Antiretroviral resistance and inadequate antiretroviral medication adherence are the most important predictors of virological failure during treatment for HIV infection [1–5]. These two factors are linked in that inadequate antiretroviral medication adherence is a major predictor of the development of antiretroviral medication resistance during HIV treatment . Antiretroviral medication adherence–resistance relationships are class-specific and, in some classes, medication-specific . Knowledge of the association between adherence and resistance may be useful in determining optimal treatment combinations and optimal ways to sequence antiretroviral regimens to improve long-term clinical outcomes.
Current analyses of class-specific antiretroviral medication adherence–resistance relationships are lacking long-term prospective follow-up and have generally included small numbers of treated individuals. This analysis was conducted to assess the relationship between antiretroviral medication adherence and the development of class-specific antiretroviral medication resistance in previously antiretroviral-naive participants enrolled in the Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA) Flexible Initial Retrovirus Suppressive Therapies (CPCRA FIRST 058) Study .
The CPCRA was a clinical trials group sponsored by the National Institutes of Health that conducted community-based HIV/AIDS research targeting underserved populations. The CPCRA FIRST Study was a randomized clinical trial assessing different initial treatment strategies for antiretroviral-naive HIV-infected adults . From 1999 through 2002, participants were randomly allocated (1: 1: 1) to one of three initial strategies: the protease inhibitor strategy (protease inhibitors + nucleoside reverse transcriptase inhibitors (NRTIs)]; the nonnucleoside reverse transcriptase inhibitor (NNRTI) strategy (an NNRTI + NRTIs); or the three class strategy (protease inhibitors + an NNRTI + NRTI(s)]. Clinicians could choose individual medications within the assigned strategy or those medications were randomly assigned for participants enrolled in class-specific substudies (14% of participants) . This analysis was limited to the protease inhibitor and NNRTI strategies in order to evaluate initial regimens that are similar to those currently recommended. One major difference is that nelfinavir was the most common protease inhibitor in the FIRST Study, and nelfinavir is not a currently recommended protease inhibitor for initial therapy . Written informed consent was obtained from all participants in order to enroll in the FIRST Study .
Data collection and definitions
Participant demographics were obtained at study entry. At baseline, months 1 and 4, and then every 4 months thereafter, the data obtained included the following: antiretroviral regimen, CD4 lymphocyte count, HIV-1 RNA level, and self-reported antiretroviral medication adherence (except at baseline). Genotypic antiretroviral resistance testing was done at the time of initial virological failure, defined as an HIV-1 RNA level more than 1000 copies/ml at or after 4 months of follow-up. Resistance was determined using the TRUGENE HIV-1 Genotyping Kit (Bayer HealthCare AG) and CPCRA interpretive algorithm v4.0. The algorithm presented drug susceptibilities for the seven NRTI, three NNRTI, and six protease inhibitor antiretroviral medications approved for use during the FIRST Study.
Adherence was assessed using the CPCRA Adherence Self-Report Form, a global 7-day recall asking participants to estimate their level of adherence to each individual antiretroviral medication in their regimen . For each medication prescribed, a value was assigned based on whether the participant indicated taking all (100%), most (80%), about half (50%), some (20%), or none (0%) of their pills during the previous 7 days. Antiretroviral adherence was calculated at each visit as the average for all components of the regimen. Fixed dose combination formulations were included in adherence calculations once. An individual not on antiretroviral therapy was assigned an adherence value of zero for that visit. An adherence report form was considered missing if a protocol required visit was missed or if a visit was attended but the adherence form was not submitted. Missing forms were not assigned an adherence value. For each participant, we calculated cumulative adherence at each visit as the average of all adherence values up to and including that visit.
This study is a post-hoc analysis of prospectively collected clinical trial data. All participants who were randomized to the protease inhibitor or NNRTI strategies of the FIRST Study who had at least one adherence measurement and for whom study-defined virological failure could be determined (i.e., had at least one HIV RNA measurement at or after the 4-month visit) were included in the analyses. CD4 cell counts and HIV RNA measurements were recorded at the screening and randomization visits; the average of the values at these two visits was considered to be the baseline measurement. Characteristics of the randomized strategy groups were compared using generalized linear models or the χ2-test.
Separately for the randomized NNRTI and protease inhibitor strategies, Cox proportional hazards regression models were used to evaluate the impact of cumulative adherence on time to virological failure with the presence of class-specific antiretroviral medication resistance. The analyses were intention to treat, with participants included in their randomized strategy even if there was a strategy switch prior to initial virological failure. Two models were created; both were adjusted for baseline characteristics (age, sex, race/ethnicity, and baseline CD4 cell count and HIV RNA levels) and had a time-updated variable for cumulative adherence. The first model evaluated cumulative adherence as a continuous-valued variable. The second model evaluated cumulative adherence as a categorical variable (0–79, 80–99, and 100%). The second model was included because the published literature suggested a nonlinear association between adherence and resistance for some medication classes . In both models, individuals without virological failure during follow-up were assumed not to have antiretroviral medication resistance. These analyses demonstrate the association between adherence and resistance in a population initiating antiretroviral therapy.
Because participants and clinicians were free to change from an initial strategy to an alternative at any time during follow-up, sensitivity analyses with censoring at first switch from randomized strategy if it occurred prior to initial virological failure were performed. In addition, Cox regression models were repeated for nonrandomized subgroups within the protease inhibitor strategy: participants who initiated antiretroviral therapy with a ritonavir-boosted protease inhibitor versus a nonboosted protease inhibitor.
Final analyses were undertaken to describe antiretroviral medication resistance in the subpopulation of participants experiencing virological failure. These calculations exclude individuals not experiencing virological failure during follow-up. In these analyses, participants who experienced virological failure and had no strategy switch prior to initial failure were classified according to their cumulative adherence at the time of initial failure (0–79, 80–99, and 100%). The proportion of individuals with class-specific resistance in each adherence category was assessed. These analyses demonstrate the association between adherence and resistance in a population failing antiretroviral therapy.
Statistical analyses were performed using SAS statistical software (version 9.1; SAS Institute Inc., Cary, North Carolina, USA). Hazard ratios are given with 95% confidence intervals (CIs). All P values presented are two-sided.
In the FIRST Study, 933 participants were randomized to the protease inhibitor (n = 470) or NNRTI (n = 463) strategies. Thirty participants (3%) were excluded from these analyses (13 in the protease inhibitor strategy; 17 in the NNRTI strategy) due to no adherence measurement, no HIV RNA measurement at or after the 4-month visit, or no resistance test results available at initial virological failure. There were no significant differences in baseline characteristics between included participants randomized to the protease inhibitor versus NNRTI strategies (data not shown). Detailed characteristics of the FIRST Study population have been published previously . In brief, the population averaged 38 years of age were racially diverse (54% black, 26% white, 17% Latino/a, and 3% other), and 22% were women. The antiretroviral-naive population assessed in these analyses had advanced HIV disease with a mean CD4 lymphocyte count of 212 cells/μl and 39% had a prior AIDS diagnosis.
Specific antiretroviral medications used in the FIRST Study have been previously reported . In summary, in the protease inhibitor strategy, the most common protease inhibitors used were nelfinavir (61%), indinavir/ritonavir (13%), indinavir (11%), and lopinavir/ritonavir (8%). In the NNRTI strategy, the most common NNRTIs used were efavirenz (63%) and nevirapine (37%). In both strategies, the most common NRTI combinations used were zidovudine/lamivudine (56%), stavudine/lamivudine (20%), abacavir/lamivudine (9%), and didanosine/stavudine (9%).
Among participants in the protease inhibitor strategy, 326 (71%) of 457 developed initial virological failure a median of 1.2 years after initiating therapy (Table 1). At the time of virological failure, antiretroviral medication resistance within at least one class was present in 135 (30%); 37 (8%) had protease inhibitor resistance and 121 (26%) had NRTI resistance. Among participants in the NNRTI strategy, 262 (59%) of 446 developed virological failure a median of 3.0 years after initiating therapy. Antiretroviral medication resistance was present in 127 (28%); 112 (25%) had NNRTI resistance and 63 (14%) had NRTI resistance. Two-class or three-class antiretroviral medication resistance at initial virological failure developed in 39 (9%) participants randomized to protease inhibitor and 50 (11%) participants randomized to NNRTI. The most common resistance mutations were for protease inhibitors the D30N (n = 26), for NNRTIs the K103N (n = 81), and for NRTIs the M184I/V (n = 171). A comprehensive list of specific resistance mutations in the FIRST Study has been reported .
Cumulative self-reported adherence at the time of initial virological failure or censoring was excellent in both the protease inhibitor [median 91.6%, interquartile range (IQR) 77.1–100] and NNRTI (median 93.3%, IQR 76.4–100) strategies (P = 0.96). Within the protease inhibitor strategy, median cumulative adherence levels at the time of initial virological failure were 89.8% (IQR 66.7–100) and 93.1% (IQR 79.1–100) for the boosted and nonboosted protease inhibitor recipients, respectively. When assessing cumulative adherence prior to initial virological failure or censoring as a categorical variable in the protease inhibitor strategy, 148 participants had 100% cumulative adherence, 187 had 80–99% adherence, and 120 had 0–79% adherence. In the NNRTI strategy, 120 participants had 100% cumulative adherence, 204 had 80–99 adherence, and 125 had 0–79% adherence.
The results of multivariate Cox regression analyses of time to initial virological failure with class-specific antiretroviral medication resistance are presented in Table 2. In the protease inhibitor strategy, no association was found between cumulative adherence (continuous variable) and protease inhibitor resistance at initial virological failure (hazard ratio 1.1, 95% CI 0.9–1.4, per 10% lower cumulative adherence). In the NNRTI strategy, lower cumulative adherence increased the risk of NNRTI resistance at initial virological failure (hazard ratio 1.2, 95% CI 1.1–1.3 per 10% lower cumulative adherence). In both strategies, lower cumulative adherence was associated with an increased risk of NRTI resistance (Table 2). The results from on-treatment analyses, with censoring at first switch from randomized strategy, were similar in direction, magnitude, and significance (data not shown).
Results from multivariate Cox proportional hazards regression analyses assessing cumulative adherence as a time-updated categorical value (0–79, 80–99, and 100%) are presented in Fig. 1. In the protease inhibitor strategy, once again, there was no association between cumulative adherence and protease inhibitor resistance, whereas in the NNRTI strategy, there was an increased risk of NNRTI resistance in the 80–99% (hazard ratio 2.3, 95% CI 1.4–3.7) and the 0–79% (hazard ratio 6.5, 95% CI 3.9–10.7) cumulative adherence groups compared with individuals with 100% cumulative adherence. In both strategies, there was an increased risk of NRTI resistance in individuals with cumulative adherence levels less than 100% (Fig. 1). The greatest risk of NNRTI resistance was in individuals with the lowest adherence (0–79%).
Among participants randomized to the protease inhibitor strategy, 116 (25%) were prescribed a ritonavir-boosted protease inhibitor at study entry, 338 (74%) were prescribed a nonboosted protease inhibitor, and three (1%) did not initiate therapy with a protease inhibitor (Table 3). Of those on a boosted protease inhibitor regimen, 52% were on indinavir/ritonavir, 31% were on lopinavir/ritonavir, and 10% were on saquinavir/ritonavir. Nonboosted versus boosted protease inhibitor use was associated with a significantly higher risk of protease inhibitor resistance (10% compared with 2%; P < 0.01) and NRTI resistance (29% compared with 19%; P = 0.03) at initial virological failure. In multivariate Cox regression analysis, there was no relationship between cumulative adherence and protease inhibitor resistance for participants initiating therapy with a nonboosted protease inhibitor. Similar analyses were not possible for those who initiated with a boosted-protease inhibitor due to the rarity of events in this group. The two individuals failing boosted protease inhibitor-based therapy with protease inhibitor resistance were no longer receiving boosted protease inhibitor-based therapy at the time of initial virological failure.
In a final analysis, participants experiencing virological failure prior to strategy switch were categorized by their cumulative adherence at initial virological failure: 0–79, 80–99, and 100%. Within each cumulative adherence category, the proportion of participants with class-specific antiretroviral medication resistance at the time of virological failure are presented in Fig. 2. The percentage of individuals failing nonboosted protease inhibitor-based therapy with protease inhibitor resistance (5–30%, depending on cumulative adherence category; Fig. 2a) was lower than the percentage of individuals failing NNRTI-based therapy with NNRTI resistance (37–75%). The percentage of individuals with NRTI resistance at the time of virological failure was higher in participants failing nonboosted protease inhibitor-based therapy (38–64%) compared with participants failing NNRTI-based therapy (7–39%) or boosted protease inhibitor-based therapy (0–41%). At the time of virological failure, NRTI resistance was highest among participants with 80–99% cumulative adherence on all three classes of therapy (Fig. 2b).
The data presented here represent one of the largest studies with the longest duration of follow-up to assess class-specific adherence–resistance relationships in antiretroviral-naive individuals initiating antiretroviral therapy. In analyses including all individuals, no association between cumulative adherence and protease inhibitor resistance was found. For NNRTIs and NRTIs, we found a significant association between lower levels of cumulative adherence and resistance at initial virological failure. In analyses that included only individuals failing antiretroviral therapy, higher levels of cumulative adherence were associated with the presence of nonboosted protease inhibitor, NNRTI, and NRTI resistance mutations. The development of protease inhibitor resistance mutations in individuals initiating boosted protease inhibitor-based therapy was rare. Finally, NRTI resistance was more common in individuals receiving nonboosted protease inhibitor-based therapy compared with those receiving NNRTI-based or boosted protease inhibitor-based therapy.
In this study, there was no apparent association between cumulative adherence to protease inhibitor-based antiretroviral therapy and failure with protease inhibitor resistance. For ritonavir-boosted protease inhibitor-based therapy the reason was clear – the development of protease inhibitor resistance was extremely rare at all adherence levels. The fact that protease inhibitor resistance is rare in individuals failing boosted protease inhibitor-based therapy has been shown previously [13,14]. For nonboosted protease inhibitors, others have shown that the highest risk of protease inhibitor resistance occurs at moderate to high levels of adherence [15,16]. We did not find a similar association. The lack of association between protease inhibitor adherence and resistance in this study may have to do with the adherence assessment tool. Self-reported adherence tends to overestimate true adherence , making it more difficult to identify associations between adherence and outcomes at higher levels of adherence. The lack of association may also be due to the short duration of follow-up prior to initial virological failure in this subpopulation (median time to failure 1.1 years). It takes longer for resistance to develop when more mutations are required for full resistance (the genetic barrier to resistance), as seen with many protease inhibitors .
For NNRTIs, at lower levels of cumulative adherence, the risk of NNRTI resistance at initial virological failure was increased. In the subset of individuals with virological failure, however, resistance was common at all adherence levels. Prior studies showed similar findings [18,19]. Thus, in a population of individuals initiating NNRTI-based antiretroviral therapy, the highest risk of failing with resistance occurs at the lowest levels of adherence. This finding is largely due to the potency of NNRTIs; where potency is defined as the likelihood that a given antiretroviral regimen will suppress HIV-1 viremia below the limits of standard assay detection . On a potent regimen, most individuals with moderate to excellent adherence have complete viral suppression and, therefore, are unlikely to develop resistance. However, in a population of individuals failing antiretroviral therapy, the risk of NNRTI resistance is high at all adherence levels. Two factors are responsible for this relationship. First, most NNRTIs have a low genetic barrier to resistance, making the development of resistance during viral replication common. Second, NNRTI-resistant virus typically retains its ability to replicate in the presence or absence of drug (fitness and replicative capacity ), allowing resistant virus to out-compete wild-type virus at most drug exposure levels .
NRTI resistance, like NNRTI resistance, was more common at lower levels of cumulative adherence. However, in individuals failing virologically, NRTI resistance was seen at high but not perfect levels of cumulative adherence, similar to other studies [16,20]. Of note, 88% of NRTI mutations led to lamivudine and emtricitabine resistance via the M184I/V mutation. Thus, these results more closely approximate the adherence–resistance relationship for deoxycytidine analogs, rather than the whole NRTI class. The deoxycytidine analog NRTI adherence–resistance relationship is defined by the fitness cost of the M184I/V mutation (making resistance at low levels of adherence less likely) and the low genetic barrier to resistance that makes resistance common in the setting of viral replication at higher adherence levels .
The potency of antiretroviral regimens is the largest single determinant of the development of antiretroviral resistance for all antiretroviral classes. The more potent the regimen, the less likely is viral replication (at all levels of adherence) and the less likely is the development of antiretroviral resistance . Although the fitness cost of resistance and the genetic barrier to resistance are important, they matter most during active viral replication (i.e., during virological failure) . The message to HIV-infected individuals should be clear – complete virological suppression remains the goal of antiretroviral therapy. The best way to achieve complete virological suppression is to optimize adherence to all components of multidrug antiretroviral therapy . It is true that there is a theoretical concern of increasing the risk of development of antiretroviral resistance in patients with active viral replication while on therapy. However, the increased chance of successful virological suppression with improved adherence appears to outweigh the risk involved as shown in our analyses.
Another important finding in this analysis is that NRTI resistance is more prevalent in individuals failing nonboosted protease inhibitor-based therapy than in those failing NNRTI or boosted protease inhibitor-based therapy. A previous comparison of nonboosted versus boosted protease inhibitor-based therapy showed similar findings . This finding now extends to NNRTIs, which showed similar rates of development of NRTI resistance compared with boosted protease inhibitors in all adherence categories. Together, these data suggest that it is the potency of the concomitant drugs, rather than their genetic barrier to resistance or fitness cost of resistance that is most important in determining the incidence of resistance to concomitant drugs. Again, these analyses support that any efforts at improving adherence will have the greatest likelihood of decreasing the development of resistance overall.
Most participants who failed antiretroviral therapy in the FIRST Study did not have antiretroviral resistance mutations. Resistance within any antiretroviral medication class at initial virological failure occurred in less than 30% of randomized participants and in less than 50% of participants with virological failure in the protease inhibitor and NNRTI strategies of the FIRST Study. One possible explanation for this finding is that patients may have completely interrupted therapy, acknowledged or not acknowledged. Also, resistance mutations arise as random mutations that occur over time and are dependent on the level, not just the presence, of viral replication. Thus, individuals with low-level viremia or intermittent viremia may not develop resistance for longer periods. Other possibilities to explain the lack of resistance in most individuals failing therapy include low-frequency mutations not picked up by population sequencing or nonsequenced resistance mutations like connection domain or RNase H mutations in reverse transcriptase or GAG mutations that affect protease inhibitor susceptibility [22–24].
There are several important limitations to this study. First, we did not have baseline resistance data. However, baseline resistance was carried out on a random sample of 491 (35%) FIRST Study participants, with rates of any resistance at about 10% . Lack of baseline resistance could have impacted our results if resistance within an individual class of medications had a differential impact on response rates to regimens containing that class. Second, due to the timing of the study, there were relatively few ritonavir-boosted protease inhibitor-based regimens, and the receipt of boosted versus nonboosted protease inhibitor was not randomized. Third, in the Cox regression analyses, the time to virological failure varied greatly between individuals and different adherence patterns likely occurred. The impact of the timing of nonadherence (early, late, and continuous) has not been well described and could have potentially affected these analyses. Finally, our analyses utilized self-reported adherence, as this was the only adherence measure used in the FIRST Study; supporting data such as serum antiretroviral medication levels were not available. Self-reported adherence likely overestimates true adherence and may make it difficult to discern adherence–resistance relationships at higher levels of adherence.
This study is one of the largest prospective analyses of class-specific adherence–resistance relationships to date. These analyses support that in a population initiating antiretroviral therapy, the higher the level of cumulative adherence, the better the outcomes. Excellent adherence and full virological suppression remain the goal of antiretroviral therapy. These analyses also demonstrate why boosted protease inhibitor-based and NNRTI-based antiretroviral therapies are preferred over nonboosted protease inhibitor-based options. Not only are rates of failure higher for nonboosted protease inhibitor-based therapy, so are rates of resistance development for the nonboosted protease inhibitor and the accompanying NRTIs. As we learn more about adherence–resistance relationships, these data can likely be applied to emerging classes of therapy . This knowledge can help provide the foundation for rational design of medication combinations and regimen sequencing to improve the longevity of currently available therapies in the era of HIV as a chronic illness .
We would like to acknowledge and thank the participants in the FIRST Study and the dedicated staff at participating CPCRA units.
The National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health grants 5U01AI042170, 5U01AI046362 and 1U01AI068641, provided financial support for this work as part of the FIRST Study (CPCRA 058) and INSIGHT Network. E.M.G. is supported by a career development award from the National Institutes of Health, National Institute of Allergy and Infectious Diseases (K01 AI067063).
E.M.G.: study design, data organization and interpretation, manuscript writer.
K.H.H.: concept design, data analysis and interpretation, and manuscript editing.
E.E.T.: concept design and manuscript editing.
S.S.: data analysis and interpretation and manuscript editing.
G.P.: data organization and analysis.
W.J.B.: concept design, data interpretation, and manuscript editing.
R.D.M.: parent study design, data interpretation, and manuscript editing.
M.C.: concept design, data interpretation, and manuscript editing.
G.F.: concept design, data interpretation, and manuscript editing.
S.B.M.: concept design, data interpretation, and manuscript editing.
This study was presented in part at the 15th Conference on Retroviruses and Opportunistic Infections, 3–6 February 2008, Boston, Massachusetts, USA.
1. Bangsberg DR, Perry S, Charlebois ED, Clark RA, Roberston M, Zolopa AR, et al
. Nonadherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS 2001; 15:1181–1183.
2. Low-Beer S, Yip B, O'Shaughnessy MV, Hogg RS, Montaner JS. Adherence to triple therapy and viral load response. J Acquir Immune Defic Syndr 2000; 23:360–361.
3. 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.
4. Mannheimer S, Friedland G, Matts J, Child C, Chesney M. The consistency of adherence to antiretroviral therapy predicts biologic outcomes for human immunodeficiency virus-infected persons in clinical trials. Clin Infect Dis 2002; 34:1115–1121.
5. Clavel F, Hance AJ. HIV drug resistance. N Engl J Med 2004; 350:1023–1035.
6. Deeks SG. Determinants of virological response to antiretroviral therapy: implications for long-term strategies. Clin Infect Dis 2000; 30(Suppl 2):S177–S184.
7. Gardner EM, Burman WJ, Steiner JF, Anderson PL, Bangsberg DR. Antiretroviral medication adherence and the development of class-specific antiretroviral resistance. AIDS 2009; 23:1035–1046.
8. MacArthur RD, Novak RM, Peng G, Chen L, Xiang Y, Hullsiek KH, et al
. A comparison of three highly active antiretroviral treatment strategies consisting of nonnucleoside reverse transcriptase inhibitors, protease inhibitors, or both in the presence of nucleoside reverse transcriptase inhibitors as initial therapy (CPCRA 058 FIRST Study): a long-term randomised trial. Lancet 2006; 368:2125–2135.
9. MacArthur RD, Chen L, Mayers DL, Besch CL, Novak R, van den Berg-Wolf M, et al
. The rationale and design of the CPCRA (Terry Beirn Community Programs for Clinical Research on AIDS) 058 FIRST (Flexible Initial Retrovirus Suppressive Therapies) trial. Control Clin Trials 2001; 22:176–190.
10. Hammer SM, Eron JJ Jr, Reiss P, Schooley RT, Thompson MA, Walmsley S, et al
. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society-USA panel. JAMA 2008; 300:555–570.
11. Mannheimer S, Thackeray L, Huppler Hullsiek K, Chesney M, Gardner EM, Wu AW, et al
. A randomized comparison of two instruments for measuring self-reported antiretroviral adherence. AIDS Care 2008; 20:161–169.
12. Kozal MJ, Hullsiek KH, Macarthur RD, Berg-Wolf M, Peng G, Xiang Y, et al
. The Incidence of HIV drug resistance and its impact on progression of HIV disease among antiretroviral-naive participants started on three different antiretroviral therapy strategies. HIV Clin Trials 2007; 8:357–370.
13. Eron J Jr, Yeni P, Gathe J Jr, Estrada V, DeJesus E, Staszewski S, et al
. The KLEAN study of fosamprenavir–ritonavir versus lopinavir–ritonavir, each in combination with abacavir–lamivudine, for initial treatment of HIV infection over 48 weeks: a randomised noninferiority trial. Lancet 2006; 368:476–482.
14. Walmsley S, Bernstein B, King M, Arribas J, Beall G, Ruane P, et al
. Lopinavir-ritonavir versus nelfinavir for the initial treatment of HIV infection. N Engl J Med 2002; 346:2039–2046.
15. 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.
16. Tam LW, Chui CK, Brumme CJ, Bangsberg DR, Montaner JS, Hogg RS, et al
. The relationship between resistance and adherence in drug-naive individuals initiating HAART is specific to individual drug classes. J Acquir Immune Defic Syndr 2008; 49:266–271.
17. Liu H, Golin CE, Miller LG, Hays RD, Beck CK, Sanandaji S, et al
. A comparison study of multiple measures of adherence to HIV protease inhibitors. Ann Intern Med 2001; 134:968–977.
18. Bangsberg DR, Acosta EP, Gupta R, Guzman D, Riley ED, Harrigan PR, et al
. Adherence-resistance relationships for protease and nonnucleoside reverse transcriptase inhibitors explained by virological fitness. AIDS 2006; 20:223–231.
19. Maggiolo F, Airoldi M, Kleinloog HD, Callegaro A, Ravasio V, Arici C, et al
. Effect of adherence to HAART on virologic outcome and on the selection of resistance-conferring mutations in NNRTI- or PI-treated patients. HIV Clin Trials 2007; 8:282–292.
20. King MS, Brun SC, Kempf DJ. Relationship between adherence and the development of resistance in antiretroviral-naive, HIV-1-infected patients receiving lopinavir/ritonavir or nelfinavir. J Infect Dis 2005; 191:2046–2052.
21. Gardner EM, Sharma S, Peng G, Hullsiek KH, Burman WJ, Macarthur RD, et al
. Differential adherence to combination antiretroviral therapy is associated with virological failure with resistance. AIDS 2008; 22:75–82.
22. Dam E, Quercia R, Glass B, Descamps D, Launay O, Duval X, et al
. Gag mutations strongly contribute to HIV-1 resistance to protease inhibitors in highly drug-experienced patients besides compensating for fitness loss. PLoS Pathog 2009; 5:e1000345.
23. Ntemgwa M, Wainberg MA, Oliveira M, Moisi D, Lalonde R, Micheli V, et al
. Variations in reverse transcriptase and RNase H domain mutations in human immunodeficiency virus type 1 clinical isolates are associated with divergent phenotypic resistance to zidovudine. Antimicrob Agents Chemother 2007; 51:3861–3869.
24. Yap SH, Sheen CW, Fahey J, Zanin M, Tyssen D, Lima VD, et al
. N348I in the connection domain of HIV-1 reverse transcriptase confers zidovudine and nevirapine resistance. PLoS Med 2007; 4:e335.
25. Novak RM, Chen L, MacArthur RD, Baxter JD, Huppler Hullsiek K, Peng G, et al
. Prevalence of antiretroviral drug resistance mutations in chronically HIV-infected, treatment-naive patients: implications for routine resistance screening before initiation of antiretroviral therapy. Clin Infect Dis 2005; 40:468–474.
26. Siegel K, Lekas HM. AIDS as a chronic illness: psychosocial implications. AIDS 2002; 16(Suppl 4):S69–S76.
© 2010 Lippincott Williams & Wilkins, Inc.