Abstract

Background. Human immunodeficiency virus type 1 (HIV-1) drug resistance may limit the benefits of antiretroviral therapy (ART). This cohort study examined patterns of drug-resistance mutations (DRMs) in individuals with virological failure on first-line ART at 13 clinical sites in 6 African countries and predicted their impact on second-line drug susceptibility.

Methods. A total of 2588 antiretroviral-naive individuals initiated ART consisting of different nucleoside reverse transcriptase inhibitor (NRTI) backbones (zidovudine, stavudine, tenofovir, or abacavir, plus lamivudine or emtricitabine) with either efavirenz or nevirapine. Population sequencing after 12 months of ART was retrospectively performed if HIV RNA was >1000 copies/mL. The 2010 International Antiviral Society–USA list was used to score major DRMs. The Stanford algorithm was used to predict drug susceptibility.

Results. HIV-1 sequences were generated for 142 participants who virologically failed ART, of whom 70% carried ≥1 DRM and 49% had dual-class resistance, with an average of 2.4 DRMs per sequence (range, 1–8). The most common DRMs were M184V (53.5%), K103N (28.9%), Y181C (15.5%), and G190A (14.1%). Thymidine analogue mutations were present in 8.5%. K65R was frequently selected by stavudine (15.0%) or tenofovir (27.7%). Among participants with ≥1 DRM, HIV-1 susceptibility was reduced in 93% for efavirenz/nevirapine, in 81% for lamivudine/emtricitabine, in 59% for etravirine/rilpivirine, in 27% for tenofovir, in 18% for stavudine, and in 10% for zidovudine.

Conclusions. Early failure detection limited the accumulation of resistance. After stavudine failure in African populations, zidovudine rather than tenofovir may be preferred in second-line ART. Strategies to prevent HIV-1 resistance are a global priority.

The rapid scale-up of access to combination antiretroviral therapy (ART) for human immunodeficiency virus type 1 (HIV-1)–infected persons in sub-Saharan Africa during the past decade, through a World Health Organization (WHO)–recommended public health approach [1], has dramatically reduced HIV-related mortality [2]. However, the widespread use of HIV clinical staging and, if available, CD4 cell counts to diagnose ART failure in resource-limited settings, rather than routine virological monitoring, is associated with the accumulation of HIV-1 drug-resistance mutations (DRMs) [3–5], which may limit subsequent drug options and constitutes a source for onward transmission. Studies from the region have reported high levels of drug resistance in individuals with prolonged first-line ART failure, including complex nucleoside reverse transcriptase inhibitor (NRTI) resistance profiles, such as K65R, Q151M, and thymidine analogue mutations (TAMs), in addition to highly prevalent M184V and nonnucleoside reverse transcriptase inhibitor (NNRTI) mutations [4–8].

Most reports on HIV-1 drug resistance deal with subtype B infections in developed countries. There is limited knowledge of resistance pathways in the different HIV-1 non-B subtypes and their clinical relevance [9], despite the fact that >90% of HIV-1 infections globally belong to non-B subtype variants [10].

The PASER-Monitoring (PASER-M) study is a prospective cohort of HIV-1–infected individuals from 13 clinical sites in 6 sub-Saharan African countries who initiated first-line ART including different NRTI backbones (zidovudine [37%], tenofovir [34%], stavudine [27%], or abacavir [3%]) and NNRTIs efavirenz [60%] or nevirapine [40%]) in accordance with national guidelines [11]. The present study assessed HIV-1 DRMs in participants with virological failure of different first-line regimens and predicted viral drug susceptibility to gain insight about the optimal strategies for second-line therapy.

METHODS

Study Population

The PASER-M study includes clinical sites in Kenya (2), Nigeria (1), South Africa (3), Uganda (3), Zambia (3), and Zimbabwe (1) [11]. We previously reported the participants’ baseline resistance profiles [12] and the effect of pretreatment drug resistance on the immunological and virological patient outcomes after the first year of ART [13].

For the present analysis, individuals were included if they were aged ≥18 years with HIV-1 infection and had initiated standard first-line NNRTI-based ART in accordance with national guidelines—that is, advanced immunodeficiency (CD4 cell count <200 cells/μL) or advanced HIV disease (WHO clinical stage 3 or 4) [13]. Individuals who were previously exposed to any antiretroviral drugs for prevention and/or treatment were excluded. Other exclusion criteria were pregnancy at study screening or, for those screened in Nigeria only, HIV type 2 coinfection.

Participants provided written informed consent at enrollment. The study protocol was approved by the appropriate national and local research ethics committees at all collaborating sites and the Academic Medical Center of the University of Amsterdam in the Netherlands.

Data Collection

Participants were followed up in accordance with local standard-of-care guidelines. A single-drug substitution, due to toxicity or intolerance, was not considered a regimen switch. Plasma was collected at the baseline visit and after 12 months of ART (time window, 11–15 months) and was stored for retrospective assessment of HIV RNA and genotypic drug resistance. Virologic failure was defined as a plasma HIV RNA value of ≥400 copies/mL. Participants who were switched to second-line ART earlier than month 12 (owing to locally diagnosed ART failure) were not included in the month 12 summary statistics.

Virological Analysis

All virological testing was conducted at 1 of 2 reference laboratories in South Africa or Uganda, as previously described [12]. In brief, HIV RNA was determined using the NucliSens EasyQ real-time assay, version 2.0 (bioMérieux, Lyon, France) or COBAS Ampliprep/COBAS TaqMan assay (Roche, Branchburg, New Jersey). Population-based genotyping of HIV-1 protease and reverse transcriptase was undertaken in specimens with HIV RNA of >1000 copies/mL, using in-house sequencing methods. HIV-1 subtypes were inferred from the pol sequences using the STAR algorithm [14] and, if required, the REGA tool, version 2.0 [15]. Genotypic drug resistance was defined as the presence of ≥1 major amino acid substitution included in the International Antiviral Society–USA mutation list of December 2010 [16]. Drug classes considered were NRTIs, NNRTIs, and protease inhibitors (PIs). HIV-1 drug susceptibility for each participant was predicted using the Stanford algorithm, version 6.1.0 [17] and was categorized as susceptible, potential low-level resistance, low-level resistance, intermediate resistance, or high-level resistance. All sequences have been deposited in GenBank (accession numbers JQ480157-JQ480298).

Statistical Analysis

Group comparisons for categorical data were done using χ2 or Fisher exact test, and for continuous data using the Kruskal-Wallis test, as appropriate. The Mantel-Haenszel method was used to compute weighted odds ratios (ORs) that adjusted for a single confounding factor. Reported P values are 2-sided, and P < .05 was considered statistically significant. All analyses were performed using Stata version 11 (StataCorp, College Station, Texas).

RESULTS

Study Population and Clinical Outcomes

Between March 2007 and September 2009, 2588 participants were enrolled. Table 1 shows the baseline characteristics of all participants (N = 2588), and of those with a resistance test result by month 12 (n = 142). Participants were initiated on ART regimens containing either zidovudine (964 [37.2%]), tenofovir (867 [33.5%]), stavudine (691 [26.7%]), or abacavir (66 [2.6%]), combined with either efavirenz (1543 [59.6%]) or nevirapine (1045 [40.4%]). Baseline resistance was detected in 140 (5.4%) participants; the proportions did not differ between the initial regimens.

Table 1.

Baseline Characteristics of Participants

Variable Baseline Month 12 P Valuea 
Overall (N = 2588) Overall (n = 142) No Resistance (n = 42) Resistance (n = 100) 
Sex     .13 
    Female 1482 (57.3) 61 (43.0) 14 (33.3) 47 (47.0)  
    Male 1106 (42.7) 81 (57.0) 28 (66.7) 53 (53.0)  
Age at initiation in years, mean (SD) 38.0 (9.0) 36.0 (8.5) 36.7 (7.9) 35.7 (8.8) .43 
Country     .80 
    Zambia 555 (21.5) 28 (19.7) 9 (21.4) 19 (19.0)  
    South Africa 593 (22.9) 26 (18.3) 9 (21.4) 17 (17.0)  
    Uganda 602 (23.3) 38 (26.8) 8 (19.1) 30 (30.0)  
    Kenya 425 (16.4) 22 (15.5) 7 (16.7) 15 (15.0)  
    Zimbabwe 211 (8.2) 12 (8.5) 3 (7.1) 9 (9.0)  
    Nigeria 17 (25.8) 16 (11.3) 6 (14.3) 10 (10.0)  
WHO clinical stage at initiation     .51 
    1–2 1015 (39.2) 52 (36.6) 17 (40.5) 35 (35.0)  
    3–4 1573 (60.8) 90 (63.4) 25 (59.5) 65 (65.0)  
Baseline HIV-1 drug-resistanceb 140 (5.4)c 15 (10.6) 1 (2.4) 14 (14.0) .10 
Initial ART regimen     .78 
    Efavirenz-based 1543 (59.6) 69 (48.6) 21 (50.0) 48 (48.0)  
     Tenofovir-containing 720 (27.8) 33 (23.4) 10 (23.8) 23 (23.0)  
     Stavudine-containing 430 (16.6) 11 (7.8) 5 (11.9) 6 (6.0)  
     Zidovudine-containing 342 (13.2) 23 (16.2) 6 (14.3) 17 (17.0)  
     Abacavir-containing 51 (2.0) 2 (1.4) 0 (0.0) 2 (2.0)  
    Nevirapine-based 1045 (40.4) 73 (51.4) 21 (50.0) 52 (52.0)  
     Zidovudine-containing 622 (24.0) 48 (33.8) 13 (31.0) 35 (35.0)  
     Stavudine-containing 261 (10.1) 9 (6.3) 4 (9.5) 5 (5.0)  
     Tenofovir-containing 147 (5.7) 14 (9.9) 3 (7.1) 11 (11.0)  
     Abacavir-containing 15 (0.6) 2 (1.4) 1 (2.4) 1 (1.0)  
CD4 cell count, median cells/μL (IQR) 133 (62–204)d 120 (52–198 157 (136) 108 (127.5) .002 
Plasma HIV RNA, median log10 copies/mL (IQR) 5.00 (4.38–5.59)e 5.23 (4.53–5.67) 5.18 (1.31) 5.24 (.99) .19 
HIV-1 subtypef     .14 
    A 611 (24.9) 37 (26.1) 14 (33.3) 23 (23.0)  
    C 1329 (54.2) 69 (48.6) 22 (52.4) 47 (47.0)  
    D 276 (11.3) 19 (13.4) 1 (2.4) 18 (18.0)  
    CRF02_AG 116 (4.7) 6 (4.2) 2 (4.8) 4 (4.0)  
    Other 118 (4.8) 11 (7.8) 3 (7.1) 8 (8.0)  
Variable Baseline Month 12 P Valuea 
Overall (N = 2588) Overall (n = 142) No Resistance (n = 42) Resistance (n = 100) 
Sex     .13 
    Female 1482 (57.3) 61 (43.0) 14 (33.3) 47 (47.0)  
    Male 1106 (42.7) 81 (57.0) 28 (66.7) 53 (53.0)  
Age at initiation in years, mean (SD) 38.0 (9.0) 36.0 (8.5) 36.7 (7.9) 35.7 (8.8) .43 
Country     .80 
    Zambia 555 (21.5) 28 (19.7) 9 (21.4) 19 (19.0)  
    South Africa 593 (22.9) 26 (18.3) 9 (21.4) 17 (17.0)  
    Uganda 602 (23.3) 38 (26.8) 8 (19.1) 30 (30.0)  
    Kenya 425 (16.4) 22 (15.5) 7 (16.7) 15 (15.0)  
    Zimbabwe 211 (8.2) 12 (8.5) 3 (7.1) 9 (9.0)  
    Nigeria 17 (25.8) 16 (11.3) 6 (14.3) 10 (10.0)  
WHO clinical stage at initiation     .51 
    1–2 1015 (39.2) 52 (36.6) 17 (40.5) 35 (35.0)  
    3–4 1573 (60.8) 90 (63.4) 25 (59.5) 65 (65.0)  
Baseline HIV-1 drug-resistanceb 140 (5.4)c 15 (10.6) 1 (2.4) 14 (14.0) .10 
Initial ART regimen     .78 
    Efavirenz-based 1543 (59.6) 69 (48.6) 21 (50.0) 48 (48.0)  
     Tenofovir-containing 720 (27.8) 33 (23.4) 10 (23.8) 23 (23.0)  
     Stavudine-containing 430 (16.6) 11 (7.8) 5 (11.9) 6 (6.0)  
     Zidovudine-containing 342 (13.2) 23 (16.2) 6 (14.3) 17 (17.0)  
     Abacavir-containing 51 (2.0) 2 (1.4) 0 (0.0) 2 (2.0)  
    Nevirapine-based 1045 (40.4) 73 (51.4) 21 (50.0) 52 (52.0)  
     Zidovudine-containing 622 (24.0) 48 (33.8) 13 (31.0) 35 (35.0)  
     Stavudine-containing 261 (10.1) 9 (6.3) 4 (9.5) 5 (5.0)  
     Tenofovir-containing 147 (5.7) 14 (9.9) 3 (7.1) 11 (11.0)  
     Abacavir-containing 15 (0.6) 2 (1.4) 1 (2.4) 1 (1.0)  
CD4 cell count, median cells/μL (IQR) 133 (62–204)d 120 (52–198 157 (136) 108 (127.5) .002 
Plasma HIV RNA, median log10 copies/mL (IQR) 5.00 (4.38–5.59)e 5.23 (4.53–5.67) 5.18 (1.31) 5.24 (.99) .19 
HIV-1 subtypef     .14 
    A 611 (24.9) 37 (26.1) 14 (33.3) 23 (23.0)  
    C 1329 (54.2) 69 (48.6) 22 (52.4) 47 (47.0)  
    D 276 (11.3) 19 (13.4) 1 (2.4) 18 (18.0)  
    CRF02_AG 116 (4.7) 6 (4.2) 2 (4.8) 4 (4.0)  
    Other 118 (4.8) 11 (7.8) 3 (7.1) 8 (8.0)  

Baseline characteristics of all participants at baseline (N = 2588) and of those who had a genotypic resistance test results by month 12 (n = 142). Data are no. (%) unless stated otherwise.

Abbreviations: ART, antiretroviral therapy; HIV-1, human immunodeficiency virus type 1; IQR, interquartile range; SD, standard deviation; WHO, World Health Organization.

a

P values for comparison between no resistance vs resistance outcome groups.

b

Defined as at least 1 drug-resistance mutation of the 2010 International Antiviral Society–USA list.

c

Data available for n = 2442.

d

Data available for n = 2578.

e

Data available for n = 2564.

f

Data available for n = 2450.

A total of 2132 (82.4%) participants were retained in care up to 12 months of follow-up, of whom 2128 (82.2%) were still on first-line ART and 4 (0.2%) were switched to second-line therapy prior to month 12 due to locally diagnosed ART failure. The remaining 456 (17.6%) participants were not retained, because they died (190 [7.3%]), were lost to follow-up (200 [7.7%]), transferred out (61 [2.4%]), or discontinued ART (5 [0.2%]) (Figure 1).

Figure 1.

Study profile. Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; PI, protease inhibitor; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor.

Figure 1.

Study profile. Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; PI, protease inhibitor; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor.

HIV RNA by month 12 was assessed for 94.6% (2014 of 2128) of participants who were still on first-line ART. For all who initiated first-line ART, 70.3% (1820 of 2588; 95% confidence interval [CI], 68.5%–72.1%) achieved viral suppression, and for those who had a 12-month HIV RNA result, 90.4% (1820 of 2014; 95% CI, 89.0%–91.6%) achieved viral suppression (Figure 1).

Drug-Resistance Mutations by Drug Regimen

Of the 166 participants with HIV RNA of >1000 copies/mL by month 12, sequence results were available for 142 (85.5%); 14 specimens failed to amplify and, results were missing for 10 specimens. One hundred (70.4%; 95% CI, 62.2–77.8) sequences harbored ≥1 DRM and 42 (29.6%) did not harbor any DRMs. The average number of DRMs (any class) per sequence was 2.4, with a range of 1–8. Detected DRMs were associated with NRTIs (82 [57.8%]) and NNRTIs (86 [60.6%]) (Figure 2). Dual-class resistance to NRTIs and NNRTIs was detected in 69 (48.6%) participants. Combinations of DRMs included M184V and NNRTI (64 [45.1%]), M184V and TAMs (10 [7.60%]), TAMs and NNRTIs (8 [5.6%]), and M184V, TAMs, and NNRTIs (7 [4.9%]) (Figure 2). In 2 (1.4%) participants, DRMs known to be associated with PIs were observed (both tipranavir-associated Q58E); no triple-class resistance was detected. Exclusion of the 16 (9.6%) participants who had ≥1 single-drug substitution during the first year of ART did not significantly change the frequencies of the DRMs (not shown).

Figure 2.

Frequencies of any drug resistance according to drug class and permutations by nucleoside reverse transcriptase inhibitor (NRTI) and by nonnucleoside reverse transcriptase inhibitor (NNRTI). Figure shows major International Antiviral Society–USA drug-resistance mutations associated with NRTIs and NNRTIs in participants experiencing virological failure. The legend includes the number of patients exposed to each antiretroviral drug. Abbreviations: DRM, drug-resistance mutation; TAM, thymidine analogue mutation.

Figure 2.

Frequencies of any drug resistance according to drug class and permutations by nucleoside reverse transcriptase inhibitor (NRTI) and by nonnucleoside reverse transcriptase inhibitor (NNRTI). Figure shows major International Antiviral Society–USA drug-resistance mutations associated with NRTIs and NNRTIs in participants experiencing virological failure. The legend includes the number of patients exposed to each antiretroviral drug. Abbreviations: DRM, drug-resistance mutation; TAM, thymidine analogue mutation.

Of the NRTI-associated DRMs (Figure 3), M184V was the most frequent (76 [53.5%]), followed by K65R (17 [12.0%]) and any TAM (12 [8.5%]). In 9 (6.3%) participants, ≥2 TAMs were detected, and in 3 (2.1%) participants, ≥3 TAMs were detected. TAM-1 mutations included M41L and T215 F/Y (L210W not observed); TAM-2 mutations included D67N, K219E, and K70R (K219Q not observed). In participants who failed regimens containing zidovudine, stavudine, or tenofovir, ≥1 TAM was detected in 12.7% (9 of 71), 5.0% (1 of 20), and 4.3% (4 of 47), respectively; K65R was detected in 0% (0 of 71), 15.0% (3 of 20), and 27.7% (13 of 47), respectively. Q151M was not detected.

Figure 3.

Frequencies of individual drug-resistance mutations by nucleoside reverse transcriptase inhibitor (NRTI) and by nonnucleoside reverse transcriptase inhibitor (NNRTI). Figure shows major International Antiviral Society–USA drug-resistance mutations associated with NRTIs and NNRTIs in participants experiencing virological failure. The legend includes the number of patients exposed to each antiretroviral drug. Abbreviation: TAM, thymidine analogue mutation.

Figure 3.

Frequencies of individual drug-resistance mutations by nucleoside reverse transcriptase inhibitor (NRTI) and by nonnucleoside reverse transcriptase inhibitor (NNRTI). Figure shows major International Antiviral Society–USA drug-resistance mutations associated with NRTIs and NNRTIs in participants experiencing virological failure. The legend includes the number of patients exposed to each antiretroviral drug. Abbreviation: TAM, thymidine analogue mutation.

Of the NNRTI-associated DRMs (Figure 3), K103N (41 [28.9%]) was the most frequent, followed by Y181C (22 [15.5%]) and G190A (20 [14.1%]). In participants failing an efavirenz-based or nevirapine-based regimen, NNRTI-associated DRMs were detected in 58.0% (40 of 69) and 63.0% (46 of 73), respectively (P = .54); K103N was detected in 33.3% (23 of 69) and 24.7% (18 of 73) participants, respectively (P = .25); Y181C was detected in 7.3% (5 of 69) and 23.3% (17 of 73) participants, respectively (P = .008); and V106M was detected in 20.3% (14 of 69) and 1.4% (1 of 73) participants, respectively (P < .001). V106A only occurred after nevirapine exposure. Sixteen (11.3%) participants harbored only DRMs associated with NNRTIs, and 33 (22.5%) participants harbored ≥2 NNRTI-associated DRMs. K103N occurred as the only NNRTI mutation in 23 of 41 participants (56.1%) and occurred in combination with other NNRTI mutations in 18 participants (with Y181C [7], P225H [3], G190A [2], V106M [1], M230L [1], V106M+Y181C [1], Y181C+G190A [1], P225H+M230L [1], and L100I+P225H [1]).

Compared with participants with virological failure by month 12 who did not harbor any DRM, those with ≥1 DRM had a marginally lower median HIV RNA (4.33 vs 4.71 log10 copies/mL; P = .04); this difference was least pronounced for NNRTI resistance (4.30 vs 4.55 log10 copies/mL; P = .05) and most pronounced for M184V (4.21 vs 4.59 log10 copies/mL, P = .02).

Of the 142 participants with a genotype result by month 12, 135 had a baseline genotype result (7 baseline specimens not collected); of these, 14 (10.4%) participants harbored ≥1 DRM before start of ART. Of the DRMs detected by month 12, 96% were newly acquired during the first year of ART.

HIV-1 Subtype Diversity

Among the 142 participants who had a resistance test result by month 12, HIV-1 subtype C was most commonly identified, followed by A, D, A/G recombinant, and other subtypes/recombinants (Table 1). The subtype distribution did not differ significantly between those with or without DRMs (P = .14). The DRMs according to subtype are shown in Figure 4. K65R was more frequent in subtype C (12 [17.4%]) than non-C (5 [6.9%]) (P = .05). After adjusting for differential tenofovir and stavudine exposure, the association with subtype was not significant (OR for C vs non-C [by Mantel-Haenszel method], 1.27; 95% CI, .41–3.93; P = .68).

Figure 4.

Frequencies of individual drug-resistance mutations by human immunodeficiency virus type 1 (HIV-1) subtype. Figure shows major International Antiviral Society–USA drug-resistance mutations associated with nucleoside reverse transcriptase inhibitors (NRTIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs) in the participants experiencing virological failure. The legend includes the number of patients infected with each HIV-1 subtype. Abbreviation: TAM, thymidine analogue mutation.

Figure 4.

Frequencies of individual drug-resistance mutations by human immunodeficiency virus type 1 (HIV-1) subtype. Figure shows major International Antiviral Society–USA drug-resistance mutations associated with nucleoside reverse transcriptase inhibitors (NRTIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs) in the participants experiencing virological failure. The legend includes the number of patients infected with each HIV-1 subtype. Abbreviation: TAM, thymidine analogue mutation.

Compared with subtype C (13 [18.8%]), K103N occurred more frequently in D (10 [52.6%]; P = .003) and A (13 [35.1%]; P = .06), but did not differ between A and D (P = .21). After adjusting for differential efavirenz and nevirapine exposure, K103N remained significantly associated with subtype (OR for C vs non-C, 0.33 by Mantel-Haenszel method; 95% CI, .15–.75; P = .005; OR for D vs non-D, 3.40; 95% CI, 1.21–9.58; P = .014). Frequencies of other DRMs, except for V106M, which was exclusively observed in subtype C, did not differ between HIV-1 subtypes (Figure 4).

Predicted Genotypic Drug Susceptibility

Of the viruses harboring ≥1 DRM (n = 100), the predicted HIV-1 susceptibility to lamivudine and emtricitabine was reduced in the majority (81%) due to the high frequency of M184V (Figure 5). For the other NRTIs, reduced HIV-1 susceptibility was predicted for abacavir in 42% of sequences, didanosine in 40%, tenofovir in 27%, stavudine in 18%, and zidovudine in 10%. For the NNRTIs, reduced HIV-1 susceptibility was frequently predicted for the first-generation drugs efavirenz and nevirapine (93%) and less frequently for the second-generation drugs etravirine and rilpivirine (59%). High-level resistance was most frequently predicted for nevirapine (92%), followed by lamivudine/emtricitabine (77%) and efavirenz (73%), but was uncommon for etravirine and rilpivirine (5%). Reduced HIV-1 susceptibility was not predicted for any of the ritonavir-boosted PIs.

Figure 5.

Predicted viral susceptibility to reverse transcriptase inhibitors. Figure includes the viruses that harbored ≥1 drug-resistance mutation (n = 100). Genotypic drug susceptibility was predicted using the Stanford HIVdb algorithm (version 6.1.0). Abbreviations: 3TC, lamivudine; ABC, abacavir; AZT, zidovudine; d4T, stavudine; ddI, didanosine; EFV, efavirenz; ETR, etravirine; FTC, emtricitabine; NVP, nevirapine; RPV, rilpivirine; TDF, tenofovir.

Figure 5.

Predicted viral susceptibility to reverse transcriptase inhibitors. Figure includes the viruses that harbored ≥1 drug-resistance mutation (n = 100). Genotypic drug susceptibility was predicted using the Stanford HIVdb algorithm (version 6.1.0). Abbreviations: 3TC, lamivudine; ABC, abacavir; AZT, zidovudine; d4T, stavudine; ddI, didanosine; EFV, efavirenz; ETR, etravirine; FTC, emtricitabine; NVP, nevirapine; RPV, rilpivirine; TDF, tenofovir.

DISCUSSION

This multicountry cohort study examined HIV-1 resistance mutation patterns in African patients with virological failure on different first-line ART regimens and predicted their impact on second-line drug susceptibility. After the first year of ART, 70% of those with virological failure had a virus with ≥1 DRM, and dual-class resistance was observed in 49%. Nearly all (96%) DRMs detected by month 12 were newly acquired in the first year of ART. Previous studies from the region have reported higher frequencies (83%–93% of those with virological failure) and complexity of DRM patterns in persons with prolonged ART failure in the absence of viral load monitoring [4–8]. By contrast, in the present study, routine viral load testing after 12 months of ART enabled the relatively early detection of virological ART failure, which, to some extent, may have prevented the accumulation of resistance. Nonetheless, observed resistance patterns were more extensive for this cohort than for cohorts that received intensive virological monitoring in South Africa [18] and resource-rich countries [19, 20]. Therefore, our data underscore the importance of implementing routine viral load monitoring in ART programs in sub-Saharan Africa to prevent drug-resistance accumulation [3–5]. The presence of DRMs was associated with lower HIV RNA values, which result from the reduced fitness of resistant variants.

In patients who failed a tenofovir-containing regimen, the K65R mutation, which is tenofovir’s signature mutation, was commonly observed. Notably, K65R, rather than TAMs, was also frequently selected after stavudine failure. This finding concurs with previous clinical studies that reported high rates of K65R in subtype C–infected Africans who failed stavudine-containing regimens [4, 5, 7] and contrasts with the low frequencies of K65R observed in subtype B–infected individuals from developed countries [21, 22]. K65R and TAMs represent antagonistic pathways of NRTI resistance [23], and K65R confers cross-resistance to all NRTIs except zidovudine.

Several mechanistic studies have demonstrated in vitro that the K65R mutation, which confers broad cross-resistance to the NRTI class, develops more readily in HIV-1 subtype C, the predominant subtype in sub-Saharan Africa, than in subtype B [24–29]. In our cohort, which included only non-B subtypes, K65R was detected more frequently in subtype C than in subtypes A or D. However, after adjusting for differential stavudine and tenofovir use across the subtypes, the association was not statistically significant. This discrepancy with the published literature is most likely explained by the limited statistical power of the present study to assess the effects of subtype on the rates of individual DRMs. Nonetheless, overall, our data support the conclusion that K65R is frequently selected after stavudine or tenofovir use in African populations receiving first-line ART. Therefore, from a virological perspective, zidovudine rather than tenofovir (as currently recommended by the WHO guidelines [30]), may be preferred in second-line therapy for non-B subtype–infected patients who fail a stavudine-containing first-line regimen in the absence of stringent viral load and resistance monitoring. This scenario would warrant enhanced toxicity monitoring, particularly for zidovudine-associated anemia [31].

WHO advises that stavudine be phased out of first-line regimens due to its serious adverse effects, to be replaced with either tenofovir or zidovudine [30]. Dose reduction (from 40 mg to 30 mg, and even 15–20 mg per day) has been suggested to limit the mitochondrial toxicities from stavudine [32]. However, suboptimal dosing of stavudine might, in theory, result in even more frequent emergence of K65R, particularly when used as part of NNRTI-based regimens. Emerging K65R will compromise future NRTI backbones, and clinical trials are ongoing to establish the effectiveness of (“simplified”) PI-based second-line therapy without an effective NRTI backbone [33].

For second-line ART, WHO recommends a ritonavir-boosted PI (preferably lopinavir or atazanavir), with a dual backbone of 2 new or recycled NRTIs, that is, tenofovir plus lamivudine/emtricitabine (after use of stavudine or zidovudine in first-line) or zidovudine plus lamivudine (after use of tenofovir in first-line) [30]. In our cohort, HIV-1 susceptibility to all PIs was preserved for all participants; zidovudine and tenofovir susceptibility was preserved in the majority of patients who failed first-line ART (90% and 73%, respectively). This means that the majority of patients who are switched empirically to second-line, PI-based therapy will receive at least 2 active drugs. Preliminary studies in Africa have indeed suggested high chance of viral resuppression with empirically prescribed second-line regimens [34–38].

The second-generation NNRTIs rilpivirine or etravirine are being considered as second-line drugs in resource-limited countries in patients previously exposed to efavirenz or nevirapine because of their different resistance profiles and suggested high genetic barrier. Susceptibility to the second-generation NNRTIs is preserved in viruses containing the K103N mutation, preferentially selected by efavirenz, but susceptibility is reduced in the presence of Y181C, preferentially selected by nevirapine [39]. In our cohort, viral susceptibility to the second-generation NNRTIs was already reduced in 59% of participants who had ≥1 DRM. Given that NRTI and NNRTI resistance mutations rapidly accumulate in the absence of viral load monitoring [4, 5], second-line regimens based on 2 NRTIs and a second-generation NNRTI are unlikely to be effective. Instead, use of second-generation NNRTIs as part of first-line ART in resource-limited countries merits further evaluation.

In our cohort, K103N occurred more frequently in subtype D infections than in subtype C infections. By contrast, previous studies in women and infants after the use of single-dose nevirapine to prevent vertical HIV-1 transmission have suggested that K103N accumulates faster in subtype C than D and faster in D than A [40–42]. Additional studies are needed to elucidate the virological mechanisms for different subtypes.

This study has some limitations. Population-based sequencing is not able to detect minority resistant viral strains, thus potentially underestimating resistance [43]. The lack of frequent serial blood sampling after ART initiation precluded a more detailed analysis of the evolution of DRMs after ART initiation.

The number of HIV-1–infected persons accessing ART in resource-limited countries is projected to increase further in view of the goal of universal access [2] and, particularly, the earlier start of ART [44] and emerging treatment for prevention [45]. Before widely implementing such strategies, mathematical modeling and empirical data are urgently needed to investigate their acceptability, adherence, and the possible consequences for the emergence and spread of drug resistance.

In conclusion, regular viral load monitoring aimed at early failure detection may limit the accumulation of HIV-1 drug resistance. The strengthening of national HIV treatment programs through robust supply chains, routine viral load monitoring, and improved access to alternative drug regimens in sub-Saharan Africa to prevent drug resistance is an urgent priority. Zidovudine rather than tenofovir may be the preferred second-line NRTI after first-line stavudine failure. The high frequencies of Y181C and accumulated NNRTI mutations limit the effectiveness of second-generation NNRTIs for use in second-line ART.

Notes

Acknowledgments.

The authors thank the study participants, the staff at the collaborating clinical sites and reference laboratories, and the support staff at PharmAccess Foundation and Contract Laboratory Services. Special thanks go to Michèle van Vugt for supervising study implementation. The PASER study is part of the LAASER program (Linking African and Asian Societies for an Enhanced Response to HIV/AIDS), a partnership of Stichting Aids Fonds, The Foundation for AIDS Research (amfAR)—TREAT Asia, PharmAccess Foundation, and International Civil Society Support.

Author contributions.

R. L. H., C. L. W., W. S., T. F. R. W., and R. S. designed the study and developed the protocol. R. L. H. and K. C. E. S. set up the study and trained and supervised study workers. C. K., M. S., F. C., M. B., K. M., M. W., and A. O. established the cohort and supervised data collection. C. L. W., C. K., and W. S. supervised the laboratory testing. R. L. H. conceived and undertook the data analyses and drafted the manuscript. K. C. E. S., A. M. W., C. L. W., and R. S. helped interpret the data and reviewed the manuscript. All authors reviewed and approved the final version of the manuscript.

Financial support.

PASER is an initiative of the PharmAccess Foundation, with financial support provided by the Ministry of Foreign Affairs of the Netherlands through a partnership with Stichting Aids Fonds (grant number 12454). The funders had no role in the study design; data collection, data analysis, or data interpretation; decision to publish; or writing of the report. The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of any of the institutions mentioned above.

Potential conflicts of interest.

All authors: No reported conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

PASER-Monitoring Collaborating Sites and Collaborators

Lusaka Trust Hospital (M. Siwale and C. Njovu), Coptic Hospital (M. Labib), KARA Clinic and Laboratory (J. Menke), Lusaka, Zambia; Muelmed Hospital, Pretoria, South Africa (M. E. Botes); Themba Lethu Clinic, Clinical HIV Research Unit (F. Conradie, P. Ive, and I. Sanne), and Department of Molecular Medicine and Haematology (C. L. Wallis, E. Letsoalo, and W. S. Stevens), University of the Witwatersrand, Johannesburg, South Africa; Acts Clinic, White River, South Africa (M. Hardman); Newlands Clinic, Harare, Zimbabwe (M. Wellington and R. Luthy); Coast Province General Hospital, International Centre for Reproductive Health Kenya, Mombasa (K. Mandaliya, S. Abdallah, and I. Jao); Mater Misericordiae Hospital, Nairobi, Kenya (M. Dolan); Joint Clinical Research Centre, Fort Portal, Mbale and Kampala, Uganda (C. Kityo, G. Namayanja, L. Nakatudde, I. Nankya, M, Kiconco, M. Abwola, and P. Mugyenyi); Lagos University Teaching Hospital, Nigeria (A. Osibogun and S. Akanmu); Department of Virology, University Medical Center, Utrecht, the Netherlands (R. Schuurman and A. M. Wensing); PharmAccess Foundation, Department of Global Health, Academic Medical Center of the University of Amsterdam, Amsterdam Institute for Global Health and Development, the Netherlands (R. L. Hamers, K. C. E. Sigaloff, E. Straatsma, F. W. Wit, J. Dekker, M. van Vugt, J. M. Lange, and T. F. Rinke de Wit).

References

1.
Gilks
CF
Crowley
S
Ekpini
R
, et al.  . 
The WHO public-health approach to antiretroviral treatment against HIV in resource-limited settings
Lancet
 , 
2006
, vol. 
368
 (pg. 
505
-
10
)
2.
World Health Organization
Towards universal access: scaling up priority HIV/AIDS interventions in the health sector: progress report 2010
 , 
2010
 
3.
Gupta
RK
Hill
A
Sawyer
AW
, et al.  . 
Virological monitoring and resistance to first-line highly active antiretroviral therapy in adults infected with HIV-1 treated under WHO guidelines: a systematic review and meta-analysis
Lancet Infect Dis
 , 
2009
, vol. 
9
 (pg. 
409
-
17
)
4.
Hosseinipour
MC
van Oosterhout
JJ
Weigel
R
, et al.  . 
The public health approach to identify antiretroviral therapy failure: high-level nucleoside reverse transcriptase inhibitor resistance among Malawians failing first-line antiretroviral therapy
AIDS
 , 
2009
, vol. 
23
 (pg. 
1127
-
34
)
5.
Sigaloff
KC
Hamers
RL
Wallis
CL
, et al.  . 
Unnecessary antiretroviral treatment switches and accumulation of HIV resistance mutations; two arguments for viral load monitoring in Africa
J Acquir Immune Defic Syndr
 , 
2011
, vol. 
58
 (pg. 
23
-
31
)
6.
Marconi
VC
Sunpath
H
Lu
Z
, et al.  . 
Prevalence of HIV-1 drug resistance after failure of a first highly active antiretroviral therapy regimen in KwaZulu Natal, South Africa
Clin Infect Dis
 , 
2008
, vol. 
46
 (pg. 
1589
-
97
)
7.
Wallis
CL
Mellors
JW
Venter
WD
Sanne
I
Stevens
W
Varied patterns of HIV-1 drug resistance on failing first-line antiretroviral therapy in South Africa
J Acquir Immune Defic Syndr
 , 
2010
, vol. 
53
 (pg. 
480
-
4
)
8.
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
 , 
2009
, vol. 
9
 pg. 
108
 
9.
Wainberg
MA
Zaharatos
GJ
Brenner
BG
Development of antiretroviral drug resistance
N Engl J Med
 , 
2011
, vol. 
365
 (pg. 
637
-
46
)
10.
Hemelaar
J
Gouws
E
Ghys
PD
Osmanov
S
Global trends in molecular epidemiology of HIV-1 during 2000–2007
AIDS
 , 
2011
, vol. 
25
 (pg. 
679
-
89
)
11.
Hamers
RL
Oyomopito
R
Kityo
C
, et al.  . 
Cohort profile: The PharmAccess African (PASER-M) and the TREAT Asia (TASER-M) monitoring studies to evaluate resistance–HIV drug resistance in sub-Saharan Africa and the Asia-Pacific [published online ahead of print 21 April 2011]
Int J Epidemiol
 , 
2011
12.
Hamers
RL
Wallis
CL
Kityo
C
, et al.  . 
HIV-1 drug resistance in antiretroviral-naive individuals in sub-Saharan Africa after rollout of antiretroviral therapy: a multicentre observational study
Lancet Infect Dis
 , 
2011
, vol. 
11
 (pg. 
750
-
9
)
13.
Hamers
RL
Schuurman
R
Sigaloff
KC
, et al.  . 
Effect of pretreatment HIV-1 drug resistance on immunological, virological, and drug-resistance outcomes of first-line antiretroviral treatment in sub-Saharan Africa: a multicentre cohort study [published online ahead of print 27 October 2011]
Lancet Infect Dis
 , 
2011
14.
Myers
RE
Gale
CV
Harrison
A
Takeuchi
Y
Kellam
P
A statistical model for HIV-1 sequence classification using the subtype analyser (STAR)
Bioinformatics
 , 
2005
, vol. 
21
 (pg. 
3535
-
40
)
15.
de Oliveira
T
Deforche
K
Cassol
S
, et al.  . 
An automated genotyping system for analysis of HIV-1 and other microbial sequences
Bioinformatics
 , 
2005
, vol. 
21
 (pg. 
3797
-
800
)
16.
Johnson
VA
Brun-Vezinet
F
Clotet
B
, et al.  . 
Update of the drug resistance mutations in HIV-1: December 2010
Top HIV Med
 , 
2010
, vol. 
18
 (pg. 
156
-
63
)
17.
Liu
TF
Shafer
RW
Web resources for HIV type 1 genotypic-resistance test interpretation
Clin Infect Dis
 , 
2006
, vol. 
42
 (pg. 
1608
-
18
)
18.
Wallis
C
Papthanasopoulos
M
Fox
M
, et al.  . 
Low rates of nucleoside reverse transcriptase inhibitor (NRTI) resistance detected in a well monitored cohort in South Africa accessing antiretroviral therapy [published online ahead of print 2 December 2011]
Antivir Ther
 , 
2011
 
doi:10.3851/IMP2010
19.
Hatano
H
Hunt
P
Weidler
J
, et al.  . 
Rate of viral evolution and risk of losing future drug options in heavily pretreated, HIV-infected patients who continue to receive a stable, partially suppressive treatment regimen
Clin Infect Dis
 , 
2006
, vol. 
43
 (pg. 
1329
-
36
)
20.
Cozzi-Lepri
A
Phillips
AN
Ruiz
L
, et al.  . 
Evolution of drug resistance in HIV-infected patients remaining on a virologically failing combination antiretroviral therapy regimen
AIDS
 , 
2007
, vol. 
21
 (pg. 
721
-
32
)
21.
Gallant
JE
Staszewski
S
Pozniak
AL
, et al.  . 
Efficacy and safety of tenofovir DF vs stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial
JAMA
 , 
2004
, vol. 
292
 (pg. 
191
-
201
)
22.
Margot
NA
Lu
B
Cheng
A
Miller
MD
Resistance development over 144 weeks in treatment-naive patients receiving tenofovir disoproxil fumarate or stavudine with lamivudine and efavirenz in study 903
HIV Med
 , 
2006
, vol. 
7
 (pg. 
442
-
50
)
23.
Parikh
UM
Bacheler
L
Koontz
D
Mellors
JW
The K65R mutation in human immunodeficiency virus type 1 reverse transcriptase exhibits bidirectional phenotypic antagonism with thymidine analog mutations
J Virol
 , 
2006
, vol. 
80
 (pg. 
4971
-
7
)
24.
Doualla-Bell
F
Avalos
A
Brenner
B
, et al.  . 
High prevalence of the K65R mutation in human immunodeficiency virus type 1 subtype C isolates from infected patients in Botswana treated with didanosine-based regimens
Antimicrob Agents Chemother
 , 
2006
, vol. 
50
 (pg. 
4182
-
5
)
25.
Brenner
BG
Oliveira
M
Doualla-Bell
F
, et al.  . 
HIV-1 subtype C viruses rapidly develop K65R resistance to tenofovir in cell culture
AIDS
 , 
2006
, vol. 
20
 (pg. 
F9
-
13
)
26.
Coutsinos
D
Invernizzi
CF
Xu
H
, et al.  . 
Template usage is responsible for the preferential acquisition of the K65R reverse transcriptase mutation in subtype C variants of human immunodeficiency virus type 1
J Virol
 , 
2009
, vol. 
83
 (pg. 
2029
-
33
)
27.
Invernizzi
CF
Coutsinos
D
Oliveira
M
Moisi
D
Brenner
BG
Wainberg
MA
Signature nucleotide polymorphisms at positions 64 and 65 in reverse transcriptase favor the selection of the K65R resistance mutation in HIV-1 subtype C
J Infect Dis
 , 
2009
, vol. 
200
 (pg. 
1202
-
6
)
28.
Coutsinos
D
Invernizzi
CF
Xu
H
Brenner
BG
Wainberg
MA
Factors affecting template usage in the development of K65R resistance in subtype C variants of HIV type-1
Antivir Chem Chemother
 , 
2010
, vol. 
20
 (pg. 
117
-
31
)
29.
Coutsinos
D
Invernizzi
CF
Moisi
D
, et al.  . 
A template-dependent dislocation mechanism potentiates K65R reverse transcriptase mutation development in subtype C variants of HIV-1
PloS One
 , 
2011
, vol. 
6
 pg. 
e20208
 
30.
World Health Organization
Antiretroviral therapy for HIV infection in adults and adolescents: Recommendations for a public health approach, 2010 revision
 , 
2010
WHO: Geneva, Switzerland
 
31.
Gallant
JE
DeJesus
E
Arribas
JR
, et al.  . 
Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV
N Engl J Med
 , 
2006
, vol. 
354
 (pg. 
251
-
60
)
32.
McComsey
GA
Lo Re
V
3rd
O’Riordan
M
, et al.  . 
Effect of reducing the dose of stavudine on body composition, bone density, and markers of mitochondrial toxicity in HIV-infected subjects: a randomized, controlled study
Clin Infect Dis
 , 
2008
, vol. 
46
 (pg. 
1290
-
6
)
33.
Bartlett
J
Aga
E
Ribaudo
H
, et al.  . 
A pilot study of LPV/r monotherapy following virologic failure of first-line NNRTI-containing regimens in resource-limited settings: the week-24 primary analysis of ACTG 5230
18th Conference on Retroviruses and Opportunistic Infections (Boston)
 , 
2011
Alexandria, VA
CROI
34.
Hosseinipour
MC
Kumwenda
JJ
Weigel
R
, et al.  . 
Second-line treatment in the Malawi antiretroviral programme: high early mortality, but good outcomes in survivors, despite extensive drug resistance at baseline
HIV Med
 , 
2010
, vol. 
11
 (pg. 
510
-
8
)
35.
Sigaloff
KC
Hamers
RL
Wallis
CL
Kityo
C
, et al.  . 
Second-line antiretroviral treatment successfully re-suppresses drug-resistant HIV-1 after first-line failure: prospective cohort in sub-Saharan Africa. J Infect Dis
2011
 
; In press
36.
Pujades-Rodriguez
M
Balkan
S
Arnould
L
Brinkhof
MA
Calmy
A
Treatment failure and mortality factors in patients receiving second-line HIV therapy in resource-limited countries
JAMA
 , 
2010
, vol. 
304
 (pg. 
303
-
12
)
37.
Fox
MP
Ive
P
Long
L
Maskew
M
Sanne
I
High rates of survival, immune reconstitution, and virologic suppression on second-line antiretroviral therapy in South Africa
J Acquir Immune Defic Syndr
 , 
2010
, vol. 
53
 (pg. 
500
-
6
)
38.
Murphy
RA
Sunpath
H
Lu
Z
, et al.  . 
Outcomes after virologic failure of first-line ART in South Africa
AIDS
 , 
2010
, vol. 
24
 (pg. 
1007
-
12
)
39.
Vingerhoets
J
Tambuyzer
L
Azijn
H
, et al.  . 
Resistance profile of etravirine: combined analysis of baseline genotypic and phenotypic data from the randomized, controlled phase III clinical studies
AIDS
 , 
2010
, vol. 
24
 (pg. 
503
-
14
)
40.
Eshleman
SH
Guay
LA
Wang
J
, et al.  . 
Distinct patterns of emergence and fading of K103N and Y181C in women with subtype A vs. D after single-dose nevirapine: HIVNET 012
J Acquir Immune Defic Syndr
 , 
2005
, vol. 
40
 (pg. 
24
-
9
)
41.
Eshleman
SH
Hoover
DR
Chen
S
, et al.  . 
Resistance after single-dose nevirapine prophylaxis emerges in a high proportion of Malawian newborns
AIDS
 , 
2005
, vol. 
19
 (pg. 
2167
-
9
)
42.
Eshleman
SH
Hoover
DR
Chen
S
, et al.  . 
Nevirapine (NVP) resistance in women with HIV-1 subtype C, compared with subtypes A and D, after the administration of single-dose NVP
J Infect Dis
 , 
2005
, vol. 
192
 (pg. 
30
-
6
)
43.
Kozal
MJ
Chiarella
J
St John
EP
, et al.  . 
Prevalence of low-level HIV-1 variants with reverse transcriptase mutation K65R and the effect of antiretroviral drug exposure on variant levels
Antivir Ther
 , 
2011
, vol. 
16
 (pg. 
925
-
9
)
44.
Severe
P
Juste
MA
Ambroise
A
, et al.  . 
Early versus standard antiretroviral therapy for HIV-infected adults in Haiti
N Engl J Med
 , 
2010
, vol. 
363
 (pg. 
257
-
65
)
45.
Cohen
MS
Chen
YQ
McCauley
M
, et al.  . 
Prevention of HIV-1 infection with early antiretroviral therapy
N Engl J Med
 , 
2011
, vol. 
365
 (pg. 
493
-
505
)

Author notes

a
PASER members are listed in the Appendix.

Comments

0 Comments