Share this article on:

Adherence–resistance relationships for protease and non-nucleoside reverse transcriptase inhibitors explained by virological fitness

Bangsberg, David Ra,b; Acosta, Edward Pc; Gupta, Reenad; Guzman, Davida; Riley, Elise Da; Harrigan, P Richardf; Parkin, Neile; Deeks, Steven Gb

doi: 10.1097/01.aids.0000199825.34241.49
Clinical Science

Objective: To compare the prevalence of resistance by adherence level in patients treated with non-nucleoside reverse transcriptase inhibitors (NNRTI) or protease inhibitors (PI). Also to examine the mechanism of differential class-specific adherence–resistance relationships, focusing on the patient-derived capacity of wild-type and drug-resistant recombinant variants to replicate in vitro in the presence of variable drug levels.

Methods: Participants received unannounced pill count measures to assess adherence, viral load monitoring, and genotypic resistance testing. The replicative capacity of drug-susceptible and drug-resistant recombinants was determined using a single-cycle recombinant phenotypic susceptibility assay. Drug exposure was estimated using population-averaged pharmacological measurements adjusted by participant-specific levels of adherence.

Results: In the NNRTI-treated group, 69% had resistance at 0–48% adherence compared to 13% at 95–100% (P = 0.01). PI resistance was less common than NNRTI resistance at 0–48% adherence (69% versus 23%; P = 0.01). In multivariate analysis, the odds for PI resistance increased (P = 0.03) while the odds for NNRTI resistance decreased (P = 0.04) with improving adherence. Individuals with drug-resistant variants were more likely to have levels of drug exposure where the resistant variant was more fit than the drug-susceptible variant in vitro, while those with drug-susceptible virus were more likely to have levels of drug exposure where the drug-susceptible virus was more fit than the drug-resistant variant (P = 0.005).

Conclusions: NNRTI resistance was more common than PI resistance at low levels of adherence. Class-specific adherence–resistance relationships are associated with the relative replicative capacity of drug-resistant versus wild-type variants to replicate in the presence of clinically relevant drug levels.

From the aEpidemiology and Prevention Interventions Center, Division of Infectious Diseases and

bPositive Health Program, San Francisco General Hospital, UCSF, San Francisco, California

cUniversity of Alabama at Birmingham, Birmingham, Alabama

dHarvard Medical School, Boston, Massachusetts

eMonogram Biosciences, Inc, South San Francisco, California, USA

fBritish Columbia Centre for Excellence in HIV/AIDS, Vancouver, Canada.

Received 11 March, 2005

Revised 25 July, 2005

Accepted 26 September, 2005

Correspondence to Dr D. R. Bangsberg, Epidemiology and Prevention Interventions Center, PO Box 1372, UCSF, San Francisco, California 94143-1372, USA. e-mail:

Back to Top | Article Outline


Non-adherence to HIV antiretroviral therapy is closely associated with incomplete viral suppression [1–3], progression to AIDS [4], and death [5,6]. Early views postulated a bell-shaped relationship between adherence and resistance to HIV therapy [7–9]. Data from cohorts with well-characterized measures of adherence have indicated that the burden of drug resistance to non-boosted protease inhibitors (PI) occurs in patients with high levels of adherence [1,10–13]. The relative absence of PI resistance at low levels of adherence may be because PI resistance mutations decrease the inherent capacity of HIV to replicate in the absence of drug pressure (‘replicative capacity’); thus, high levels of adherence are necessary to create sufficient selective pressure to allow the PI-resistant virus to outcompete the drug-susceptible virus [14]. This mechanism, however, has not been addressed experimentally.

Non-nucleoside reverse transcriptase inhibitors (NNRTI) may have a different adherence–resistance relationship. Unlike most PI drugs, only a single mutation is required to create high-level resistance to NNRTI. Because these mutations occur distant to the active site of the reverse transcriptase enzyme [15], most clinically relevant mutations have limited impact on replicative capacity [16,17]. This may explain why a single dose of nevirapine monotherapy (analogous to very low levels of adherence) can commonly lead to NNRTI drug resistance [18].

Based on the variation in the relative replicative capacity barriers of these drug classes, it was hypothesized that NNRTI resistance may occur at lower of levels of adherence than PI resistance [19]. This hypothesis was examined by comparing the prevalence of resistance at various levels of adherence in patients taking antiretroviral therapy based on either single (non-boosted) PI or NNRTI. Unannounced pill counts at the participant's usual place of residence was used for 6-12 months prior to genotyping as an objective adherence measure. This measure has been closely associated with electronic pill cap monitoring [20], viral suppression [1], drug resistance [11], and progression to AIDS [4].

In order to define the biological mechanisms that might account for epidemiological differences in PI and NNRTI adherence–resistance relationships, drug-susceptible and drug-resistant variants were obtained from this cohort and the relative replicative capacity of the virus was assessed in vitro in the presence of clinically relevant drug concentrations. The analysis addressed the hypothesis that drug-susceptible virus would replicate in vitro more efficiently than drug-resistant virus at low levels of PI exposure, and that the opposite would be true for the NNRTI.

Back to Top | Article Outline


Study design and participant recruitment

Participants were identified from the Research on Access to Care in the Homeless (REACH) cohort, a systematic sample of HIV-positive adults recruited from San Francisco homeless shelters, free-meal programs, and low-income single-room-occupancy hotels [21]. The REACH cohort recruited 330 HIV-positive participants between July 1996 and April 2000. The University of California, San Francisco Committee on Human Subjects Research approved all study procedures.

Individuals were included in the analysis if they were had been on stable antiretroviral drug regimens for a minimum of 9 months. A stable regimen was defined as having at least three antiretroviral drugs with either a single PI or a single NNRTI and at least three monthly adherence measures prior to the genotype measurement. Individuals on regimens with greater than one PI or a combination of a PI and a NNRTI were excluded from the analysis.

Back to Top | Article Outline

Periodic adherence assessments

Every 3 to 6 weeks over 12 months, on an unannounced day, pill counts were conducted on all antiretroviral medications at the participant's usual place of residence [1]. In order to determine the percentage of doses taken, the calculated number of pills taken was divided by the total number of prescribed tablets during the same period. Mean monthly adherence was calculated over the 3 months prior to resistance testing.

Back to Top | Article Outline

Biological measurements

Blood was collected monthly and plasma was processed and stored at −20°C within 6 h of collection. Plasma HIV RNA levels were determined monthly using the HIV-1 Amplicor Monitor Version 1.0 ultrasensitive assay (Roche Molecular Systems, Alameda, California, USA). Genotypic and phenotypic drug resistance was analyzed on the last sample with plasma HIV RNA levels > 50 copies/ml during adherence monitoring. Genotypic HIV drug resistance was determined using the TruGene HIV-1 Resistance Kit according to the manufacturer's recommendations (Visible Genetics, Toronto, Canada) with the following modification. Stored plasma samples were thawed, and 500 μl was centrifuged at 220 000 × g for 1 h at 4°C. HIV RNA was extracted using a commercial RNA extraction kit from the resulting 140 μl viral pellet (Qiagen, Valencia, California, USA). The entire HIV-1 protease gene and codons 38–235 of the reverse transcriptase gene were interrogated and analyzed using Gene Objects software (Visible Genetics).

Genotypic resistance was defined based on the presence of at least one major mutation to the NNRTI or PI used in the subject's regimen [22]. The following codon changes were considered evidence of resistance: for NNRTI, A98G, L100I, K101E/P, K103N/S, V106A/M, V108I, Y181C/I, Y188C/H/l, G190A/C/E/Q/S/T/V, P225H, F227L, M230L, P236L; for PI, D30N, V32I, M46I/l, I47A, G48V, I50L/V, V82A/F/S/T, I84A/C/V, N88S/T, L90M. Phenotypic resistance was assayed using PhenoSense (Monogram Biosciences, South San Francisco, California, USA) [23]. Phenotypic resistance was defined as fold changes in drug susceptibility above the recently defined biological ‘cut-offs’ for wild-type HIV (i.e., the 99th percentile of the fold-change in IC50 for wild-type HIV) [24].

Back to Top | Article Outline

Replicative capacity and generation of fitness curves

The capacity of patient-derived virus to replicate in the presence of increasing levels of drug was measured using a modification of the Phenosense® HIV [23] assay. Briefly, patient-derived reverse transcriptase and protease sequences were inserted into a viral vector containing a luciferase reporter gene. Luciferase activity was measured after a single round of replication in the presence of serial dilutions of drug. The final drug concentrations were selected to maximize curve-fitting accuracy of the assay over ranges of maximal change in IC50 for both drug-susceptible and drug-resistant viruses. The luciferase activity of the vector containing patient-derived gene sequences was compared with luciferase activity of the NL4-3 ‘wild-type’ control at each drug concentration. This allowed generation of fitness curves in which the relative capacity of the resistant and control virus to replicate was defined against a range of drug concentrations [25]. A value of 1 implies an equal ability of the two virus constructs to replicate, while > 1 implies that the patient-derived resistant variant replicates at higher levels than the wild-type control.

An average fitness curve was generated for the three most commonly used drugs in this cohort: nelfinavir, nevirapine, and efavirenz. To generate these resistance curves, luciferase data were utilized from all subjects who had fold-changes in drug susceptibility greater than the upper biological cut-off (99th percentile of wild-type HIV) (3.6 for nelfinavir, 4.5 for nevirapine, and 3.0 for efavirenz) [24], and these data were compared with the luciferase data for wild-type control. These curves (see Figs. 2 and 3 below), therefore, reflect the average relative differences in replication for resistant versus wild-type virus at varying drug concentrations.

Back to Top | Article Outline

Estimation of drug exposure in vivo and in vitro

An estimation of drug exposure in vivo was needed to test the hypothesis that NNRTI-resistant variants were more likely to replicate in the presence of low levels of drug exposure than PI-resistant variants. It had previously been demonstrated that single untimed drug levels were an insensitive measure of drug exposure over time [26]; consequently, adherence measurements were used to estimate the average minimum plasma concentrations of drug (Cmin) during the preceding month. Hence, average adherence (0 to 100%) was multiplied by average Cmin previously established for the three most commonly used drugs in the cohort (nevirapine 3.73 mg/l; efavirenz, 1.77 mg/l; and nelfinavir, 0.7 mg/l [27–29]) to derive a patient-specific measure of drug exposure in vivo.

These patient-specific levels of drug exposure were plotted on the fitness curves. Because the effect of drug on replication in vitro was measured in 10% fetal bovine serum, the concentration of drug used in the phenotypic assay was extrapolated for 100% human serum using previously defined correction factors (nevirapine 1.2; efavirenz, 25.5; and nelfinavir, 90.6 [30]). Fitness curves expressing the replication rates of drug-resistant relative to drug-susceptible viruses over a broad spectrum of protein-adjusted drug concentrations were then plotted, and the relative adherence-adjusted Cmin value for each participant was added to the final plot.

Back to Top | Article Outline

Statistical analysis

Virological outcomes were defined by average monthly plasma HIV RNA levels during the period of adherence monitoring. Resistance outcomes were defined based on the presence or absence of genotypic resistance mutations at the last time point this was measurable during adherence monitoring. The presence of any reverse transcriptase or major protease mutation was used as the primary outcome. Participants with complete suppression were assumed to have no resistance.

Adherence categories were categorized by quartile. Associations between adherence and viral suppression were tested with Spearman's correlation coefficient, χ2-square for trend, and logistic regression. An interaction between regimen type (PI versus NNRTI) and adherence was tested by logistic regression. The multivariable model controlled for treatment duration, prior nucleoside reverse transcriptase inhibitor (NRTI) exposure, and baseline CD4 cell count. All data were analyzed using the SAS statistical analysis software (SAS Institute, Cary, North Carolina, USA).

Back to Top | Article Outline


Study population

Of 110 eligible participants, 74 (67%) had HIV RNA levels > 50 copies/ml. Genotypic resistance testing was performed successfully on samples obtained from 72 (98%) of these individuals. The two individuals without genotype results were censored from this analysis. No participants died or were lost to follow-up during the 12-month observation period.

Of the 108 participants included in the analysis, 54 were on an NNRTI-based regimen and 54 were on a single PI-based regimen (Table 1). There were no significant differences between participants on NNRTI- or PI-based regimens in terms of age, race, gender, housing status, history of injection drug use, baseline CD4 T cell count, nadir CD4 T cell count, or mean pill count adherence. There was no significant difference between the groups in the prevalence of one or more mutation giving resistance to NRTI (41% for NNRTI group versus 30% for PI group; P = 0.45).

Table 1

Table 1

Back to Top | Article Outline

Virological outcomes

Higher levels of adherence were significantly associated with improved viral suppression in participants treated with either NNRTI (r = -0.52; P < 0.0001) or PI (r = −0.47; P = 0.0004) drugs. NNRTI-treated participants were significantly more likely to have evidence of viral suppression to < 50 copies/ml than PI-treated participants (50% versus 22%; P = 0.005; Fig. 1).

Fig. 1

Fig. 1

Back to Top | Article Outline

Class-specific genotypic resistance mutations

Among individuals with incomplete viral suppression, 36% of PI-treated individuals had evidence of drug resistance compared with 74% of NNTRI-treated individuals (P = 0.001). PI resistance was less common than NNRTI resistance in the lowest quartile of adherence (69% versus 23%; P = 0.01). The prevalence of NNRTI resistance declined with improving levels of adherence (P = 0.006; Fig. 1). For example, 69% of individuals at the lowest quartile of adherence (0–48%) had an NNRTI resistance mutation, compared with 13% of individuals in the highest quartile of adherence (95–100%). There was no trend in the prevalence of PI resistance by level of adherence in univariate analyses.

Predictors of drug resistance were assessed using separate multivariable analyses for NNRTI- and PI-treated participants. Each model controlled for CD4 T cell nadir, duration of prior antiretroviral treatment, and prior exposure to mono/dual NRTI therapy. For the NNRTI model, each 10% increase in adherence was associated with a 25% decrease in the odds of NNRTI resistance (Table 2; P = 0.04). For the PI model, each 10% increase in adherence was associated with a 41% increase in the odds of PI resistance (Table 2; P = 0.03). When NNRTI- and PI-treated individuals were combined in a single multivariable model, there was a significant interaction between regimen type and adherence (P = 0.0002), consistent with opposite directions in the association between adherence and resistance for PI versus NNRTI.

Table 2

Table 2

Back to Top | Article Outline

Generation of fitness curves for nevirapine, efavirenz and nelfinavir

Phenotypic resistance testing was successfully performed on 54 patients who had detectable viremia in the presence of drug. Of these, 29 had phenotypic evidence of resistance to their PI or NNRTI (phenotypic resistance was present in 14 of 18 nevirapine-treated patients, 5 of 7 efavirenz treated patients, 6 of 18 nelfinavir-treated patients, 2 of 8 indinavir-treated patients, and 2 of 3 ritonavir- or saquinavir-treated patients). The majority (76%) of NNRTI-treated individuals had phenotypic evidence of resistance to these drugs while a smaller proportion (34%) of PI-treated participants had phenotypic evidence of resistance to nelfinavir (P = 0.005).

A series of fitness curves were generated in which the luciferase activity of vectors containing patient-derived reverse transcriptase/protease was plotted against all tested drug concentrations and compared with the replication of drug-susceptible reference virus at these same drug concentrations (Fig. 2). These fitness curves, therefore, represented the level of luciferase activity (a surrogate for replication) of the patient-derived drug-resistant variant compared with the level of luciferase activity for the drug-susceptible reference virus at each drug concentration. A ratio of > 1 implied that the resistant virus was more fit than the wild-type virus at that level of drug exposure. Samples for this analysis were obtained from all individuals with phenotypic evidence of drug resistance to the drugs for which there were the most data (five taking efavirenz, 14 taking nevirapine, and six taking nelfinavir). All individuals had at least one primary resistance mutation. As shown in Fig. 3, drug-resistant variants replicated at greater levels than the drug-susceptible reference virus at in-vitro protein-adjusted concentrations greater than 0.07 mg/l (nevirapine), 0.02 mg/l (efavirenz), and 0.56 mg/l (nelfinavir).

Fig. 2

Fig. 2

Back to Top | Article Outline

Replicative capacity measured in the presence of drug predicts resistance outcomes

In order to determine if replicative capacity (as measured in presence of drug) predicted resistance outcomes in our cohort, the average fitness curves for each drug were used to estimate the ratio of drug-resistant HIV replication to drug-susceptible HIV replication for each participant's predicted adherence-adjusted Cmin. All participants with phenotypic and genotypic resistance data were included in this analysis (seven taking efavirenz, eight taking nevirapine, and 18 taking nelfinavir). The adherence-adjusted Cmin values were determined for each subject and plotted against the corresponding protein-adjusted drug concentration used in the replication capacity assays (Fig. 3). Individual adherence- and protein-adjusted Cmin values ranged from 0.94 to 3.73 mg/l for nevirapine, 0.51 to 1.74 mg/l for efavirenz, and 0.08 to 0.70 mg/l for nelfinavir (Fig. 3).

Fig. 3

Fig. 3

As shown in Fig. 3, the NNRTI-resistant virus was more fit than the NNRTI-susceptible virus at all protein-adjusted participant-specific Cmin values. In contrast, the nelfinavir-resistant virus was often less fit than the nelfinavir-susceptible virus at subject-specific drug concentrations (Fig. 3). When all data were included, subjects who harbored drug-resistant variants were more likely to have levels of drug exposure where the resistant variant was more fit than the drug-susceptible variant (i.e., where the ratio was > 1). Those who harbored drug-susceptible virus were significantly more likely to have levels of drug exposure where the drug-susceptible virus was more fit than the drug-resistant variant (i.e., where the ratio was < 1) (P = 0.005).

Back to Top | Article Outline


In a diverse cohort using objective adherence measures, we found that the prevalence of NNRTI resistance was significantly higher than the prevalence of PI resistance at low levels of adherence. Multivariable analyses controlling for prior treatment history and stage of disease confirmed that, with lower levels of adherence, the odds of NNRTI resistance increased while the odds of PI resistance decreased. These data are consistent with our hypothesis that the risk of NNRTI resistance is highest at low levels of adherence, while the risk of single PI resistance is highest at levels of adherence just short of complete viral suppression [19]. Our NNRTI adherence–resistance finding is also consistent with previous observations that NNRTI resistance occurs at low levels of adherence [31] and non-boosted PI resistance occurs at high levels of adherence [10–12].

Viral fitness generally refers to the ability of one virus to compete against a second virus in a defined environment, while replicative capacity refers to the inherent ability of a virus to replicate ex vivo in absence of any drug or immunological pressure [32,33]. Most in-vitro studies have used a drug-free environment to define the relative replicative capacity values of drug-susceptible and drug-resistant variants. Here, we considered the capacity of HIV to replicate in the presence of low versus high concentrations of drug. Theoretically, these measurements simultaneously incorporate the impact treatment-selected mutations have on drug susceptibility (resistance) and on the inherent capacity of the mutated enzyme (reverse transcriptase or protease) to function efficiently [25,34]. Using a single-cycle recombinant virus assay to measure replication of patient-derived viruses and drug-susceptible reference virus, we observed that the NNRTI-resistant variant generally exhibited higher replication than reference virus at all clinically relevant levels of drug exposure. Therefore, at even low levels of drug adherence in vivo, the NNRTI-resistant virus should be more fit than the drug-susceptible variant.

Our approach allows estimation of the minimum level of adherence that will allow selection of resistant virus. For NNRTI, as little as 1–2% adherence may result in resistance (Fig. 3), whereas 85% adherence is necessary to select for nelfinavir resistance. This observation is consistent with in-vitro studies, which clearly suggest that the impact of most commonly observed NNRTI resistance mutations (e.g., Y181C, K103N) on reverse transcriptase function and/or viral replicative capacity is minimal [35–37]. In contrast, we observed that PI-susceptible and PI-resistant variants generally had comparable levels of replication at high levels of drug exposure, at least for nelfinavir (Fig. 3). We believe that these observations at least partially explain why many patients fail non-boosted PI regimens without having PI-associated mutations [38].

Although not stated explicitly, inherent in our replicative capacity studies is the impact of class-specific pharmacokinetic properties. NNRTI have very long-half lives and, therefore, relatively high Cmin values at the end of the dosing period. In contrast, non-boosted PI have very short half-lives and low Cmin values at the end of the dosing period. Therefore, the NNRTI drug concentrations in the marginally adherent patient will generally remain to the right on our fitness curves (where the predicted resistant/susceptible ratios are > 1), while the opposite will be true for PI drug concentrations. Ritonavir-mediated boosting of PI partially ameliorates this affect by shifting the predicted Cmin values to the right on our fitness curves. Since these levels of PI exposure appear to be fully suppressive, most replication of HIV in presence of the inhibitor will occur at predicted Cmin values where the ratio of resistant/wild-type variant will be < 1. As suggested by others, these concepts may explain why most patients failing boosted PI regimens harbor PI-susceptible virus [13,39].

Although replicative capacity barriers may affect the risk of resistance in patients with poorer adherence, the most effective mechanism to avoid resistance is to achieve and maintain complete or near complete suppression of viral replication [40,41]. Indeed, NNRTI resistance was less common than PI resistance in patients with moderate to high levels of adherence in large part because rates of viral suppression were far higher in the NNRTI-treated group.

There are several limitations to our study. First, our patients were generally antiretroviral experienced. We attempted to control for prior treatment history with regression analysis; nonetheless, residual confounding is possible. Second, several studies have demonstrated that individual variations in pharmacokinetics are an important predictor of treatment failure [42]. Here, we did not account for participant-specific differences in drug absorption or disposition but rather assumed that these differences were distributed randomly in the population and that adherence was a major factor in determining overall drug exposure. For example, our in-vitro studies would suggest that NNRTI-resistant virus should dominate in nearly all subjects taking at least some NNRTI therapy; however, this was not seen in all participants in our study. Our estimation of Cmin based on adherence level should be considered a first-order approximation of a more complex relationship. Future studies should incorporate both rigorous drug-level monitoring and analysis of time-specific patterns of adherence. Third, we assumed that resistance was not present in those with plasma HIV RNA < 50 copies/ml. We believe that this assumption is valid as most studies suggest limited viral evolution in such patients [43]. Fourth, while we cannot exclude a Hawthorne effect of intensive adherence measurement, others have failed to detect an independent effect of intensive measurement on adherence behavior [44]. Finally, we only included nelfinavir as the candidate PI and did not have sufficient numbers of patients on other PI to analyze replication capacity–resistance relationships. We also did not include patients on the more recent ritonavir-boosted PI regimens, which lead to higher rates of viral suppression and likely have different adherence–resistance relationships [13,39].

In summary, we found that moderate levels of adherence led to higher rates of viral suppression in NNRTI-treated individuals compared with individuals treated with a single PI. While use of an NNRTI can lead to a greater proportion of individuals with viral suppression, the vast majority of NNRTI-treated individuals with levels of adherence too low for viral suppression developed resistance. In contrast, few individuals on single PI therapy with low–moderate levels of adherence developed PI resistance. It is apparent that patients with low levels of adherence to NNRTI therapy are at high risk for resistance, creating a precarious balance between viral suppression and drug resistance.

Sponsorship: This work was supported in part by the NIH (MH54907, AA015287, AI052745, A1058696 and AI058696), the UCSF Center for AIDS Research (P30 MH59037) and Abbott Pharmaceuticals. HIV viral load assays were donated by Roche.

We thank Monogram Biosciences Clinical Reference Laboratory for performance of phenosense® assays and Mojyan Haddad for assistance with data analysis.

Back to Top | Article Outline


1. Bangsberg DR, Hecht FM, Charlebois ED, Zolopa AR, Holodniy M, Sheiner L, et al. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS 2000; 14:357–366.
2. Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000; 133:21–30.
3. Arnsten JH, Demas PA, Farzadegan H, Grant RW, Gourevitch MN, Chang CJ, et al. Antiretroviral therapy adherence and viral suppression in HIV-infected drug users: comparison of self-report and electronic monitoring. Clin Infect Dis 2001; 33:1417–1423.
4. Bangsberg DR, Perry S, Charlebois ED, Clark RA, Roberston M, Zolopa AR, et al. Non-adherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS 2001; 15:1181–1183.
5. Hogg R, Heath K, Bangsberg DR, Yip B, Press N, O'Shaughnessy MV, et al. Intermittent use of triple combination therapy is predictive of mortality at baseline and after one year of follow-up AIDS. AIDS 2002; 16:1051–1058.
6. Garcia de Olalla P, Knobel H, Carmona A, Guelar A, Lopez-Colomes JL, Cayla JA. Impact of adherence and highly active antiretroviral therapy on survival in HIV-infected patients. J Acquir Immune Defic Syndr 2002; 30:105–110.
7. Friedland GH, Williams A. Attaining higher goals in HIV treatment: the central importance of adherence. AIDS 1999; 13(Suppl 1):S61–S72.
8. Mayers DL. Drug-resistant HIV-1: the virus strikes back. JAMA 1998; 279:2000–2002.
9. Wainberg MA, Friedland G. Public health implications of antiretroviral therapy and HIV drug resistance. JAMA 1998; 279:1977–1983.
10. Walsh JC, Pozniak AL, Nelson MR, Mandalia S, Gazzard BG. Virologic rebound on HAART in the context of low treatment adherence is associated with a low prevalence of antiretroviral drug resistance. J Acquir Immune Defic Syndr 2002; 30:278–287.
11. Bangsberg DR, Charlebois ED, Grant RM, Holodniy M, Deeks SG, Perry S, et al. High levels of adherence do not prevent accumulation of HIV drug resistance mutations. AIDS 2003; 17:1925–1932.
12. Harrigan PR, Hogg RS, Dong WW, Yip B, Wynhoven B, Woodward J, et al. Predictors of HIV drug-resistance mutations in a large antiretroviral-naive cohort initiating triple antiretroviral therapy. J Infect Dis 2005; 191:339–347.
13. King MS, Brun SC, Kempf DJ. Relationship between adherence and the development of resistance in antiretroviral-naïve, HIV-1-infected patients receiving lopinavir/ritonavir or nelfinavir. J Infect Dis 2005; 191:2046–2052.
14. Bangsberg DR, Porco TC, Kagay C, Charlebois ED, Deeks SG, Guzman D, et al. Modeling the HIV protease inhibitor adherence-resistance curve by use of empirically derived estimates. J Infect Dis 2004; 190:162–165.
15. Smerdon SJ, Jager J, Wang J, Kohlstaedt LA, Chirino AJ, Friedman JM, et al. Structure of the binding site for nonnucleoside inhibitors of the reverse transcriptase of human immunodeficiency virus type 1. Proc Natl Acad Sci USA 1994; 91:3911–3915.
16. Dykes C, Fox K, Lloyd A, Chiulli M, Morse E, Demeter LM. Impact of clinical reverse transcriptase sequences on the replication capacity of HIV-1 drug-resistant mutants. Virology 2001; 285:193–203.
17. Joly V, Descamps D, Peytavin G, Touati F, Mentre F, Duval X, et al. Evolution of human immunodeficiency virus type 1 (HIV-1) resistance mutations in nonnucleoside reverse transcriptase inhibitors (NNRTIs) in HIV-1-infected patients switched to antiretroviral therapy without NNRTIs. Antimicrob Agents Chemother 2004; 48:172–175.
18. Jackson JB, Becker-Pergola G, Guay LA, Musoke P, Mracna M, Fowler MG, et al. Identification of the K103N resistance mutation in Ugandan women receiving nevirapine to prevent HIV-1 vertical transmission. AIDS 2000; 14:F111–F115.
19. Bangsberg DR, Moss AR, Deeks SG. Paradoxes of adherence and drug resistance to HIV antiretroviral therapy. J Antimicrob Chemother 2004; 53:696–699.
20. Bangsberg D, Hecht F, Charlebois E, Chesney M, Moss A. Comparing objectives measures of adherence to HIV antiretroviral therapy: electronic medication monitors and unannounced pill counts. AIDS Behav 2001; 5:275–281.
21. Moss AR, Hahn JA, Perry S, Charlebois ED, Guzman D, Clark RA, et al. Adherence to highly active antiretroviral therapy in the homeless population in San Francisco: a prospective study. Clin Infect Dis 2004; 39:1190–1198.
22. Johnson VA, Brun-Vezinet F, Clotet B, Conway B, D'Aquila RT, Demeter LM, et al. Update of the drug resistance mutations in HIV-1: 2004. Top HIV Med 2004; 12:119–124.
23. Petropoulos CJ, Parkin NT, Limoli KL, Lie YS, Wrin T, Huang W, et al. A novel phenotypic drug susceptibility assay for human immunodeficiency virus type 1. Antimicrob Agents Chemother 2000; 44:920–928.
24. Parkin NT, Hellmann NS, Whitcomb JM, Kiss L, Chappey C, Petropoulos CJ. Natural variation of drug susceptibility in wild-type human immunodeficiency virus type 1. Antimicrob Agents Chemother 2004; 48:437–443.
25. Perrin V, Mammano F. Parameters driving the selection of nelfinavir-resistant human immunodeficiency virus type 1 variants. J Virol 2003; 77:10172–10175.
26. Liechty CA, Alexander CS, Harrigan PR, Guzman JD, Charlebois ED, Moss AR, et al. Are untimed antiretroviral drug levels useful predictors of adherence behavior? AIDS 2004; 18:127–129.
27. van Heeswijk RP, Veldkamp AI, Mulder JW, Meenhorst PL, Wit FW, Lange JM, et al. The steady-state pharmacokinetics of nevirapine during once daily and twice daily dosing in HIV-1-infected individuals. AIDS 2000; 14:F77–F82.
28. Agouron Viracept (Nelfinavir). [package insert] La Jolla, CA: Agouron; 2003.
29. Bristol Myers Squibb. Sustiva (Efavirenz). [package insert] New York Bristol Myers Squibb; 2003.
30. Limoli KL, Trinh LH, Heilek-Snyder GM, Hellmann NS, Petropoulos CJ. Effects of human serum (HS) on protease inhibitor (PI) and non-nucleoside reverse transcriptase inhibitor (NNRTI) activity in a phenotypic drug susceptibility assay. XIV International Conference on AIDS. Barcelona, July 2002 [abstract].
31. Sethi AK, Celentano DD, Gange SJ, Moore RD, Gallant JE. Association between adherence to antiretroviral therapy and human immunodeficiency virus drug resistance. Clin Infect Dis 2003; 37:1112–1118.
32. Nijhuis M, Deeks S, Boucher C. Implications of antiretroviral resistance on viral fitness. Curr Opin Infect Dis 2001; 14:23–28.
33. Quinones-Mateu ME, Arts EJ. HIV-1 fitness: implications for drug resistance, disease progression, and global epidemic evolution. In HIV Sequence Compendium 2001. Edited by Korber B. Los Alamos, NM: Theoretical Biology and Biophysics Group, Los Alamos National Laboratory; 2001:134–170.
34. Harrigan PR, Bloor S, Larder BA. Relative replicative fitness of zidovudine-resistant human immunodeficiency virus type 1 isolates in vitro. J Virol 1998; 72:3773–3778.
35. Gerondelis P, Archer RH, Palaniappan C, Reichman RC, Fay PJ, Bambara RA, et al. The P236L delavirdine-resistant human immunodeficiency virus type 1 mutant is replication defective and demonstrates alterations in both RNA 5′-end- and DNA 3′-end-directed RNase H activities. J Virol 1999; 73:5803–5813.
36. Iglesias-Ussel MD, Casado C, Yuste E, Olivares I, Lopez-Galindez C. In vitro analysis of human immunodeficiency virus type 1 resistance to nevirapine and fitness determination of resistant variants. J Gen Virol 2002; 83:93–101.
37. Collins JA, Thompson MG, Paintsil E, Ricketts M, Gedzior J, Alexander L. Competitive fitness of nevirapine-resistant human immunodeficiency virus type 1 mutants. J Virol 2004; 78:603–611.
38. Havlir DV, Hellmann NS, Petropoulos CJ, Whitcomb JM, Collier AC, Hirsch MS, et al. Drug susceptibility in HIV infection after viral rebound in patients receiving indinavir-containing regimens. JAMA 2000; 283:229–234.
39. Kempf DJ, King MS, Bernstein B, Cernohous P, Bauer E, Moseley J, et al. Incidence of resistance in a double-blind study comparing lopinavir/ritonavir plus stavudine and lamivudine to nelfinavir plus stavudine and lamivudine. J Infect Dis 2004; 189:51–60.
40. Yeni PG, Hammer SM, Hirsch MS, Saag MS, Schechter M, Carpenter CC, et al. Treatment for adult HIV infection: 2004 recommendations of the International AIDS Society-USA Panel. JAMA 2004; 292:251–265.
41. Dybul M, Fauci AS, Bartlett JG, Kaplan JE, Pau AK. Guidelines for using antiretroviral agents among HIV-infected adults and adolescents. Ann Intern Med 2002; 137:381–433.
42. Acosta EP, Kakuda TN, Brundage RC, Anderson PL, Fletcher CV. Pharmacodynamics of human immunodeficiency virus type 1 protease inhibitors. Clin Infect Dis 2000; 30(Suppl 2):S151–S159.
43. Hermankova M, Ray SC, Ruff C, Powell-Davis M, Ingersoll R, D'Aquila RT, et al. HIV-1 drug resistance profiles in children and adults with viral load of < 50 copies/ml receiving combination therapy. JAMA 2001; 286:196–207.
44. Wagner GJ, Ghosh-Dastidar B. Electronic monitoring: adherence assessment or intervention? HIV Clin Trials 2002; 3:45–51.

adherence; replicative capacity; resistance; non-nucleoside reverse transcriptase inhibitors; protease inhibitors

© 2006 Lippincott Williams & Wilkins, Inc.