Based on phylogenetic analysis, HIV is classified into 3 distinct groups designated M (major), N (new, non M, or non O), and O (outlier). Group M is divided into subtypes A-K and is responsible for the majority of HIV infections worldwide.1 B subtype viruses are most common in the industrialized world whereas >70% of all HIV strains globally are nonsubtype B. Recombination between different subtypes has further increased the genetic diversity leading to various circulating recombinant forms (CRF01-CRF32).1 Despite the vast majority of HIV strains being nonsubtype B and the large prevalence of certain Circulating recombinant form (CRF) in many African countries, most drug screening and susceptibility studies have been done using subtype B viruses.
The availability of highly active antiretroviral therapy for people living with HIV is increasing rapidly in sub-Saharan Africa,2 and this has also led to emergence of resistance against antiretrovirals (ARVs). Earlier studies before large-scale ARV introduction had reported the emergence of resistance in individuals undergoing treatment in some African countries including Côte d'Ivoire, Uganda, South Africa, and Zimbabwe.3-7 Because of the recent scale-up, however, most reports on resistance have been preliminary, involving small cohorts of drug-exposed subjects benefiting from free ARV supplies. The prevalence of drug resistance mutations after 10 months of therapy ranged from 16.2% in Cameroon to 23%-34% in Côte d'Ivoire among adults and children, respectively.8,9 To date, nevirapine (NVP)-associated mutations (K103N and Y181C) are most commonly reported, mostly because NVP has been extensively used for prevention of HIV mother-to-child transmission in sub-Saharan Africa.10 Several recent studies have also reported the presence of drug resistance mutations among treatment-naive subjects, suggesting transmission of resistant viruses.11-14
Burkina Faso is a West African country which reports an HIV prevalence of <3% in rural and urban areas.15 In recent years, ARV treatment has been substantially increased in Burkina Faso (13,290 as of May 2007) through support by the Global Fund and private initiatives. However, there is presently no sequencing facility to monitor the development of drug resistance in ARV-exposed patients in this country and the drug resistance pattern is therefore unknown. We have performed genotypic resistance analysis for 75 patients with 11 of them providing at least 2 sequential samples at the Centre Hospitalier Yalgado Ouedraogo (CHUYO) in Ouagadougou, Burkina Faso. Drug resistant viruses were analyzed after sequencing the relevant regions in the protease (PR) and reverse transcriptase (RT) of pol. Phylogenetic analysis was performed to classify the sequences into different subtypes. Further, we compared the drug resistance profiles between CRF02_AG and CRF06_cpx strains, the 2 main viral strains circulating in Burkina Faso, and observed a significantly higher frequency of certain resistance mutations and of multiresistant strains for CRF06_cpx strains.
MATERIALS AND METHODS
Subjects and Sample Collection
Samples were obtained from HIV-infected drug-exposed individuals failing therapy (viral load >1000 copies/mL or a decrease in CD4 cell count of >50 cells/mm3) presenting at CHUYO in Ouagadougou, Burkina Faso. Informed consent was obtained from all subjects participating in the study. Whole blood was collected from the study participants, and plasma was separated from cells by centrifugation at 3000 rpm for 5 minutes and stored at −80°C in several aliquots.
CD4 Count and Viral Load Determination
Whole blood less than 6 hours old was used to determine the CD4 cell counts of infected subjects using a FACS count flow cytometer (Becton Dickinson, San Jose, CA) at the Virology Unit of CHUYO. Quantification of viral RNA was carried out at the same site using the LcX HIV RNA Quantitative Probe system (Abbott Diagnostics, Rungis Cedex, France). All assays were done as stipulated by the manufacturers. Frozen plasma was shipped to the Department of Virology, University Clinic Heidelberg, for further analysis. Viral load was again determined after shipment for all samples using the Amplicor HIV-1 Monitor assay, version 1.5 (Roche Diagnostics, Mannheim, Germany) before reverse transcriptase-polymerase chain reaction amplification.
RNA Extraction, Polymerase Chain Reaction, and Genotypic Analysis
RNA extraction and reverse transcriptase-polymerase chain reaction were performed using the Viroseq kit (Abbott Diagnostics, Wiesbaden, Germany) according to the standard protocol. Polymerase chain reaction products were separated on agarose gels, and amplified fragments >1.2 kb (containing the PR and 5′RT regions) were purified and sequenced using the dye termination method on an ABI 3100 sequencer (Applied Biosystem, Foster City, CA). Putative PR and RT drug resistance mutations were determined using the Drug Resistance Algorithm from the Stanford HIV database.16 This tool compares the input sequences with sequences of HIV-1 subtype B isolates shown to confer resistance to anti-HIV drugs.
For phylogenetic analysis, sequence alignment of the gene fragments was performed automatically using CLUSTAL-X with minor manual adjustments taking into consideration protein-coding sequences.17 New sequences were aligned with the recent compilation of nonrecombinant and CRF HIV-1 sequences from the Los Alamos database. Phylogenetic analyses of these sequences were performed using the neighbor-joining method of the CLUSTAL-X program.18 To prevent the possibility of new sequences obscuring the genetic relationships, unknown sequences were phylogenetically analyzed one at a time. During the final analysis, gaps which were introduced to create the alignment were stripped. The statistical robustness of the neighbor-joining tree and reliability of the branching pattern was confirmed by bootstrapping (1000 replicates).
Statistical analysis of differences in mutation frequencies between genotypes was carried out using the Fisher exact test. The P values <0.05 were considered to be statistically significant.
Study Participants and Clinical Data
This study covered a 2-year period from November 2004 to November 2006. Inclusion of subjects was based on a random selection of HIV-infected patients presenting at CHUYO in Ouagadougou who either showed a viral load >1000 copies/mL or a decrease in CD4 cell count of >50 cells/mm3 after their last visit. During the study period, a total of 290 samples from both drug-naive and drug-exposed subjects were shipped from Ouagadougou to Heidelberg. Resistance profiles could be obtained for 105 samples from drug-naive and 87 from drug-exposed individuals. Twelve of the latter were follow-up samples collected from 11 subjects between 4 and 22 months after the initial sampling. Table 1 summarizes the epidemiological and clinical history of the 75 drug-exposed patients, results for drug-naive patients will be presented elsewhere. Numbers of males and females were roughly equal with a median age of 33 years (range 9-61). Viral loads ranged from 1000 to 4 × 106 copies/mL with a median viral load of 8.2 × 104 copies/mL. The CD4 counts ranged from 2 to 469 cells/mm3 with a median of 104. All patients were either in Center for Disease Control (CDC) stage B (33%) or C (77%). Some of the subjects had started ARV treatment before 2004 when this study began. However, most patients [51 of 75 (68%)] were on their first regimen.
Genotypic Analysis of HIV-1 Strains
The PR and 5′RT HIV coding regions of all 87 samples were analyzed using the Viroseq kit. The genotypic distribution was as follows: 36 (48%) CRF06_cpx, 30 (40%) CRF02_AG, 6 (8%) unclassifiable (U), and 1 each of subtype A, B, and CRF09_cpx (Table 1). Predicted resistance mutations in the PR [for protease inhibitors (PIs)] and 5′RT coding regions [for nucleoside reverse transcriptase inhibitors (NRTIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs)] were classified based on the Stanford Drug Resistance Database16 and are shown according to subtype (CRF06_cpx, CRF02_AG or others) in Table 2. For reference purposes, we also show the frequency of the respective mutations in B subtype-infected subjects exposed to the respective class of drugs and failing therapy from the Stanford database.16 The statistical significance of subtype-specific differences between the 2 predominant subtypes (CRF06_cpx and CRF02_AG) was determined using Fisher exact test and is also depicted in Table 2.
The most prevalent predicted resistance mutation in PR affected position M46, which was changed to Ile, Leu, or Val in 13, 2, and 1 cases (37.2% of total), respectively (Table 2). Other relevant mutations in PR included I54V (11 cases, 25.6%, all but two in CRF06_cpx strains), V82A/T/F (13 cases, 30.2%), I84V (8 cases, 18.6%), and L90M (8 cases, 18.6%). Interestingly, the prevalence of predicted resistance mutations in PR was higher in CRF06_cpx strains for all these positions except for I84V, and this subtype-dependent difference was statistically significant for I54V, V82A/T/F, and L90M, respectively (Table 2).
Among the NRTI resistance mutations, M184V was most common and was observed in 43 patients (57.3%). Also common were the thymidine analogue mutations (TAMs): M41L (37.3%), D67N (34.6%), K70R (14.6%), L210W (34.6%), and T215F/Y (48%). Again, these predicted resistance mutations were more common among the CRF06_cpx strains, and this was statistically significant for the TAMs D67N, K70R, and L210W and for the TAM-associated mutations V75M and V118I (P < 0.05). Multiresistant viruses are commonly associated with the Q151M mutation [causing resistance to azidothymidine (AZT), abacavir, zalcitabine, stavudine (d4T)] or with insertions of several amino acids (mainly Ser) at codon 69 (causing resistance to NRTI and tenofovir). We observed 4 sequences with the Q151M mutation, which were distributed through the subtype groups. No insertions at position 69 were detected.
Among the NNRTI-associated resistance mutations, K103N was most frequent (33 cases, 44.4%). Other common predicted resistance mutations included Y181C/I (12 cases, 16%) and Y188C/L (8 cases, 10.6%). Although no statistically significant difference between the 2 most prevalent subtypes for NNRTI resistance mutations was found, there was a higher prevalence of these 3 resistance mutations among CRF06_cpx strains. The M230L mutation, which causes resistance to the entire NNRTI class, was found in 4 cases (5.3%). For comparison, this mutation is found in 1.3% of subtype B NNRTI-experienced patients (Table 2).
ARV Treatment History and Longitudinal Follow-up
We compared the treatment history of these patients with the prevalence of different mutations. Fifty-one patients (68%) were on their first regimen whereas 24 (32%) were on their second or more regimen (Table 1). The recommended first-choice regimen of ARV in Burkina Faso is AZT, lamivudine (3TC), and efavirenz (EFV). The most frequently used ARVs are NRTI followed by NNRTI (Figs. 1A and 1B). NRTIs were used in the following order of frequency: 3TC>AZT>d4T>ddI>ABC>TDF>FTC (Fig. 1A). PIs are still rare in this country, but there were about 40% of individuals who had received indinavir (IDV) at one point during their treatment history (Fig. 1B). Less than 10% had received liponavir/ritonavir (LPV/r) or nelfinavir (NFV) whereas no patient had been exposed to amprenavir, atazanavir, and saquinavir. The combinations 3TC+AZT+EFV (n = 12); 3TC+AZT+IDV (n = 10); 3TC+AZT+NVP (n = 7); and 3TC+d4T+NVP (n = 7) were the most frequently administered first-line treatments among the patients in our study.
Mutational patterns in individual samples were classified as low, intermediate, or high predicted resistance against individual compounds based on the Stanford Database (Fig. 1). More than 80% of all samples showed predicted resistance against 3TC, abacavir, didanosine, and emtricitabine (FTC), respectively, and the numbers for AZT and d4T were only slightly lower. Some predicted resistance against emtricitabine was detected in 66% of samples (Fig. 1A). High-level resistance was most commonly predicted against 3TC (72%), FTC (69%), and AZT/d4T (45/39%, respectively) but was 5% for Tenofovir (TDF) (Fig. 1A). EFV and NVP are the 2 NNRTIs commonly available in Burkina Faso, and resistance frequencies of ∼80% were also observed for these with 60%-70% of all samples predicted to be highly resistant (Fig. 1B). As expected, there was a high prevalence (68%) of cross-resistance to delavirdine, the third drug in the NNRTI class. Approximately 40% of samples showed some predicted resistance to the PIs (Fig. 1B), although high-level resistance was relatively rare for most compounds with 21% high-level resistance against NFV and ∼10% against IDV.
Figure 2A compares the relative number of mutations associated with resistance to PIs, NNRTIs, and NRTIs in the different HIV-1 subtypes. A major difference between CRF02_AG and CRF06_cpx strains was observed for NRTI, where 69% of all CRF06_cpx samples exhibited >3 NRTI resistance mutations compared with 20% of CRF02_AG samples. This was even more pronounced for the TAMs where the majority of CRF06_cpx strains (25 sequences; 69%) carried ≥3 TAMs, although this was only the case in 3 (10%) of the CRF02_AG viruses. The cumulative single and double mutations (25% and 26.6%), respectively, were similar between both genotypes, however (data not shown). A trend to a higher percentage of CRF06_cpx samples carrying multiple resistance mutations compared with CRF02_AG samples was also observed for resistance against NNRTI and PI but was much less pronounced. In total, 58% of all samples had no predicted resistance mutations against PIs and this number was 73% in CRF02_AG samples. Samples with no predicted resistance mutations were much less frequent for NNRTI (25%) and NRTI (17%), corresponding to the more common use of these drugs. Subtype-dependent differences became also evident when we determined the relative number of samples showing resistance to more than one drug class. Almost 90% of the CRF06_cpx sequences carried major resistance mutations associated with resistance to 2 or all 3 drug classes with 44% of samples carrying at least one major resistance mutation associated with each class. The corresponding numbers for CRF02_AG samples were 52% with mutations associated with resistance to 2 or 3 classes and 26% with at least one resistance-associated mutation against each class (Fig. 2B).
Two or more samples for resistance analysis were available from 36 patients. Seventeen of these responded favorably to treatment (viral load < 40 copies/mL) whereas 6 had not yet started therapy. In 12 cases (from 11 patients), viral loads were above 1000 copies/mL, signifying insufficient suppression of viral replication. The drug exposure, viral load and CD4 count, and drug resistance patterns for these sequential samples are outlined in Table 3. Among the 11 subjects, 6 were infected with CRF06_cpx, 4 with CRF02_AG strains, and 1 with subtype B. All second samples were collected between 4 and 22 months after the first sample and the only third sample was collected after 23 months (Table 3). The NRTIs (3TC and AZT), the NNRTIs (NVP and EFV), and the PIs (IDV and LPV/r) were the most commonly used drugs among this group of follow-up patients. Switching therapy (which in most cases involved the replacement of an NNRTI with a PI within a regimen) did not significantly change the resistance pattern observed in the previous sample. In most cases, this only increased the number of mutations observed. The M184V mutation was the first major drug resistance mutation to develop and among the most common. The Type I TAMs (M41L, L210W, and T215Y) were also common among the follow-up patients and highest for the CRF06_cpx group with 5 of 6 patients carrying these mutations (Table 3). Together with M184V, they persisted despite changes in treatment combination (Table 3). We also observed the complete disappearance of resistance mutations in the population sequence in 1 individual who interrupted treatment about 6 months before sampling (Table 3). In general, most of the patients showed a 1-2 log reduction in viral load and a very moderate increase in CD4 cell counts between 2 sampling points (Table 3).
This study provides information regarding the prevalence and pattern of drug resistance mutations among ARV-exposed patients failing therapy in Ouagadougou, Burkina Faso. NRTIs and NNRTIs constitute the first line of therapy in sub-Saharan Africa,8,18 and the prevalence of predicted resistance mutations in our study corresponded to their respective usage: NRTIs 85%, NNRTIs 76%, and PIs 40%. The relative prevalence of the most common mutations was M184V>T215F/Y>K103N in RT, again corresponding to the frequency of the drugs administered. The percentage of drug-resistant viruses in the current study is high and similar to reports from Malaysia (PI 77.8%, NRTI 52.8%, and NNRTI 63.9%) and India (NRTI 78.8% and NNRTI 80.6%).18,19 High percentages of resistant samples are of course expected given the selection of patients failing therapy, and such numbers therefore need to be interpreted with caution.
The combination 3TC+AZT/d4T+EFV is the first-line therapy in Burkina Faso. Most potential PIs and some NRTIs which could be used as second-choice regimen are still not readily available. Among the different drug classes, amprenavir, atazanavir, and saquinavir (PI); FTC (NRTI); and delavirdine (NNRTI) were never administered to any of the study subjects. However, there was a high prevalence of FTC resistance mutations (82%), which could be expected because of cross-resistance between 3TC and FTC.20 Similarly, cross-resistance to delavirdine was expected due to the common use of other NNRTIs. Interestingly, NFV showed the highest number of predicted high-level resistance of all PIs (21%) even though it was used least. NFV, liponavir, and IDV share some drug resistance patterns, but the effect of specific mutations varies for the different PIs. For example, M46I and L90M cause low-level resistance against either liponavir or IDV but intermediate and high-level resistance, respectively, against NFV,16 therefore making this drug a poor choice for second-line treatment in Burkina Faso.
In the cohort of longitudinal follow-up patients, several NNRTI and NRTI mutations persisted in about 50% of the subjects, approximately 1 year or more after treatment was changed. Switching treatment mostly involved replacing an NNRTI with a PI and maintenance of the NRTIs. Accordingly, most of the TAMs observed on initial failure were maintained after treatment change. This was also reflected by the viral load ranging between 104 and 105 copies/mL, down by approximately 1 order of magnitude from before therapy switch, but clearly not well controlled. This poor response was primarily due to lack of alternative treatment options but highlights the danger of selecting multiple resistance mutations by suboptimal treatment adaptations. Genotypic resistance analysis is clearly needed within the country to complement monitoring of viral load and CD4 counts and improve patient management. This is of particular importance given the wider accessibility of multiple ARV drugs in recent years.
Given the extraordinary degree of genetic diversity of HIV-1, different subtypes and CRFs may evolve in different pathways when they are under selective pressure. Determining any subtype-specific difference in the pattern of resistance mutations would obviously be of major importance for therapy management and for our understanding of resistance mechanisms. Several previous studies have reported subtype-specific differences in the drug resistance mutation patterns,21-26 but none of these studies included CRF06_cpx, which is the second most prevalent CRF in West Africa. In the present study, we found a significant difference in the mutational pattern and the number of resistance mutations observed between CRF06_cpx and CRF02_AG, the most dominant HIV strains in West Africa. These differences primarily affected PI and NRTI resistance mutations belonging to the TAMs, but some differences were also found for NNRTI resistance mutations.
Patients included in this study were randomly selected from all patients presenting at CHUYO, and their sex and age characteristics and disease stage matched well, making it unlikely that the higher frequencies of certain mutations and the higher number of multiclass resistances in CRF06_cpx strains were caused by patient selection. However, effects caused by the duration of treatment cannot be excluded at present. Another possible explanation for the differences would be a higher frequency of preexisting drug resistance-associated polymorphisms in therapy-naive patients, paving the way for more rapid resistance development. This was indeed observed in a previous study, where CRF06_cpx strains were reported to carry more natural drug resistance-associated polymorphisms.12 In our analysis of drug-naive patients presenting at the same service in Ouagadougou, however, no significant strain-specific differences in drug resistance-associated polymorphisms were detected (Tebit et al., manuscript in preparation), making this explanation less likely. Alternatively, differences could be caused by immune pressure if strain-specific HLA epitopes map to resistance positions. However, no known HLA epitopes have been described for the position of the TAMs in either CRF06_cpx or CRF02_AG arguing against this hypothesis. Although the findings of the present report need to be corroborated in larger studies and alternative explanations are not excluded, the results would be most easily explained by differences in the mutational pathways to resistance for CRF02_AG and CRF06_cpx viruses.
A striking difference between CRF02_AG and CRF06_cpx was observed regarding the frequency of several TAMs (D67N, K70R, and L210W) with a statistically not significant trend for other TAMs (M41L, T215F/Y, and K219Q/E). Furthermore, almost two thirds of all CRF06_cpx samples carried more than 3 TAMs whereas the corresponding number for CRF02_AG strains was 10%. TAMs are selected predominantly by the thymidine analogs zidovudine and d4T but cause cross-resistance to other NRTIs. Apart from 1 patient infected with a CRF06_cpx virus, all patients in this study had received AZT or d4T at some point during their treatment history. Novitsky et al26 recently observed that subtype C viruses in Botswana used a unique pathway of resistance with mutations 67N, 70R, and 215Y but wild-type codons at positions 41, 210, and 219 during treatment failure. This pathway is different from both TAMs 1 and 2 observed in subtype B.27 The V118I mutation, which contributes to NRTI resistance and is specifically known as a TAM-associated mutation, was also found significantly more often in CRF06_cpx viruses in our study. This mutation is not common in CRF06_cpx-naive subjects27 and may play a role in resistance development for this strain. A recent study suggested that the V118I mutation was the only independent genotypic predictor of therapy failure in a d4T-containing regimen.28
In summary, this study highlights a high prevalence of multicross class drug resistance mutations among patients failing therapy in Burkina Faso. To date, this is the largest study concerning the pattern of drug-resistant mutations in CRF06_cpx-infected patients failing therapy. Furthermore, no drug resistance studies have been performed with HIV-1 subtype K strains, which constitute a major part of the CRF06_cpx pol region. Larger prospective studies and in vitro drug susceptibility and fitness experiments are required to test our preliminary observation that TAMs and some other PI and NNRTI mutations might develop more rapidly and are therefore more prevalent among CRF06_cpx-infected individuals. Such differences would be clinically important because CRF06_cpx viruses are prevalent not only in Burkina Faso but also in the neighboring countries Mali and Niger and in other West African countries.
2. Koenig SP, Kuritzkes DR, Hirsch MS, et al. Monitoring HIV treatment in developing countries. BMJ
3. Petrella M, Brenner B, Loemba H, et al. HIV drug resistance and implications for the introduction of antiretroviral therapy in resource-poor countries. Drug Resist Updat
4. Adje C, Cheingsong R, Roels TH, et al. High prevalence of genotypic and phenotypic HIV-1 drug-resistant strains among patients receiving antiretroviral therapy in Abidjan, Cote d'Ivoire. J Acquir Immune Defic Syndr
5. Toni T, Masquelier B, Bonard D, et al. Primary HIV-1 drug resistance in Abidjan (Côte d'Ivoire): a genotypic and phenotypic study. AIDS
6. Kantor R, Zijenah LS, Shafer RW, et al. HIV-1 subtype C reverse transcriptase and protease genotypes in Zimbabwean patients failing antiretroviral therapy. AIDS Res Hum Retroviruses
7. Richard N, Juntilla M, Abraha A, et al. High prevalence of antiretroviral resistance in treated Ugandans infected with non-subtype B human immunodeficiency virus type 1. AIDS Res Hum Retroviruses
8. Laurent C, Kouanfack C, Vergne L, et al. Antiretroviral drug resistance and routine therapy, Cameroon. Emerg Infect Dis
9. Chaix ML, Rouet F, Kouakoussui KA, et al. Genotypic human immunodeficiency virus type 1 drug resistance in highly active antiretroviral therapy-treated children in Abidjan, Côte d'Ivoire. Pediatr Infect Dis J
10. Arrivé E, Newell ML, Ekouevi DK, et al. Prevalence of resistance to nevirapine in mothers and children after single-dose exposure to prevent vertical transmission of HIV-1: a meta-analysis. Int J Epidemiol
11. Koizumi Y, Ndembi N, Miyashita M, et al. Emergence of antiretroviral therapy resistance-associated primary mutations among drug-naive HIV-1-infected individuals in rural western Cameroon. J Acquir Immune Defic Syndr
12. Ojesina AI, Sankale JL, Odaibo G, et al. Subtype-specific patterns in HIV Type 1 reverse transcriptase and protease in Oyo State, Nigeria: implications for drug resistance and host response. AIDS Res Hum Retroviruses
13. Vergne L, Diagbouga S, Kouanfack C, et al. HIV-1 drug-resistance mutations among newly diagnosed patients before scaling-up programmes in Burkina Faso and Cameroon. Antivir Ther
14. Toni T, Masquelier B, Minga A, et al. HIV-1 antiretroviral drug resistance in recently infected patients in Abidjan, Côte d'Ivoire: a 4-year survey, 2002-2006. AIDS Res Hum Retroviruses
17. Thompson JD, Gibson TJ, Plewniak F, et al. The CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res
18. Tee KK, Kamarulzaman A, Ng KP. Prevalence and pattern of drug resistance mutations among antiretroviral-treated HIV-1 patients with suboptimal virological response in Malaysia. Med Microbiol Immunol
19. Sen S, Tripathy SP, Patil AA, et al. High prevalence of human immunodeficiency virus type 1 drug resistance mutations in antiretroviral treatment-experienced patients from Pune, India. AIDS Res Hum Retroviruses
20. Borroto-Esoda K, Parkin N, Miller MD. A comparison of the phenotypic susceptibility profiles of emtricitabine and lamivudine. Antivir Chem Chemother
21. Turner D, Schapiro JM, Brenner BG, et al. The influence of protease inhibitor resistance profiles on selection of HIV therapy in treatment-naive patients. Antivir Ther
22. Carobene MG, Rubio AE, Carrillo MG, et al. Differences in frequencies of drug resistance-associated mutations in the HIV-1 pol
gene of B subtype and BF intersubtype recombinant samples. J Acquir Immune Defic Syndr
23. Dumans AT, Soares MA, Machado ES, et al. Synonymous genetic polymorphisms within Brazilian human immunodeficiency virus type 1 subtypes may influence mutational routes to drug resistance. J Infect Dis
24. Gupta RK, Chrystie IL, O'Shea S, et al. K65R and Y181C are less prevalent in HAART-experienced HIV-1 subtype A patients. AIDS
25. Kantor R, Katzenstein DA, Efron B, et al. Impact of HIV-1 subtype and antiretroviral therapy on protease and reverse transcriptase genotype:results of a global collaboration. PLoS Med
26. Novitsky V, Wester CW, DeGruttola V, et al. The reverse transcriptase 67N 70R 215Y genotype is the predominant TAM pathway associated with virologic failure among HIV type 1C-infected adults treated with ZDV/ddI-containing HAART in southern Africa. AIDS Res Hum Retroviruses
27. Grossman Z, Vardinon N, Chemtob D, et al. Genotypic variation of HIV-1 reverse transcriptase and protease: comparative analysis of clade C and clade B. AIDS
28. Gianotti N, Galli L, Boeri E, et al. The 118I reverse transcriptase mutation is the only independent genotypic predictor of virologic failure to a stavudine-containing salvage therapy in HIV-1-infected patients. J Acquir Immune Defic Syndr