY Trial.
Clin Infect Dis 2019; 69:1498–1505.
9. Beck IA, Crowell C, Kittoe R, Bredell H, Machaba M, Willamson C, et al.
Optimization of the oligonucleotide ligation assay, a rapid and inexpensive test for detection of HIV-1 drug resistance mutations, for non-North American variants.
J Acquir Immune Defic Syndr 2008; 48:418–427.
10. Beck IA, Mahalanabis M, Pepper G, Wright A, Hamilton S, Langston E, et al.
Rapid and sensitive oligonucleotide ligation assay for detection of mutations in human immunodeficiency virus type 1 associated with high-level resistance to protease inhibitors.
J Clin Microbiol 2002; 40:1413–1419.
11. Panpradist N, Beck IA, Vrana J, Higa N, McIntyre D, Ruth PS, et al.
OLA-Simple: a software-guided HIV-1 drug resistance test for low-resource laboratories.
EBiomedicine 2019; 50:34–44.
12. Keller B, Vrana JD, Miller A, Newman G, Klavins E.
Aquarium: the laboratory operating system (Version v2.5.0).
Zenodo 2019.
13. Margarita M-FAC, Santiago A-R, Mehta SR, Hector E.P.-J., Manuel A.B.-R., Silvia A.,
et al.
Pretreatment HIV drug resistance spread within transmission clusters in Mexico City.
J Antimicrob Chemother (in press).
14. Ávila-Ríos S, García-Morales C, Matías-Florentino M, Romero-Mora KA, Tapia-Trejo D, Quiroz-Morales VS, et al.
Pretreatment HIV-drug resistance in Mexico and its impact on the effectiveness of first-line antiretroviral therapy: a nationally representative 2015 WHO survey.
Lancet HIV 2016; 3:e579–e591.
15. Taylor T, Lee ER, Nykoluk M, Enns E, Liang B, Capina R, et al.
A MiSeq-HyDRA platform for enhanced HIV drug resistance genotyping and surveillance.
Sci Rep 2019; 9:8970.
16. Shafer RW.
Rationale and uses of a public HIV drug-resistance database.
J Infect Dis 2006; 194:S51–58.
https://hivdb.stanford.edu/
17. Casadella M, Paredes R.
Deep sequencing for HIV-1 clinical management.
Virus Res 2017; 239:69–81.
18. Perrier M, Desire N, Storto A, Todesco E, Rodriguez C, Bertine M, et al.
Evaluation of different analysis pipelines for the detection of HIV-1 minority resistant variants.
PLoS One 2018; 13:e0198334.
19. Panpradist N, Beck IA, Chung MH, Kiarie JN, Frenkel LM, Lutz BR.
Simplified paper format for detecting HIV drug resistance in clinical specimens by oligonucleotide ligation.
PLoS One 2016; 11:e0145962.
20. Gunthard HF, Wong JK, Ignacio CC, Havlir DV, Richman DD.
Comparative performance of high-density oligonucleotide sequencing and dideoxynucleotide sequencing of HIV type 1 pol from clinical samples.
AIDS Res Hum Retroviruses 1998; 14:869–876.
21. Schuurman R, Demeter L, Reichelderfer P, Tijnagel J, de Groot T, Boucher C.
Worldwide evaluation of DNA sequencing approaches for identification of drug resistance mutations in the human immunodeficiency virus type 1 reverse transcriptase.
J Clin Microbiol 1999; 37:2291–2296.
22. Avila-Rios S, Garcia-Morales C, Garrido-Rodriguez D, Ormsby CE, Hernandez-Juan R, Andrade-Villanueva J, et al.
National prevalence and trends of HIV transmitted drug resistance in Mexico.
PLoS One 2011; 6:e27812.
23. Avila-Rios S, Garcia-Morales C, Matias-Florentino M, Romero-Mora KA, Tapia-Trejo D, Quiroz-Morales VS, et al.
Pretreatment HIV-drug resistance in Mexico and its impact on the effectiveness of first-line antiretroviral therapy: a nationally representative 2015 WHO survey.
Lancet HIV 2016; 3:e579–e591.
24. Avila-Rios S, Garcia-Morales C, Valenzuela-Lara M, Chaillon A, Tapia-Trejo D, Perez-Garcia M, et al.
HIV-1 drug resistance before initiation or re-initiation of first-line ART in eight regions of Mexico: a sub-nationally representative survey.
J Antimicrob Chemother 2019; 74:1044–1055.
25. Barany F.
Genetic disease detection and DNA amplification using cloned thermostable ligase.
Proc Natl Acad Sci U S A 1991; 88:189–193.
26. Stanford University. HIV drug resistance database. 2019.
27. Inzaule SC, Hamers RL, Noguera-Julian M, Casadella M, Parera M, Kityo C, et al.
Clinically relevant thresholds for ultrasensitive HIV drug resistance testing: a multicountry nested case-control study.
Lancet HIV 2018; 5:e638–e646.
28. Panpradist N, Kline E, Beck I, Oreskovic A, Hull I, Purfield M, et al.
Rapid semi-quantitative viral load assay and reflex drug resistance test for management of HIV antiretroviral therapy.
XXVIII the International Workshop on HIV Drug Resistance and Treatment Strategies. Johannesburg, South Africa; 2019.
29. CENSIDA.
Guía de Manejo Antirretroviral de las Personas con VIH. Mexico; 2019.
