Adherence to antiretroviral therapy (ART) has been shown to be a strong determinant of virologic suppression in HIV-infected patients.1,2 Adherence is the most important predictor of success on ART and prevents the development of resistance in the long term.3-5 HIV care providers worldwide face the challenges of identifying those patients who do not take their medications regularly in the first few months after ART initiation and improving their adherence.6,7
There is no easily measurable gold standard for detecting poor adherence.8,9 However, estimating adherence is particularly difficult in resource-limited countries, for 2 reasons.
First, plasma HIV-1 RNA tests, which can identify patients who do not take their medication early in treatment, are not routinely available in many settings. In contrast, in high-income countries HIV-1 RNA is measured on average every 3 months, and patients who have detectable HIV-1 RNA in the first months of ART are targeted for adherence interventions.10
Second, even when patients are willing to take their medications regularly, they may not receive them on time. In some settings, external obstacles to drug delivery can be as common as the intrinsic factors associated with nonadherence.2,6,11-13 These logistical obstacles make it more complicated to monitor adherence, both because they cannot always be distinguished from patients' reluctance to pick up their drugs and because focusing on adherence may not be productive when the real problem lies with the drug delivery system.
In this context, a method for determining the amount of medications a patient has received can serve to simultaneously measure the health care system's ability to provide medications and some aspects of adherence. In contrast to pill intake, medication delivery can be measured accurately, even in low-resource settings, where an increasing number of HIV care centers use computerized prescription databases.14-20 These databases can be used to measure pill possession on an individual basis, provided they are prospectively managed through standardized procedures.21
In 2006, we launched a prospective cohort study of HIV-infected adults who initiated ART at 3 HIV care centers equipped with computerized prescription databases in Abidjan, the economic capital of Côte d'Ivoire, west Africa. Our objective was to describe the association between medication possession and virologic outcomes in the first year of ART.
The VOLTART Cohort
We conducted a prospective cohort study of long-term virologic outcomes on ART (VOLTART cohort). HIV-infected adults who started ART between February 2006 and May 2007 at one of 3 HIV outpatient clinics in Abidjan and showed up for their 6-month visit were eligible for the study. The 3 clinics were the Centre de Prise en Charge et de Formation (CePReF),22 the Yopougon Attié clinic, and the HIV care center affiliated with the National Center for Blood Transfusion (CNTS).23 Study subjects received the same standard care and treatment as other HIV-infected patients on ART at their respective clinics. In addition, they received free plasma HIV-1 RNA and genotype tests every 6 months. In each study center, a research coordinator prospectively managed and monitored the prescription database.
Standard Care and Treatment
The standard of HIV care for HIV-infected adults on ART in Côte d'Ivoire has been described elsewhere.22 During the study period, all patients followed in the 3 study clinics initiated ART according to the 2003-2006 World Health Organization (WHO) criteria: WHO clinical stage 4 regardless of CD4 count, CD4 count ≤200 per cubic millimeter regardless of WHO stage, or WHO stage 3 and CD4 count 200-350 per cubic millimeter.24 Chemistry examinations (serum creatinine and transaminases) were performed before ART initiation. When patients were HIV-1 infected, first-line ART consisted of 2 nucleoside reverse transcriptase inhibitors (NRTIs) and 1 nonnucleoside reverse transcriptase inhibitor (NNRTI). When patients were HIV-2 or HIV-1 and HIV-2 infected, first-line ART consisted of 2 NRTIs and 1 protease inhibitor. CD4 counts and complete blood counts were measured every 6 months. Patients paid a fixed rate of US $2 per month for antiretroviral drugs and laboratory tests until August 2008, when the national HIV program made them available for free. All patients with CD4 counts ≤500 per cubic millimeter were also given cotrimoxazole prophylaxis. Isoniazid prophylaxis was not recommended, as it is not part of the national treatment guidelines. Support groups were organized to encourage patients to adhere to therapy, and a community-based team made telephone calls or home visits when patients did not show up for clinic visits or to pick up antiretroviral drugs.25
Monitoring and Tracking System
All 3 study centers used the same standardized forms to record the following variables at routine visits: (1) initial visit: date, sex, date of birth (or age), height, weight, and HIV type (HIV-1, HIV-2, or both); (2) follow-up visit: date and weight; (3) ART initiation visit: date, WHO clinical stage, and weight; (4) drug prescription (antiretroviral or other): date, name, and quantity of drugs delivered; (5) CD4 count and complete blood count measurement: date, CD4 count, CD4 percentage, hemoglobin level, platelet count, and granulocyte and leukocyte counts; (6) telephone call and home visit: dates at which patients were contacted, and vital status on that date; (7) patients known to have died: date of death.
The care provided to patients who agreed to participate in the study differed from that provided to other patients at the same clinic in 3 ways: (1) plasma HIV-1 RNA tests were performed every 6 months (French National Agency for Research on AIDS and Viral Hepatitis real-time PCR; Biocentric, Bandol, France; threshold for detection, 300 copies/mL)26; (2) if HIV RNA exceeded 300 copies per milliliter, we used the French National Agency for Research on AIDS and Viral Hepatitis consensus technique27 to perform an automated population full-sequence analysis of the reverse transcriptase and protease genes and interpreted genotype results using the French resistance algorithm (www.hivfrenchresistance.org); (3) a research coordinator was devoted to monitoring and managing the cohort data and helping track patients by telephone and/or home visit. Genotype tests were performed in the virology laboratory of the Necker hospital in Paris, France. This laboratory undergoes an annual external quality assurance evaluation.28
Patients were defined as lost to follow up if (1) their last contact with study team was <month 12; (2) they were not known to be dead or transferred out before month 12; (3) no further information on their vital status could be obtained within the 6 months after study endpoint (ie, between month 12 and month 18).
The medication possession ratio (MPR) was defined as the number of daily doses of antiretroviral drugs dispensed by the pharmacy to each patient, divided by that patient's total follow-up time in days since ART initiation. We used generalized logistic regression analysis to estimate the association between MPR from baseline to month 12 and (1) virologic failure and wild-type HIV-1, defined as HIV-1 RNA ≥300 copies per milliliter at month 12 and no resistance mutations or (2) virologic failure and resistance, defined as HIV-1 RNA ≥300 copies per milliliter at month 12 and at least 1 resistance mutation. Results were adjusted for sex and baseline CD4 count. Analyses were performed with SAS software, version 9.1 (SAS Institute, Inc, Cary, NC).
Overall, 1573 adults started ART at the 3 study centers between February 2006 and May 2007. At month 6, 1206 patients were still alive and in care. Of these, 996 gave written informed consent to participate in the study. Median age was 36 years [interquartile range (IQR), 30-43 years] and 747 patients (75%) were women. At month 6 and month 12, 100% and 98% of study subjects had an HIV-1 RNA test done (Fig. 1).
Table 1 presents the main characteristics of the study participants at ART initiation, month 6, and month 12. The 210 patients who were in care at month 6 but refused to participate in the study differed significantly from the 996 patients who enrolled with respect to several baseline and follow-up characteristics. At baseline, patients not included were more likely than patients who enrolled to be men (32% vs. 25%; P = 0.05), less likely to receive d4T-3TC-NVP (48% vs. 58%) and more likely to receive ZDV-3TC-EFV (29% vs. 26%), d4T-3TC-EFV (10% vs. 8%), or other regimens (13% vs. 9%) (P = 0.02). We did not find a significant difference in clinical stage (P = 0.20), CD4 count (P = 0.43), age (P = 0.29), or body mass index (P = 0.62) distributions. During follow-up, patients not included were more likely to die (6.2% vs. 1.3%; P < 0.001), transfer out (3.3% vs. 0.7%; P < 0.005), or be lost to follow-up (18.1% vs. 3.3%; P < 0.001) before month 12 than patients who enrolled.
Plasma HIV-1 RNA
We obtained HIV-1 RNA test results for 996 patients at month 6 and 925 patients at month 12. At month 6, 80% of patients had undetectable viral load (VL) and 20% had HIV-1 RNA ≥300 copies per milliliter, including 14% with HIV-1 RNA ≥1000 copies per milliliter. At month 12, 75% of patients had undetectable VL and 25% had HIV-1 RNA ≥300 copies per milliliter, including 21% with HIV-1 RNA ≥1000 copies per milliliter (Table 1).
Of the 168 patients who had HIV-1 RNA ≥300 copies per milliliter at month 6 and HIV-1 RNA test results at month 12, 44% had undetectable VL at month 12. Of the 757 patients who had undetectable VL at month 6 and had HIV-1 RNA test results at month 12, 18% had HIV-1 RNA ≥300 copies per milliliter at month 12. Of the 232 patients who had HIV-1 RNA ≥300 copies per milliliter at month 12, 40% already had HIV-1 RNA ≥300 copies per milliliter at month 6 and thus remained on a failed ART regimen for more than 6 months. The other 60% patients had undetectable VL at month 6 and thus remained on failed ART for fewer than 6 months (Table 2).
Resistance to Antiretroviral Medications
At month 6, 69 patients were infected with HIV-1 strains that had at least 1 resistance mutation. These patients represented 7% of the 996 patients study subjects, 35% of the 197 patients who had detectable HIV-1 RNA, and 40% of the 173 patients who had available genotype test results. Of the 69 resistant virus, all but one (99%) were resistant to lamivudine/emtricitabine and/or nevirapine/efavirenz (lamivudine/emtricitabine only, n = 4; nevirapine/efavirenz only, n = 28; both n = 36). Only one of these strains had thymidine analogue mutations (TAMs), accounting for 0.1% of the 996 study subjects and for 1% of the 69 patients harboring a virus with at least 1 mutation (see Table 1A, Supplemental Digital Content 1, http://links.lww.com/QAI/A118).
At month 12, 106 patients were infected with HIV strains that had at least 1 resistance mutation. These patients represented 11% of the 925 study subjects who were still in active follow-up, 46% of the 232 patients who had detectable HIV RNA, and 51% of the 207 patients who had available genotype test results. Of the 106 resistant virus, all but one (99%) were resistant to lamivudine/emtricitabine and/or nevirapine/efavirenz (lamivudine/emtricitabine only, n = 8; nevirapine/efavirenz only, n = 33; both n = 64). Eight patients had a virus with TAMs, accounting for 0.9% of the 925 patients included in the study and for 7.5% of the 106 patients harboring a virus with at least 1 mutation (see Table 2A, Supplemental Digital Content 1, http://links.lww.com/QAI/A118). No virus was found resistant to the new generation of NNRTIs (etravirine) either at month 6 or at month 12.
Of the 11% of patients (n = 106/925) who were found to have at least 1 resistance mutation at month 12, 4.5% (n = 42/925) already had at least 1 resistance mutation at month 6, 1.7% (n = 15/925) had detectable HIV-1 RNA but no mutations at month 6, and 5.2% (n = 48/925) had undetectable HIV-1 RNA at month 6 (Table 2). Finally, among the 8 patients who had TAMs at month 12, 1 (12.5%) had undetectable HIV-1 RNA at month 6.
Medication Possession Ratio
Median MPR was 95% (IQR, 81%-99%; range, 0%-161%) from ART initiation to month 6 and 90% (IQR, 78%-98%; range, 0%-139%) from ART initiation to month 12. As MPR increased, the proportion of patients with HIV-1 RNA ≥300 copies per milliliter regardless of resistance decreased.
At month 6, 62% of patients whose MPR was <50% had detectable HIV-1 RNA, of whom 20% had at least 1 resistance mutation. These proportions decreased to 9% and 2% when MPRs were >95% (Fig. 2A). At month 12, 85% of patients whose MPR was <50% had detectable HIV-1 RNA, of whom 37% had at least 1 resistance mutation. These percentages were 9% and 4% when MPRs were >95% (Fig. 2B).
At month 6, the median MPR among patients with undetectable HIV-1 RNA was 96% (IQR, 86%-100%) and was significantly higher than for patients with detectable HIV-1 RNA and no resistance mutations (median, 77%; IQR, 53%-96%; P < 0.0001) and for patients with both detectable HIV-1 RNA and resistance mutations (median, 75%; IQR, 61%-86%; P < 0.0001). Among patients with detectable HIV-1 RNA, there was no significant difference in MPR distribution between patients with and without resistance (Fig. 3A). In patients with detectable VL at month 6, the median MPR increased with decreasing levels of VL (median MPR, 0.91, 0.77, 0.73, and 0.64 in patients with HIV-1 RNA <3 log10 copies/mL, 3-3.9 log10 copies/mL, 4-4.9 log10 copies/mL, and ≥5 log10 copies/mL, respectively; P < 0.0001).
At month 12, the median MPR for patients with undetectable HIV-1 RNA was 94% (IQR, 86%-99%) and again was significantly higher than for patients with detectable HIV-1 RNA and no resistance mutations (median, 73%; IQR, 55%-88%; P < 0.0001) and for patients with both detectable HIV-1 RNA and resistance mutations (median, 71%; IQR, 55%-84%; P < 0.0001). Among patients with detectable HIV-1 RNA at month 12, there was no significant difference between patients with and without resistance (Fig. 3B). In patients with detectable VL at month 12, the median MPR increased with decreasing levels of VL (median MPR, 0.89, 0.69, 0.74, and 0.65 in patients with HIV-1 RNA <3 log10 copies/mL, 3-3.9 log10 copies/mL, 4-4.9 log10 copies/mL, and ≥5 log10 copies/mL, respectively; P < 0.0001).
In multivariate analysis, MPR was associated with both detectable HIV-1 RNA with wild-type virus and detectable HIV-1 RNA with at least 1 resistance mutation at month 12, after adjusting for baseline CD4 count and sex (Table 3). There was no interaction between regimen and MPR for the association with virologic status (P = 0.98).
We conducted a prospective cohort study nested in 3 large HIV care programs in Abidjan, Côte d'Ivoire. The study participants received the same care and treatment as nonparticipants at the same HIV care centers. In addition to providing regular follow-up, they also had HIV-1 RNA done every 6 months, with genotype resistance tests when HIV-1 RNA was detectable. We recorded virologic and genotype data on these patients in their first year of ART. We had 4 main findings.
First, 20% and 25% of patients had plasma HIV-1 RNA ≥300 copies at month 6 and month 12, respectively. The percentage of patients infected with resistant HIV-1 strains was 7% at month 6 and 11% at month 12. These rates are at the lower bound of estimates reported in others studies on patients on ART in sub-Saharan Africa.29-36
Second, we found that viral suppression was a dynamic process. In other words, virologic success at month 6 did not guarantee virologic success at month 12. Among patients who had detectable HIV-1 RNA at month 12, 60% had undetectable VL at month 6. It is likely that these patients did not adhere to their medications well enough to remain undetectable beyond month 6.
Third, 49% of the participants who had detectable HIV-1 RNA and available genotype test results at month 12 had no resistance mutations. Preventing resistance from developing is clearly one important objective in the early phases of HIV treatment. This very large proportion of patients with measurable HIV-1 RNA levels, but no resistance, suggests that challenges for adherence persist through month 12. A routine HIV-1 RNA test at month 12 could thus help identify many patients who might benefit from an intervention to improve adherence and prevent resistance. However, it also indicates that half of those patients with detectable HIV-1 RNA at month 12 already had resistant HIV strains and that HIV-1 RNA testing in the absence of genotype testing, therefore, cannot differentiate between patients who should switch to second-line ART due to resistance and patients who could still benefit from improved adherence to first-line ART.
Fourth, as expected, most of the resistance mutations that developed in the first year of NNRTI-based first-line ART were to lamivudine/emtricitabine and/or to nevirapine/efavirenz. TAMs were very rare at month 6, but their prevalence increased to 7.5% of all resistant strains and 0.9% of the overall study population at month 12. This trend is important because it illustrates the difficulty of targeting patients who are resistant to lamivudine/emtricitabine and/or nevirapine/efavirenz and then switching them to a second-line regimen before they develop resistance to other NRTIs. This goal is not achievable in settings where genotype tests are not available, and even less so when HIV-1 RNA tests are not available for routine care. Furthermore, one patient in our study who had TAMs at month 12 had undetectable HIV RNA at month 6, suggesting that even if genotype tests were available, a genotype test 6 months after ART initiation would not be sufficient. Of note, very few patients had virus resistant to abacavir, ddI, or tenofovir at month 12, and none had resistance to etravirine.
This study has several limitations. First, we did not have genotype test results before ART initiation. Virologic resistance at 6 months may in part be due to primary resistance.31,37 However, the prevalence of resistance mutations in adults with recent HIV infection in Côte d'Ivoire was recently estimated to be lower than 5%.38 Second, some data are missing because 13% of the patients who were in active follow-up at month 6 refused to enroll in the study, 2% of the patients who enrolled at month 6 did not receive HIV RNA measurements at month 12, and 11% of patients who had detectable HIV RNA did not have available genotype test results. As compared with patients who enrolled, those who refused to enroll were less likely to receive d4T-3TC-NVP and had higher rates of death and loss to follow-up between month 6 and month 12. Third, the threshold of HIV RNA detectability was ≥300 copies per milliliter, which may have led us to underestimate the proportion of patients who were not completely virologically suppressed and thus the proportion of patients with resistance. Fourth, nearly one third of the patients started on ART in these programs either died, transferred out, or were lost to follow-up before month 6. However, their outcomes were likely worse than those of patients who enrolled in the study, and the association we found between MPR and virologic outcomes in patients who enrolled is likely to also exist in patients who did not.
In conclusion, in these patients still alive and in care after 6 months of ART, MPR was strongly associated with virologic outcomes during the first year on ART, and half of patients with detectable plasma HIV-1 RNA at month 12 had no resistance mutations. In the absence of HIV RNA tests, MPR should be used at month 6 to identify patients who might benefit from early interventions to reinforce adherence.
1. Chene G, Sterne JA, May M, et al. Prognostic importance of initial response in HIV-1 infected patients starting potent antiretroviral therapy: analysis of prospective studies. Lancet
2. Fielding KL, Charalambous S, Stenson AL, et al. Risk factors for poor virological outcome at 12 months in a workplace-based antiretroviral therapy programme in South Africa: a cohort study. BMC Infect Dis
3. Hanson DL, Adje-Toure C, Talla-Nzussouo N, et al. HIV type 1 drug resistance in adults receiving highly active antiretroviral therapy in Abidjan, Cote d'Ivoire. AIDS Res Hum Retroviruses
4. Bangsberg DR. Preventing HIV antiretroviral resistance through better monitoring of treatment adherence. J Infect Dis
. 2008;197(Suppl 3):S272-S278.
5. Danel C, Gabillard D, Inwoley A, et al. Medium-term probability of success of antiretroviral treatment after early warning signs of treatment failure in West African adults. AIDS Res Hum Retroviruses
6. Carrieri P, Cailleton V, Le Moing V, et al. The dynamic of adherence to highly active antiretroviral therapy: results from the French National APROCO cohort. J Acquir Immune Defic Syndr
7. Rueda S, Park-Wyllie LY, Bayoumi AM, et al. Patient support and education for promoting adherence to highly active antiretroviral therapy for HIV/AIDS. Cochrane Database Syst Rev
8. Miller LG, Hays RD. Adherence to combination antiretroviral therapy: synthesis of the literature and clinical implications. AIDS Read
9. Kouanfack C, Laurent C, Peytavin G, et al. Adherence to antiretroviral therapy assessed by drug level monitoring and self-report in cameroon. J Acquir Immune Defic Syndr
10. Kredo T, Van der Walt JS, Siegfried N, et al. Therapeutic drug monitoring of antiretrovirals for people with HIV. Cochrane Database Syst Rev
11. Diabate S, Alary M, Koffi CK. Determinants of adherence to highly active antiretroviral therapy among HIV-1-infected patients in Cote d'Ivoire. AIDS
12. Ramadhani HO, Thielman NM, Landman KZ, et al. Predictors of incomplete adherence, virologic failure, and antiviral drug resistance among HIV-infected adults receiving antiretroviral therapy in Tanzania. Clin Infect Dis
13. Eholie SP, Tanon A, Polneau S, et al. Field adherence to highly active antiretroviral therapy in HIV-infected adults in Abidjan, Cote d'Ivoire. J Acquir Immune Defic Syndr
14. Bisson GP, Gross R, Bellamy S, et al. Pharmacy refill adherence compared with CD4 count changes for monitoring HIV-infected adults on antiretroviral therapy. PLoS Med
15. Kitahata MM, Reed SD, Dillingham PW, et al. Pharmacy-based assessment of adherence to HAART predicts virologic and immunologic treatment response and clinical progression to AIDS and death. Int J STD AIDS
16. Nachega JB, Hislop M, Dowdy DW, et al. Adherence to highly active antiretroviral therapy assessed by pharmacy claims predicts survival in HIV-infected South African adults. J Acquir Immune Defic Syndr
17. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol
18. Goldman JD, Cantrell RA, Mulenga LB, et al. Simple adherence assessments to predict virologic failure among HIV-infected adults with discordant immunologic and clinical responses to antiretroviral therapy. AIDS Res Hum Retroviruses
19. Nachega JB, Hislop M, Dowdy DW, et al. Adherence to nonnucleoside reverse transcriptase inhibitor-based HIV therapy and virologic outcomes. Ann Intern Med
20. Gross R, Yip B, Lo Re V III, et al. A simple, dynamic measure of antiretroviral therapy adherence predicts failure to maintain HIV-1 suppression. J Infect Dis
21. Acri T, Tenhave TR, Chapman JC, et al. Lack of association between retrospectively collected pharmacy refill data and electronic drug monitoring of antiretroviral adherence. AIDS Behav
22. Messou E, Anglaret X, Duvignac J, et al. Antiretroviral treatment changes in adults from Cote d'Ivoire: the roles of tuberculosis and pregnancy. AIDS
23. Minga A, Danel C, Abo Y, et al. Progression to WHO criteria for antiretroviral therapy in a 7-year cohort of adult HIV-1 seroconverters in Abidjan, Cote d'Ivoire. Bull World Health Organ
25. Toure S, Kouadio B, Seyler C, et al. Rapid scaling-up of antiretroviral therapy in 10,000 adults in Cote d'Ivoire: 2-year outcomes and determinants. AIDS
26. Rouet F, Ekouevi DK, Chaix ML, et al. Transfer and evaluation of an automated, low-cost real-time reverse transcription-PCR test for diagnosis and monitoring of human immunodeficiency virus type 1 infection in a West African resource-limited setting. J Clin Microbiol
27. Pasquier C, Millot N, Njouom R, et al. HIV-1 subtyping using phylogenetic analysis of pol gene sequences. J Virol Methods
28. Descamps D, Delaugerre C, Masquelier B, et al. Repeated HIV-1 resistance genotyping external quality assessments improve virology laboratory performance. J Med Virol
29. Hamers RL, Derdelinckx I, van Vugt M, et al. The status of HIV-1 resistance to antiretroviral drugs in sub-Saharan Africa. Antiviral Ther
30. Vergne L, Kane CT, Laurent C, et al. Low rate of genotypic HIV-1 drug-resistant strains in the Senegalese government initiative of access to antiretroviral therapy. AIDS
. 2003;17(Suppl 3):S31-S38.
31. Adje-Toure C, Celestin B, Hanson D, et al. Prevalence of genotypic and phenotypic HIV-1 drug-resistant strains among patients who have rebound in viral load while receiving antiretroviral therapy in the UNAIDS-Drug Access Initiative in Abidjan, Cote d'Ivoire. AIDS
. 2003;17(Suppl 3):S23-S29.
32. 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
33. Barth RE, van der Loeff MF, Schuurman R, et al. Virological follow-up of adult patients in antiretroviral treatment programmes in sub-Saharan Africa: a systematic review. Lancet
. Infect Dis
34. Bile EC, Adje-Toure C, Borget MY, et al. Performance of drug-resistance genotypic assays among HIV-1 infected patients with predominantly CRF02_AG strains of HIV-1 in Abidjan, Cote d'Ivoire. J Clin Virol
35. Johannessen A, Naman E, Kivuyo SL, et al. Virological efficacy and emergence of drug resistance in adults on antiretroviral treatment in rural Tanzania. BMC Infect Dis
36. Seyler C, Adje-Toure C, Messou E, et al. Impact of genotypic drug resistance mutations on clinical and immunological outcomes in HIV-infected adults on HAART in West Africa. AIDS
37. Toni T, Masquelier B, Minga A, et al. HIV-1 antiretroviral drug resistance in recently infected patients in Abidjan, Cote d'Ivoire: A 4-year survey, 2002-2006. AIDS Res Hum Retroviruses
38. Ayouba A, Lien TT, Nouhin J, et al. Low prevalence of HIV type 1 drug resistance mutations in untreated, recently infected patients from Burkina Faso, Cote d'Ivoire, Senegal, Thailand, and Vietnam: the ANRS 12134 study. AIDS Res Hum Retroviruses