Share this article on:

Low level of the K103N HIV-1 above a threshold is associated with virological failure in treatment-naive individuals undergoing efavirenz-containing therapy

Goodman, Derrick Da; Zhou, Yunb; Margot, Nicolas Ab; McColl, Damian Jb; Zhong, Lijieb; Borroto-Esoda, Katynaa; Miller, Michael Db; Svarovskaia, Evguenia Sb

doi: 10.1097/QAD.0b013e3283427dcb
Clinical Science

Background: Study GS-01-934 was a randomized open-label phase III study comparing efavirenz and tenofovir/emtricitabine to efavirenz and zidovudine/lamivudine in treatment-naive HIV-1-infected individuals. Through 144 weeks, 50 of 487 participants without baseline nonnucleoside reverse transcriptase inhibitor resistance by population sequencing (efavirenz/tenofovir/emtricitabine, n = 19; efavirenz/zidovudine/lamivudine, n = 31) experienced virologic failure (>400 copies/ml). Here, we analyzed whether the presence of low levels of K103N at baseline correlated with virologic failure.

Methods: Available baseline plasma samples (n = 485) were amplified and tested for K103N using an allele-specific PCR (AS-PCR) assay with a lower detection cut-off of 0.5%.

Results: Sixteen of 476 (3.4%) evaluable participants had low-level K103N at baseline by AS-PCR (0.8–15%). The abundance of the K103N subpopulation at baseline distinguished individuals with virologic failure from those who responded durably to efavirenz-containing therapy. Among six participants with at least 2000 K103N copies/ml before treatment, five experienced virologic failure, compared with only one virologic failure among 10 who had less than 2000 K103N copies/ml (P = 0.008). Multivariate logistic regression analysis showed that K103N viral load at least 2000 copies/ml increased the risk of virologic failure with an odds ratio of 47.4 (95% confidence interval 5.2–429.2, P = 0.0006).

Conclusion: The presence of K103N mutant virus in plasma above 2000 copies/ml prior to therapy in treatment-naive individuals correlated with increased risk of virologic failure of these efavirenz-containing triple-drug regimens.

aGilead Sciences Inc., Durham, North Carolina, USA

bGilead Sciences Inc., Foster City, Carolina, USA.

Received 31 August, 2010

Revised 2 November, 2010

Accepted 9 November, 2010

Correspondence to Evguenia S. Svarovskaia, PhD, Gilead Sciences Inc., 333 Lakeside Drive, Foster City, CA 94404, USA. Tel: +1 650 522 1382; fax: +1 650 522 5890; e-mail:

Back to Top | Article Outline


HIV-1 infection is currently effectively managed with antiretroviral therapies (ARTs), including highly active three-drug combinations that suppress viral replication to very low levels. However, development of drug resistance still presents a major obstacle to the remarkable success of combination therapies. Several studies demonstrated that failure of ART is correlated with the presence of pretreatment drug-resistant variants that are present as majority or minority viral variants [1–9]. These minority drug-resistant variants can be generated within an HIV-1-infected individual due to the high level of virus turnover and the high mutation rate of HIV-1 [10–12]. Low levels of drug-resistant variants can also be a result of transmission of resistant viruses. Even though recent studies suggest that mucosal transmission of HIV-1 minority mutants is unlikely due to the establishment of infection by a single viral variant rather than a subset of viruses [13–16], other routes of HIV-1 transmission still can contribute to transmission of low-level resistant strains [17].

Use of standard HIV-1 genotyping assays provides a convenient tool to assess the presence of drug-resistant variants that exist in the viral population above approximately 20%. More sensitive assays such as deep sequencing technologies, allele-specific PCR (AS-PCR), single-genome sequencing, and others are essential to detect resistant variants at levels lower than that can be detected by population sequencing. Utilization of these assays demonstrated that the presence of minority drug-resistant mutants prior to initiating ART often correlates with treatment failure [1–9]. One of the strongest correlations was reported by Johnson et al. [5] suggesting that baseline minority mutants were associated with an 11-fold increased risk of virologic failure. In that study, reverse transcriptase mutations at positions 103, 181, and/or 184 were detected in seven of 95 patients who subsequently failed antiretroviral treatment and two of 221 with successful treatment outcome resulting in a strong association between preexisting minority resistance mutations and virologic failure (P = 0.0038). Geretti et al.[2] also found that seven of 18 patients who experienced virologic failure and none of 75 who maintained virologic suppression showed pretreatment nonnucleoside reverse transcriptase inhibitor (NNRTI) resistance as detected by standard and sensitive methods combined. Both high-frequency and low-frequency baseline mutants were significantly associated with virologic failure. Even though the conclusions of most of the reports with sensitive detection of resistance mutations were in agreement with each other, several studies did not reach statistical significance for the correlation of the presence of minority resistance and treatment outcome most likely due to the low number of patients analyzed [1,4]. Another comprehensive study by Paredes et al.[7] evaluated the presence of K103N and Y181C individuals from a substudy of efavirenz (EFV) recipients in AIDS Clinical Trial Group protocol A5095. The authors concluded that the presence of baseline minority mutants more than tripled virologic failure risk of first-line EFV-based ART. The authors of this and other studies suggested the need for refinement of a mutant level threshold that identifies individuals at greater risk of virologic failure; however, they were unable to define these threshold levels in their studies. It is likely that a failure threshold for minority resistance mutants may depend on the specific therapy regimen and the mutant sensitivity to the drug combination. A controlled study with a large number of individuals may provide further evidence to establish a failure threshold for the pretreatment burden of minority mutants.

Clinical study GS-01-934 (Study 934) compared two EFV-containing triple-drug combinations: emtricitabine + tenofovir disoproxil fumarate + EFV (FTC + TDF + EFV) and lamivudine + zidovudine + efavirenz (3TC + ZDV + EFV) in antiretroviral treatment-naive HIV-1-infected individuals. The primary safety and efficacy results of Study 934 have been previously described [18–20]. Among 509 individuals who enrolled in the study, there were 22 individuals who had NNRTI resistance at baseline as detected by standard population sequencing. Those individuals were excluded from the primary efficacy analysis of the study and are not included in the present analyses. Among the remaining individuals, the majority of those who experienced virologic failure in both arms developed NNRTI resistance with the K103N mutation present in eight of 19 virologic failure individuals in the TDF + FTC + EFV arm and in 18 of 29 virologic failure individuals in the ZDV + 3TC + EFV arm [20]. Other developing mutations included EFV resistance mutations (K101E, K103E, V108I, V179D, Y188H, G190A/S, or M230L), M184V/I, and thymidine analogue mutations (TAMs). No participant in either arm developed the K65R mutation.

In the study described here, we sought to investigate whether the presence of low levels of the K103N mutation at baseline that were undetectable by standard population sequencing contributed to virologic failure during Study 934 and propose a threshold for the low levels of K103N at baseline that is predictive of virologic failure. All available (476 of 487) baseline samples were tested for the presence of K103N minor populations with an AS-PCR assay that can detect variants representing more than 0.5% of the viral population.

Back to Top | Article Outline

Participants and methods

Study 934 design and sample selection

The participants included in this retrospective analysis were enrolled in study GS-01-934 (Study 934) [20]. Study 934 was a 144-week randomized, open-label study comparing FTC, TDF, and EFV, all dosed once daily, with a control arm composed of the twice-daily fixed-dose combination of 3TC + ZDV (Combivir; GlaxoSmithKline, Research Triangle Park, North Carolina, USA) and once-daily EFV. Antiretroviral treatment-naive, HIV-1-infected individuals with HIV-1 RNA levels greater than 10 000 copies/ml at study entry were randomized 1: 1 to each of the regimens. Participants were stratified on the basis of screening CD4+ cell count (less than or greater than 200 cells/μl). The intent-to-treat (ITT) population consisted of a total of 509 individuals. Upon genotypic analysis at baseline, 22 participants harboring one or more primary NNRTI-R mutations were excluded from the primary efficacy and resistance analyses. The remaining 487 participants, with no detectable primary NNRTI-R mutations at baseline, formed the modified-ITT (mITT) population, the primary efficacy analysis population for Study 934 (FTC + TDF + EFV group, n = 244; 3TC + ZDV + EFV group, n = 243). K103N AS-PCR testing was performed on all available plasma baseline samples of the mITT population (n = 485). K103N AS-PCR testing was also performed at virologic failure time points wherever applicable.

Back to Top | Article Outline

Definition of virologic failure

For the purpose of this study, participants were considered a virologic failure if one of the following criteria was observed: two consecutive HIV-1 viral load readings more than 400 RNA copies/ml after achieving a confirmed less than 400 RNA copies/ml; or last HIV-1 RNA more than 400 RNA copies per ml followed by discontinuation of study; or never achieved a confirmed HIV-1 RNA less than 400 copies/ml. Participants who discontinued Study 934 before week 8 were not considered as virologic failure regardless of viral load at discontinuation.

Back to Top | Article Outline

RT-PCR amplification of plasma samples

Viral RNA was extracted from up to 1 ml of plasma using the QIAmp viral RNA kit (Qiagen, Valencia, California, USA). RNA was transcribed at 42°C for 60 min using HIV-1-specific primer 5′-TATCTGGTTGTGCTTGAATGATTCCTAATGCAT and Super Script II RT (Invitrogen, Carlsbad, California, USA) and then used as template for amplification using high fidelity rTth DNA polymerase XL (Applied Biosystems, Foster City, California, USA) with the following PCR conditions: one cycle of 95°C for 3 min; 35 cycles of 95°C for 15 s, 66°C for 40 s, and 72°C for 2 min; and one cycle at 72°C for 10 min. Forward and reverse primers for the amplification were 5′-GAAATGATGACAGCATGTCAGGGAGT and 5′-TATCTGGTTGTGCTTGAATGATTCCTAATGCAT. The resultant PCR fragment covered protease and most of RT. This preamplified PCR product was subjected to AS-PCR for detection of low levels of K103N.

Back to Top | Article Outline

Clonal analyses

Clonal analyses were performed on selected samples to confirm the presence of resistance mutations. RT-PCR product for clonal analysis was generated as described above for AS-PCR. Cloning of the RT-PCR products was performed using the pCR2.1-TOPO-TA cloning kit (Invitrogen) according to the manufacturer's instructions. A total of 20–300 clones were sequenced (Elim Biopharmaceuticals, Hayward, California, USA).

Back to Top | Article Outline

MultiCode RTx PCR assay for detection of K103N HIV-1 RT mutant viral subpopulations

Plasma samples were PCR amplified and tested for low levels of K103N using MultiCode-RTx technology as previously described with modifications [21,22]. AS-PCR primers were used as follows: curative primer, 5′-GCATCACCCACATCCAGTACTGTTACTGATTT; K103K (AAA) primer, 5′-HEX-GTGCGACAGTACTGTTACTGATTTT; K103K (AAG) primer, 5′-HEX-GTGCGACAGTACTGTTACTGATTTC; K103N (AAC) primer, 5′-FAM-ACCTCGAGTACTGTTACTGATTTG; K103N (AAT) primer, 5′-FAM-ACCTCGCAGTACTGTTACTGATTTA; and reverse primer, 5′-TGGAGAAAATTAGTAGATTTCAGAG. AS-PCR was performed on the Roche LightCycler 480 in 96-well plates using the following cycling conditions: 2 min at 95°C; three cycles of 5 s at 95°C, 5 s at 61°C, and 10 s at 72°C; one cycle of 5 s at 95°C, 5 s at 45°C, and 20 s at 72°C; 80 cycles of 5 s at 95°C, 5 s at 55°C, and 20 s at 72°C and single acquisition of fluorescent signal; melt cycle of 60–95°C. Fluorescence data were exported and analyzed with MultiCode-RTx analysis software (EraGen Biosciences, Madison, Wisconsin, USA). Delta threshold cycle (Ct) values were calculated by Ct value of K103N minus the Ct value of K103K amplification. Quantification of K103N present in plasma samples was determined by using delta Ct values of a standard curve (SigmaPlot/SigmaStat 9.01; SYSTAT Software, San Jose, California, USA). Additional information for plasma samples amplification, AS-PCR, standard curve preparation, and clonal analysis is available upon request.

Back to Top | Article Outline

Statistical analyses

All statistical analyses were performed using SAS software (version 9.1.3; SAS Institute, Cary, North Carolina, USA). Multivariate logistic regression analyses were performed to evaluate the effect of K103N level, baseline HIV RNA level, and baseline CD4 level on the risk of developing virologic failure. Fisher's exact test was used for categorical variable comparisons. P values from two-sided tests that were less than 0.05 were considered statistically significant. P values were not adjusted for multiple comparisons. Smoothing plots of the risk of virologic failure versus baseline percentage of K103N or log10 copies/ml of K103N HIV-1 mutants were implemented using the SAS LOESS procedure.

Back to Top | Article Outline


Low-level detection of K103N by allele-specific PCR

The AS-PCR primers were designed to detect different codon combinations at HIV-1 RT amino acid position K103: AAG and AAA for K103 wild-type, and AAC and AAT for K103N. The cut-off of the AS-PCR assay was determined for each wild-type/mutant codon combination using plasmid mixtures to establish performance of the AS-PCR primers and assay conditions as well as using preamplified plasmid mixtures to establish the background mutation rate introduced during the PCR preamplification step. Using plasmid mixtures as a template, the performance of each primer set combination demonstrated a sensitivity of the assay down to 0.001% with background lower than 0.001%. Using preamplified plasmid mixtures, the background signal ranged from less than 0.1 to 0.1% for the various primer pairs limiting the overall sensitivity of the assay to 0.5% to allow for the highest background observed with the four primer pair sets.

Back to Top | Article Outline

Presence of low-level K103N at baseline associated with virologic failure

All available baseline samples from participants in Study 934 that did not harbor resistance mutations by population sequencing were subjected to AS-PCR for sensitive detection of minor populations of K103N RT mutants with an assay cut-off at 0.5%. Baseline characteristics of antiretroviral treatment-naive individuals enrolled in this study were as follows: the majority were infected with HIV-1 subtype B (>94%) with a median age of 37 years and 14% were women. The race and ethnic distribution were as follows: 58% white, 23% black, 15% Hispanic, and 4% defined as other. HIV infection risk factors included heterosexual contact, homosexual contact, intravenous drug use, blood transfusion, and other risks. AS-PCR results were obtained for 476 of the 487 samples from the mITT population of baseline samples (Table 1). The mITT population excluded 22 participants harboring one or more primary NNRTI-R mutations at baseline with 17 of 22 containing K103N. Two plasma samples were previously depleted and nine samples failed amplification by AS-PCR. AS-PCR detected the presence of low-level K103N in 16 of 476 participants (3.4%). The baseline characteristics, treatment arm, and treatment outcome data through week 144 among these participants are summarized in Table 2. Among the 16 participant baseline samples that were positive for low levels of K103N by AS-PCR, 11 were from the 3TC + ZDV + EFV arm and five were from the FTC + TDF + EFV arm (Tables 1 and 2). There was no significant difference in the incidence of low-level K103N at baseline between the two arms. Six of 16 (38%) participants positive for low levels of K103N at baseline experienced virologic failure, five of whom were in the 3TC + ZDV + EFV arm. Four of these six virologic failure participants developed K103N as detected by population sequencing, whereas two of the six virologic failure participants (participants 1279 and 1452) did not have K103N at failure by population sequencing. However, participant 1279 did show the presence of K103N by AS-PCR (16% K103N) at failure. Participant 1452 showed the G190S NNRTI resistance mutation at virologic failure. This participant had only 20 CD4+cells/μl at baseline and also preexisting low-level M184V (as detected by clonal analysis), which may have contributed to virologic failure.

Table 1

Table 1

Table 2

Table 2

The response to EFV-containing therapy can be described as suboptimal in all six participants who were positive for low levels of K103N at baseline and experienced virologic failure. Even though all six participants achieved a two-log response by at least 8 weeks, only one of six ever achieved confirmed HIV RNA levels of less than 50 copies/ml (participant 1497; Table 2). Two of six participants never experienced viral load suppression to less than 400 copies/ml. Furthermore, three of six participants experienced confirmed rebound above 400 copies/ml or had confirmed more than one log increases from nadir in HIV RNA within the first 4 weeks of the study. CD4 cell responses in this group were also suboptimal as compared to the rest of the participants in the study. Only one of the six participants exhibited a significant increase in CD4 cell count on therapy (participant 1318; Table 2), whereas the other five participants only showed marginal increases in their CD4 cell counts. It is likely that the presence of the resistant minority mutants compromised the efficacy of the triple-drug combination resulting in the observed suboptimal virologic response.

Among all study participants who were negative for the presence of K103N at baseline by AS-PCR, only 9% experienced virologic failure as compared to 38% for the positive participants (P = 0.003, Fisher's exact test). A similar observation was made in the ZDV + 3TC + EFV arm, in which 45% of participants who had low levels of K103N at baseline experienced virologic failure as compared to only 11% virologic failure among the participants who did not have K103N at baseline (P = 0.005, Fisher's exact test). Due to insufficient numbers of participants with virologic failure and low-level K103N in the FTC + TDF + EFV arm, we were unable to determine whether such a correlation with virologic failure existed in this arm.

We also assessed the effect of low-level K103N on predicting virologic failure with multivariate logistic regression analysis. The results showed a strong correlation between pretreatment %K103N and K103N copy number with virologic failure (Table 3). Adjusting for baseline viral load, baseline CD4 cell counts, and treatment arm, the virologic failure odds ratio for %K103N as a continuous variable was 2.33 [95% confidence interval (CI) 1.34–4.08, P = 0.0028] and for the K103N copy number (log10) was 1.79 (95% CI 1.29–2.48, P = 0.0005). It was also interesting to determine whether the time to loss of virologic response was different among participants with low levels of K103N versus the participants who did not harbor K103N at baseline and failed therapy. There was no significant difference observed in the time to loss of virologic response in participants with low levels of K103N versus participants with undetectable K103N (median 111 versus 162 days, respectively; P = 0.42 Mann–Whitney test).

Table 3

Table 3

Back to Top | Article Outline

Threshold quantity of K103N at baseline predictive of virologic failure

Not all participants with low-level K103N experienced virologic failure. Therefore, to determine how the magnitude of K103N may correlate with treatment response, the %K103N was plotted and grouped by treatment response: participants who experienced virologic failure and participants who did not experience virologic failure (nonvirologic failure) by week 144 of the study (Fig. 1a). It appeared that the %K103N was higher among the virologic failure participants. The K103N copy number was also calculated at baseline based on percentage of K103N by AS-PCR and the viral load (Table 2). Figure 1b shows the distribution of the K103N copy numbers per ml among virologic failures and nonvirologic failures. Similarly, the higher levels of K103N appear to partition with virologic failure outcome. To verify this observation, further statistical analysis was performed. Figure 2 shows a smoothing plot of the risk of virologic failure versus the baseline %K103N or log10 copies per ml of K103N HIV-1 mutants (Fig. 2a and 2b, respectively). The actual risk is represented by a solid line with a shaded 95% CI. The risk of virologic failure increases as levels of K103N increase, particularly when the K103N mutation is present at levels above 2% or 2000 copies/ml. Based on the proposed threshold values of K103N at 2% or 2000 copies/ml, we performed multivariate logistic regression analyses that assessed the independent variables as shown in Table 3. The results indicated that if an individual had K103N above 2000 copies/ml at baseline, the chances of failing the triple-drug combinations containing EFV were 47 times greater as compared to the chances for participants who had undetectable K103N (odds ratio 47.4, 95% CI 5.2–429.2, P = 0.0006). In the same analysis, K103N less than 2000 copies/ml, baseline HIV-1 RNA more than 100 000 copies/ml, and baseline CD4+T-cell count at least 200 cells/μl did not predict failure. Similar results were observed within the ZDV+3TC+EFV arm alone (odds ratio 45.5, 95% CI 5.0–411.4, P = 0.0007, data not shown); however, there were insufficient numbers of participants with virologic failure and low-level K103N in the FTC+TDF+EFV arm to determine the specific risk associated with preexisting K103N in that regimen.

Fig. 1

Fig. 1

Fig. 2

Fig. 2

In addition, K103N was categorized using 2% as a threshold value. When K103N was present above 2% at baseline, the chances of failing the triple-drug combinations containing EFV were more than 25 times greater as compared to the chances for participants who had undetectable K103N (odds ratio 25.5, 95% CI 4.6–142.1, P = 0.0002).

In this study, we defined virologic failure as confirmed HIV-1 RNA more than 400 copies/ml at week 144 or early discontinuation, which was identical to the virologic failure definition used for the resistance analysis performed for Study 934 [20]. This virologic failure definition excludes participants who discontinued prior to week 8 of Study 934 regardless of the viral load at discontinuation. However, in addition, we also evaluated the correlation between preexisting K103N and treatment outcome using a more conservative definition of pure virologic failure, which included all participants who discontinued prior to week 8 as a virologic failure if they had more than 400 copies/ml of HIV-1 RNA upon discontinuation. There were 12 and 15 participants in the FTC+TDF+EFV and ZDV+3TC+EFV arms, respectively, discontinuing their study regimen prior to week 8. None of these additional virologic failure participants had low levels of K103N detected in their baseline samples. Using these criteria for virologic failure, multivariate analysis again revealed that preexisting low levels of K103N above 2000 copies/ml were highly predictive of virologic failure (odds ratio 32.8, 95% CI 3.7–291.4, P = 0.0018, data not shown).

The prediction of virologic failure described here was based only on low levels of preexisting K103N and did not account for the participants who contained K103N as detected by standard population sequencing, which has an approximate detection cut-off of 20%. We also performed a combined analysis that would take into account all participants with K103N regardless of the detection method. For this analysis, we added the 10 participants with K103N detected by population sequencing at baseline who remained on originally assigned therapy from Study 934 and repeated the evaluation of the failure threshold for preexisting K103N. The levels of K103N were confirmed to be above 50% using AS-PCR supporting the observation by population-based sequencing. Interestingly, a similar threshold value was obtained ranging between 1260 and 6300 K103N copies/ml depending on the definition of virologic failure. Multivariate analysis using the categorical parameters described above was also performed for all participants, including the ones with K103N by population sequencing at baseline. Again, the results indicated that preexisting K103N strongly correlated with the virologic failure outcome of the EFV-containing triple-dug regimen for both arms combined (odds ratio 51.0, 95% CI 13.9–188.0, P < 0.001, data not shown) as well as for the ZDV + 3TC + EFV arm alone (odds ratio 52.5, 95% CI 10.4–264.2, P < 0.001, data not shown).

Back to Top | Article Outline


The observations described herein reinforce the clinical importance of minor drug-resistant variants of HIV-1 that are often missed by currently used genotypic and phenotypic resistance tests. Preexisting K103N NNRTI-resistant mutant viruses, as detected by a sensitive AS-PCR assay, correlated with virologic failure in treatment-naive HIV-1-infected participants starting an EFV-containing therapy. Based on statistical analyses of the Study 934 dataset (n = 487), a failure threshold for low-frequency pretreatment K103N virus was proposed. We found that the presence of K103N mutant virus in plasma above 2000 copies/ml prior to initiating an EFV-containing triple-drug regimen predicted virologic failure outcome with an odds ratio of 47 (95% CI 5.2–429.2). This failure threshold was proposed based on the treatment outcome of participants enrolled in both arms of clinical study GS-01-934 that compared two EFV-containing regimens: TDF+FTC+EFV and ZDV+3TC+EFV. It is important to point out that this study was limited to the detection of K103N minor variants only and did not include the potential presence of other NNRTI-associated mutants. The potential presence of those unmeasured mutations may have affected the accuracy of the determined threshold. Further studies are needed to verify the relevance of the proposed threshold quantities of drug-resistant mutants for the specific TDF+FTC+EFV arm and other drug regimens.

In this study, 16 of 476 (3.4%) participants had detectable K103N by AS-PCR but not by population sequencing, increasing the percentage of treatment-naive participants with NNRTI resistance to 7.5% (38 of 509) when added to those who had NNRTI resistance by population sequencing. Among 16 participants who were positive for K103N at baseline, only six harbored K103N above the proposed threshold of 2000 copies/ml with five of six failing EFV-containing therapy. All of these five participants enriched K103N at the failure time point as detected by population sequencing (four of five) or AS-PCR (one of five, 16% K103N). Interestingly, three out of five participants also developed M184V suggesting that suboptimal virologic suppression of the EFV component of the therapy allowed development of resistance to 3TC. The observation that minority quasispecies of drug-resistant viruses detected at baseline can rapidly outgrow and become the major virus population resulting in therapy failure was also reported by others [6]. In contrast, another group reported that although the presence of baseline Y181C mutants was associated with a greater risk of virologic failure, other EFV resistance mutations were more commonly found at the time of virologic failure [7].

Ten of 16 (63%) participants positive for K103N at baseline as detected by AS-PCR with a cut-off of 0.5% did respond to ART in this study. This observation is consistent with others reporting that about 70% of participants with pretreatment minority mutants were successfully suppressed on ART [7]. In our study, nine of 10 participants who responded to therapy contained K103N at levels below the proposed threshold that differentiated participants with increased risk for virologic failure. Multivariate logistic regression determined that the presence of K103N at more than 2000 copies/ml increased the risk of virologic failure nearly 50 times, whereas K103N less than 2000 copies/ml did not predict failure. These results suggest a threshold amount of K103N mutant virus that limits the effectiveness of EFV-based triple-drug therapy. The K103N failure threshold is likely to be dependent on the potency of the EFV-accompanying antiretrovirals with more potent and forgiving regimens resulting in higher thresholds. Given the low frequency of low-level resistance mutations and the overall high virologic efficacy of current regimens, multiple studies will be necessary to more precisely define such thresholds for specific regimens. The determined low-level resistance thresholds could be directly applicable and further evaluated in cohorts of women who received single-dose nevirapine as a preventive measure for mother-to-child HIV-1 transmission in resource-limited settings.

Back to Top | Article Outline


D.D.G. and E.S.S. contributed to study design, performance of experiments, data analysis and interpretation, statistics, and manuscript preparation. N.A.M., Y.Z., and L.Z. contributed to study design, data analysis and interpretation, statistics, and manuscript preparation. M.D.M., D.J.M., and K.B.E. contributed to study design, data analysis and interpretation, and manuscript preparation. All authors have participated in manuscript revisions and approved the final version. The present study is supported by Gilead Sciences, Inc. All authors are employees of Gilead Sciences Inc.

The study has previously been presented in part at the XVIII International HIV Drug Resistance Workshop, Fort Myers, Florida, USA, 9–13 June 2009.

Back to Top | Article Outline


1. Balduin M, Oette M, Daumer MP, Hoffmann D, Pfister HJ, Kaiser R. Prevalence of minor variants of HIV strains at reverse transcriptase position 103 in therapy-naive patients and their impact on the virological failure. J Clin Virol 2009; 45:34–38.
2. Geretti AM, Fox ZV, Booth CL, Smith CJ, Phillips AN, Johnson M, et al. Low-frequency K103N strengthens the impact of transmitted drug resistance on virologic responses to first-line efavirenz or nevirapine-based highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2009; 52:569–573.
3. Halvas EK, Wiegand A, Boltz VF, Kearney M, Nissley D, Wantman M, et al. Low frequency nonnucleoside reverse-transcriptase inhibitor-resistant variants contribute to failure of efavirenz-containing regimens in treatment-experienced patients. J Infect Dis 2010; 201:672–680.
4. Jakobsen MR, Tolstrup M, Sogaard OS, Jorgensen LB, Gorry PR, Laursen A, Ostergaard L. Transmission of HIV-1 drug-resistant variants: prevalence and effect on treatment outcome. Clin Infect Dis 2010; 50:566–573.
5. Johnson JA, Li JF, Wei X, Lipscomb J, Irlbeck D, Craig C, et al. Minority HIV-1 drug resistance mutations are present in antiretroviral treatment-naive populations and associate with reduced treatment efficacy. PLoS Med 2008; 5:e158.
6. Metzner KJ, Giulieri SG, Knoepfel SA, Rauch P, Burgisser P, Yerly S, et al. Minority quasispecies of drug-resistant HIV-1 that lead to early therapy failure in treatment-naive and -adherent patients. Clin Infect Dis 2009; 48:239–247.
7. Paredes R, Lalama CM, Ribaudo HJ, Schackman BR, Shikuma C, Giguel F, et al. Preexisting minority drug-resistant HIV-1 variants, adherence, and risk of antiretroviral treatment failure. J Infect Dis 2010; 201:662–671.
8. Peuchant O, Thiebaut R, Capdepont S, Lavignolle-Aurillac V, Neau D, Morlat P, et al. Transmission of HIV-1 minority-resistant variants and response to first-line antiretroviral therapy. AIDS 2008; 22:1417–1423.
9. Simen BB, Simons JF, Hullsiek KH, Novak RM, Macarthur RD, Baxter JD, et al. Low-abundance drug-resistant viral variants in chronically HIV-infected, antiretroviral treatment-naive patients significantly impact treatment outcomes. J Infect Dis 2009; 199:693–701.
10. Coffin JM. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 1995; 267:483–489.
11. Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 1995; 373:123–126.
12. Mansky LM, Temin HM. Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase. J Virol 1995; 69:5087–5094.
13. Keele BF, Giorgi EE, Salazar-Gonzalez JF, Decker JM, Pham KT, Salazar MG, et al. Identification and characterization of transmitted and early founder virus envelopes in primary HIV-1 infection. Proc Natl Acad Sci U S A 2008; 105:7552–7557.
14. Poss M, Rodrigo AG, Gosink JJ, Learn GH, de Vange Panteleeff D, Martin HL Jr, et al. Evolution of envelope sequences from the genital tract and peripheral blood of women infected with clade A human immunodeficiency virus type 1. J Virol 1998; 72:8240–8251.
15. Salazar-Gonzalez JF, Bailes E, Pham KT, Salazar MG, Guffey MB, Keele BF, et al. Deciphering human immunodeficiency virus type 1 transmission and early envelope diversification by single-genome amplification and sequencing. J Virol 2008; 82:3952–3970.
16. Yang OO, Daar ES, Jamieson BD, Balamurugan A, Smith DM, Pitt JA, et al. Human immunodeficiency virus type 1 clade B superinfection: evidence for differential immune containment of distinct clade B strains. J Virol 2005; 79:860–868.
17. Bar KJ, Li H, Chamberland A, Tremblay C, Routy JP, Grayson T, et al. Wide variation in the multiplicity of HIV-1 infection among injection drug users. J Virol 2010; 84:6241–6247.
18. Arribas JR, Pozniak AL, Gallant JE, Dejesus E, Gazzard B, Campo RE, et al. Tenofovir disoproxil fumarate, emtricitabine, and efavirenz compared with zidovudine/lamivudine and efavirenz in treatment-naive patients: 144-week analysis. J Acquir Immune Defic Syndr 2008; 47:74–78.
19. Gallant JE, DeJesus E, Arribas JR, Pozniak AL, Gazzard B, Campo RE, et al. Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med 2006; 354:251–260.
20. Margot NA, Enejosa J, Cheng AK, Miller MD, McColl DJ. Development of HIV-1 drug resistance through 144 weeks in antiretroviral-naive subjects on emtricitabine, tenofovir disoproxil fumarate, and efavirenz compared with lamivudine/zidovudine and efavirenz in Study GS-01-934. J Acquir Immune Defic Syndr 2009; 52:209–221.
21. Svarovskaia ES, Margot NA, Bae AS, Waters JM, Goodman D, Zhong L, et al. Low-level K65R mutation in HIV-1 reverse transcriptase of treatment-experienced patients exposed to abacavir or didanosine. J Acquir Immune Defic Syndr 2007; 46:174–180.
22. Svarovskaia ES, Moser MJ, Bae AS, Prudent JR, Miller MD, Borroto-Esoda K. MultiCode-RTx real-time PCR system for detection of subpopulations of K65R human immunodeficiency virus type 1 reverse transcriptase mutant viruses in clinical samples. J Clin Microbiol 2006; 44:4237–4241.

efavirenz; HIV-1 reverse transcriptase; K103N; minority mutants threshold; minority quasispecies; nonnucleoside reverse transcriptase inhibitor resistance; Study GS-01-934

© 2011 Lippincott Williams & Wilkins, Inc.