Secondary Logo

Journal Logo


High levels of adherence do not prevent accumulation of HIV drug resistance mutations

Bangsberg, David Ra,b; Charlebois, Edwin Da; Grant, Robert Mc; Holodniy, Marke; Deeks, Steven Gb; Perry, Sharonf; Conroy, Kathleen Nugentf; Clark, Richardd; Guzman, Davida; Zolopa, Andrewf; Moss, Andrewd

Author Information



Recent studies have suggested a high level of antiretroviral drug resistance among both recently and chronically HIV-infected patients [1–3]. Non- adherence to therapy is closely associated with incomplete viral suppression [4–6] and disease progression [7–9] and is thought to be a risk factor for the development of drug resistance [10–13]. Studies with objective measures of adherence indicate that average adherence is 70% [4–6,14–17].

The relationship between adherence to HIV combination antiretroviral therapy (HAART) and drug resistance, however, has not yet been clearly defined. Support for the stance that non-adherence promotes drug resistance is based on observations with protease inhibitor (PI) monotherapy and on studies with incompletely characterized adherence measures to combination therapy based on non-nucleoside reverse transcriptase inhibitor drugs (NNRTI) [18,19]. More recent cross-sectional data suggest that resistance actually occurs at moderate to high levels of adherence [4,20–22]. These cross-sectional studies are limited by the assumption that adherence over the measured period is temporally related to the observed drug-resistant mutations.

The present study examines whether HIV antiretroviral adherence is prospectively and temporally related to the development of new drug resistance mutations. Adherence and drug resistance was investigated in a prospective cohort of HIV-positive urban poor individuals with objective measurements of adherence to antiretroviral therapy. Among those with incomplete viral suppression, it was assumed that the development of drug resistance was a function of treatment duration and the rate of accumulation of resistance mutations. It was also assumed that those with viral suppression to < 50 copies/ml did not develop drug resistance. Based on these assumptions, the study examined (i) the relationship between adherence and the duration of therapy, (ii) the relationship between adherence and rates of viral suppression to < 50 copies/ml and (iii) the rate of accumulation of new drug resistance mutations in the HIV-1 protease and reverse transcriptase gene with incomplete viral suppression (viral load > 50 copies/ml) while on therapy.

Adherence has been assessed in various ways. Unannounced pill counts at the participant's usual place of residence [23] has a close association with concurrent viral load [4], electronic pill cap adherence assessment [23], and progression to AIDS [7]. Unlike electronic pill cap assessment, however, unannounced pill count does not interfere with the use of ‘Mediset’ pillbox organizers and does not require multiple devices to measure adherence to all antiretroviral medications. Furthermore, it has the advantage over clinic-based pill count because it results in a more complete count of pills in the participant's possession; the unannounced nature of the visit makes it more difficult for participants to empty their pill bottles (or ‘pill dump') prior to the assessment.


Study design and subject 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 [24, 25]. The REACH cohort recruited 330 HIV-positive subjects between July 1996 and April 2000. The University of California San Francisco Committee on Human Subjects Research approved all procedures.

Adherence monitoring component

Beginning in January 1998, all participants taking three or more antiretroviral medications were invited to participate in adherence monitoring. Every 3 to 6 weeks over a 12-month period, on an unannounced day as previously described [23], pill counts were conducted on all antiretroviral medications at the subject's usual place of residence [4].

Treatment duration over 12 months

Treatment duration over 12 months was assessed with a structured questionnaire administered monthly and confirmed with pill count visits.

Specimen collection

Phlebotomy was conducted monthly. Plasma was processed and stored at −40°C within 6 h of collection. HIV viral loads levels were determined monthly, and specimens for genotyping were selected as described below. CD4 cell count was determined at the baseline adherence-monitoring visit.

Eligibility for genotyping

The genotyping strategy was designed to identify the rate of accumulation of new drug resistance mutations over a fixed interval in patients on stable therapy with detectable viremia. Two specimens (G1, G2) separated by a 6-month interval were selected for genotyping the individuals who had (i) at least 1 month of stable combination antiretroviral therapy prior to G1, (ii) no change in antiretroviral regimen between G1 and G2, (iii) a viral load of > 50 copies/ml at G1 and G2, and (iv) a minimum of 3 months of adherence monitoring between the two specimens. These criteria were defined in order to determine the rate of acquiring new drug resistance mutations on stable therapy in those with detectable viremia and well-characterized adherence between genotype tests.

Viral load and genotyping

HIV-1 viral load was measured using the HIV-1 Amplicor Monitor Version 1.0 ultrasensitive assay (Roche Molecular Systems, Alameda, California, USA). Genotypic HIV drug resistance was determined from plasma-associated HIV RNA using the TruGene HIV-1 Resistance Kit according to the manufacturer's recommendations (Visible Genetics, Toronto, Canada) with the following modification. Plasma samples were thawed, and 500 μl was centrifuged at 22 000 × 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). Known primary and secondary antiretroviral mutations in both genes, as well as polymorphic changes compared with an HIV-1 subtype B consensus sequence, were recorded for each patient's baseline plasma sample. Primary and secondary mutations for drug resistance were defined according to the IAS-USA consensus statement [26], and the number of total drug resistance mutations was used as the primary outcome [27].

Statistical analysis

Adherence categories were defined by quintiles among all people receiving adherence monitoring. Differences in mean treatment duration during the 12-month adherence-monitoring period for each adherence quintile was tested by analysis of variance. The relationship between the accumulation of new drug resistance mutations between G1 and G2 was examined with univariate and multivariate Poisson regression using percentage adherence as a continuous measure. Total drug resistance mutations, PI mutations (primary and secondary), nucleoside analog reverse transcriptase inhibitor (NRTI) resistance mutations, NNRTI resistance mutations and mutation at codon 184 were examined as separate outcomes for subsets of people on selected regimens. The analysis controlled for viral load, CD4 cell count, time on therapy, prior mono or dual NRTI exposure, and number of drug resistance mutations at G1 to assess the independent relationship between adherence and accumulation of drug resistance in a multivariate model. All analyses were conducted with SAS (SAS Institute, Cary, North Carolina, USA).

Estimation of proportion of drug resistance by adherence strata

To estimate the population burden of new drug resistance mutations by adherence quintile over a standard calendar year of observation, the measurements of viral suppression, treatment duration, and rates of new drug resistance mutations were combined. Measured rates of viral suppression was used to determine the proportion of people with viral suppression (< 50 copies/ml). The number of months that a participant received antiretroviral therapy during the 12 months of observation was used to estimate treatment duration over a calendar year. The measured proportion of individuals with viral suppression was multiplied by the rate of accumulation drug resistance mutations and treatment duration in order to estimate total number of new drug resistance mutations in each adherence quintile over 12 months. This assumed that individuals with complete viral suppression (viral load < 50 copies/ml) do not develop drug resistance mutations [28,29].


Participant characteristics

Of the 330 people followed in the cohort, 182 (54%) received at least 1 month of a three drug combination antiretroviral therapy, and 148 (81%) consented to unannounced adherence monitoring at their usual place of residence. The study sample primarily comprised people of color (56.8%), with a high proportion of injection drug users (37.2%). Most individuals were on PI-based therapy (53.4%), and many had a history of mono or dual NRTI exposure prior to the initial PI-based regimen (48.0%) (Table 1).

Table 1
Table 1:
Patient and treatment characteristics.

Participants were selected for longitudinal genotype analyses if they had two specimens with a viral load > 50 copies/ml over 6 months, no change in therapy, more than 30 days of stable therapy before the first genotype test and at least 90 days of adherence monitoring between the two genotype tests. Of the 148 people who received adherence monitoring, 42 (28%) were excluded from genotyping because their viral load was < 50 copies/ml at either G1 or G2; 25 (17%) were excluded because they discontinued therapy; six (4%) were excluded because they changed therapy between G1 and G2; 13 (9%) were excluded because of insufficient adherence data (less than 90 days covered by pill counts), and one person was excluded owing to an error in specimen handling. Of the remaining 61 eligible individuals, genotypes were obtained in 95% (116/122) of all specimens. Genotyping was obtained for both G1 and G2 in 57 (93%) individuals. Patient and regimens characteristics of the genotype subsample and adherence-monitoring sample were similar (Table 1).

Adherence and viral suppression

Median adherence in the 148 individuals over the 12-month observation period was 64.7%, and it was closely associated with mean viral load (r = −0.54; P < 0.0001); a mean viral load of < 50 copies/ml viral load was found for 37 (25%). Mean adherence in those with < 50 copies/ml was 82%, compared with 58% in those with > 50 copies/ml (P < 0.0001). Mean adherence between G1 and G2 in the subset receiving genotyping was 63.5%. Other characteristics of this group are given in Table 1.

Treatment duration

The number of months of treatment across the 12 months of adherence monitoring was closely associated with adherence level. Individuals in the 0–41, 42–57, 58–78, 79–91, and 92–100% adherence quintiles received a mean of 6.4, 9.8, 10.1, 11.0, and 11.1 months of HAART, respectively (P < 0.0001).

Prevalence of drug resistance mutations at baseline

Participants had an average of 3.2 drug resistance mutations at G1. For individual antiretroviral medication classes there were 1.0 mean primary PI mutations, 1.1 mean secondary PI mutations (n = 44), 1.0 mean NRTI mutation (n = 57), 0.77 mean NNRTI mutations (n = 18), and 0.3 mean M184V mutations (n = 43) per individual. In a multivariate Poisson regression, the number of total drug resistance mutations at G1 was positively associated with months of therapy prior to G1 (P = 0.02) and prior mono or dual NRTI exposure (P = 0.006).

Incidence of drug resistance mutations over 6 months

Participants who remained viremic on a stable regimen developed a mean of 0.93 new primary or secondary drug resistance mutations between G1 and G2 (0.15 per month). Of the 57 people with genotype data for G1 and G2, 22 (39%) individuals gained at least one net new mutation; 30 (52%) had no net change in drug resistance mutations, and five (9%) had a net loss of drug resistance mutations between G1 and G2.

The number of new drug resistance mutations over 6 months was significantly and positively associated with level of adherence between G1 and G2 (P = 0.004). The number of resistance mutations was also positively associated with the number of antiretroviral treatment months (mean, 16) prior to G1 (P = 0.005). The number of new drug resistance mutations was not associated with viral load, number of drug resistance mutations at baseline or prior NRTI exposure. Pill-count adherence remained positively associated with the number of new drug resistance mutations (P = 0.0002) in a multivariate model including prior months of HAART, baseline CD4 cell count, baseline viral load, baseline number of drug resistance mutations, and prior mono or dual NRTI exposure (Table 2).

Table 2
Table 2:
Predictors of new drug resistance mutations by antiretroviral class.

For specific antiretroviral drug classes, adherence was associated with accumulation of new PI drug mutations (P < 0.0001), new primary PI drug mutations (P < 0.0001) and new secondary PI mutations (P = 0.03) in the 44 people receiving a PI. Adherence was not significantly associated with new NRTI resistance mutations (n = 57), new NNRTI mutations (n = 18), or M184V mutations (n = 43) in patients receiving these drugs (Table 2; Fig.1).

Fig. 1.
Fig. 1.:
 Observed rates of new drug resistance mutations over 6 months in viremic patients (geometric mean > 50 copies/ml) by adherence quintile. DRM, drug resistance mutations; PI total, all protease inhibitor mutations; PI primary, primary protease inhibitor mutations; PI 2ndary, secondary protease inhibitor mutations; NAM, nucleoside analog reverse transcriptase mutations; NNRTI, non-nucleoside analog reverse transcriptase mutations; M184V, mutations associated with lamivudine. **P < 0001 and *P < 0.01 by Poisson regression.

Estimation of proportions of drug resistance mutations by adherence strata in the treated population

In each adherence stratum, the proportion of drug resistance mutations expected over 12 months was estimated by combining the subgroup with detectable viremia (measured rate of new drug resistance mutations) and the subgroup with undetectable viremia (new drug resistance mutation rate assumed to be zero) in their observed proportions. It was estimated that, during 12 months of observation at the population level, 23% of all drug resistance mutations would occur in the 92–100% adherence stratum, compared with 30%, 15%, 20%, and 12% in the 79–91%, 58–78%, 42–57%, and 0–41% strata, respectively (Table 3).

Table 3
Table 3:
Estimated numbers and proportions of new drug resistance mutations generated in the treated population over a 1-year period by adherence quintile.


Our observations suggest that 23% of drug resistance mutations occur in individuals in the top quintile of adherence (92–100%), and over 50% of all drug resistance occurs in the 40% most adherent patients (79–100%). There are several components to this conclusion. First, the rate of accumulating drug resistance mutations increased with increasing adherence among patients with incomplete viral suppression. Second, individuals with high levels of adherence sustained therapy over a longer period than people with low levels of adherence. Prolonged therapy in the absence of complete viral suppression is more likely to create high-level drug resistance to multiple classes than brief periods of therapy. Third, the proportion of people with sustained viral loads < 50 copies/ml increased with improving adherence. We estimate that these phenomena combine such that the combined effect of prolonged treatment duration and more rapid accumulation of drug resistance mutations in viremic patients balances higher rates of viral suppression in highly adherent patients. The net effect at the population level is that a high proportion of drug resistance mutations are in patients with high levels of adherence.

This relationship between adherence and drug resistance was strongest for PI mutations. We did not detect a relationship between adherence and the development of NRTI mutations. This is in contrast to both our earlier report of more common NRTI resistance mutations in highly adherent patients [4] and similar findings by Kuritzkes et al. [22]. This discrepancy is likely to result from the fact that half of the participants in our prospective study had pre-existing NRTI resistance mutations at baseline in the setting of prior mono or dual NRTI exposure, and that the rate of subsequent NRTI mutation accumulation may be slower in variants that have already developed high-level drug resistance. Fifty-four percent of baseline NRTI resistance mutations occurred in the highest adherent quintile (data not shown) and theses individuals may be saturated with NRTI mutations at G1 [22].

We also did not see a relationship between adherence and new drug resistance mutations for NNRTI (n = 18) or the M184V mutation associated with lamivudine resistance (n = 42). We cannot exclude a relationship because of the small number of patients; however, different antiretroviral medications and codon mutations likely have different adherenceresistance relationships. Rapid development of resistance has been observed after single-dose nevirapine therapy [30] and short-course lamivudine therapy given for the prevention of mother-to-child HIV transmission [31], suggesting that resistance to these medications may occur rapidly and at low levels of adherence. These mutations are readily selected during therapy because they have potent effects on drug susceptibility compared with their effects on replicative capacity. In contrast, protease mutations are typically insufficient to abrogate PI activity when they occur as single mutations [32], and multiple mutations are associated with substantial costs to replicative capacity [33]. Our data suggest that high levels of adherence may be necessary to create sufficient drug pressure to select and maintain PI drug-resistant virus.

Our finding is consistent with several cross-sectional studies. In a previous study, our group found that HIV-1 PI drug resistance mutations were present only in individuals with 65–100% adherence, and drug resistance mutations in the reverse transcriptase gene were also more common in patients with higher levels of adherence [4]. In separate reports, Walsh et al. [20] and Howard et al. [34] demonstrated that the number of drug resistance mutations was associated with increasing adherence as measured with electronic medication monitors. Kuritzkes et al. [22] found that highly adherent patients were more likely to have baseline drug resistance. Finally, Gallego et al. [21] found that resistance was limited to individuals reporting ≥ 90% adherence on an indinavir-based regimen. However, these cross-sectional analyses remain limited by the possibility that resistance occurred prior to adherence assessment. Our prospective analysis over a defined interval overcomes this limitation. Collectively, these findings suggest that the greatest risk for resistance is in patients with high levels of adherence and incomplete viral suppression, and this relationship is strongest for PI-based therapy.

There are several limitations to our findings. First, we measured rates of drug resistance mutations over 6 months in a selected subset of individuals on stable therapy in order to avoid changes in mutational pattern caused by starting, stopping, or changing therapy. It is possible that people who change therapy may change because of resistance. However, relatively few people (4%) changed therapy in the group we studied. Second, we assumed that individuals with a viral plasma load < 50 copies/ml did not develop drug resistance based on observations that resistance is uncommon in these individuals [28,29]. If resistance mutations do develop in such individuals, the proportion of all drug resistance at high levels of adherence would be greater than our estimates. Finally, few patients were antiretroviral drug naive and few patients received ritonavir-boosted PI therapy. Our results may not be generalizable to such patients.

These findings have important implications for the practice of using expected adherence as a criterion to determine candidacy for treatment. Based on the expectation that low levels of adherence will lead to drug resistance, current guidelines indicate that the likelihood of patient adherence should be considered in the decision to initiate antiretroviral therapy [35]. If, as we suggest, resistance does not occur more frequently in patients with low levels of adherence, trial therapy in patients who are expected not to adhere gives some patients the opportunity to prove the clinician's expectation of non-adherence mistaken and derive clinical benefit that would be otherwise withheld. Our results suggest that the resistance costs of this approach may be relatively low, especially with respect to protease resistance.

While our findings indicate that resistance is most likely to occur in individuals with high levels of adherence and detectable viremia, these findings do not suggest that low levels of adherence should be advocated in the effort to avoid drug resistance. High levels of adherence predict delayed progression to both AIDS and death [7–9]. Our findings combined with these studies suggest that levels of adherence too low to generate drug resistance are also too low to delay AIDS progression or death. Conversely, high levels of adherence, even with drug-resistant virus, will likely provide more clinical benefit than either low levels of adherence or no treatment [33,36,37]. Consequently, the goal remains for patients and providers to strive for exceptional levels of adherence both to delay progression to AIDS and death and to achieve reliable and durable viral suppression in order to limit the development of drug resistance as much as possible.

Sponsorship: This study was funded by NIMH grants 54907 and 63011, the University Wide Aids Research Program of the State of California, the AIDS Clinical Research Center of the University of California, and the University of California, San Francisco-Gladstone Institute of Virology and Immunology Center for AIDS Research, P30 MH59037. Viral load kits were donated by Roche. Dr Bangsberg received additional funding from The Doris Duke Charitable Foundation.

Note: The Epidemiology and Prevention Interventions Center, Positive Health Program and Center for AIDS Prevention Studies are programs of the UCSF AIDS Research Institute.


1. Little SJ, Holte S, Routy JP, Daar ES, Markowitz M, Collier AC, et al. Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med 2002, 347:385–394.
2. UK Collaborative Group on Monitoring the Transmission of HIV Drug Resistance.Analysis of prevalence of HIV-1 drug resistance in primary infections in the United Kingdom. BMJ 2001, 322:1087–1088.
3. Richman D, Bozzette S, Morton S, Wrin T, Dawson K, Morton S, et al. The prevalence of antiretroviral drug resistance in the US.Interscience Conference on Antimicrobial Agents and Chemotherapy. Chicago, 2001 [abstract LB-17].
4. Bangsberg DR, Hecht FM, Charlebois ED, Zolop 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.
5. 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.
6. 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.
7. 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.
8. Hogg RS, Heath K, Bangsberg D, Yip B, Press N, O'Shaughnessy MV, et al. Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS 2002, 16:1051–1058.
9. de Olalla P, Knobel H, Carmona A, Guelar A, Lopez-Colomes J, Cayla J. Impact of adherence on highly active antiretroviral therapy on survival in HIV-infected patients. J AIDS 2001, 30:105–110.
10. Wainberg MA, Friedland G. Public health implications of antiretroviral therapy and HIV drug resistance. [See comments.] JAMA 1998, 279:1977–1983.
11. Mayers DL. Drug-resistant HIV-1: the virus strikes back [editorial; comment]. JAMA 1998, 279:2000–2002.
12. Durant J, Clevenbergh P, Halfon P, Delgiudice P, Porsin S, Simonet P, et al. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. [See comments; published erratum appears in Lancet 1999, 354:1128.] Lancet 1999, 353:2195–2199.
13. Friedland GH, Williams A. Attaining higher goals in HIV treatment: the central importance of adherence. AIDS 1999, 13(Suppl 1):S61–S72.
14. McNabb J, Ross JW, Abriola K, Turley C, Nightingale CH, Nicolau DP. Adherence to highly active antiretroviral therapy predicts virologic outcome at an inner-city human immunodeficiency virus clinic. Clin Infect Dis 2001, 33:700–705.
15. 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.
16. Gross R, Bilker WB, Friedman HM, Strom BL. Effect of adherence to newly initiated antiretroviral therapy on plasma viral load. AIDS 2001, 15:2109–2117.
17. Walsh JC, Mandalia S, Gazzard BG. Responses to a 1 month self-report on adherence to antiretroviral therapy are consistent with electronic data and virological treatment outcome. AIDS 2002, 16:269–277.
18. Montaner JS, Reiss P, Cooper D, Vella S, Harris M, Conway B, et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: the INCAS Trial. Italy, The Netherlands, Canada and Australia Study. [See comments.] JAMA 1998, 279:930–937.
19. Vanhove GF, Schapiro JM, Winters MA, Merigan TC, Blaschke TF. Patient compliance and drug failure in protease inhibitor monotherapy.JAMA 1996, 276:1955–1956.
20. 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.
21. Gallego O, de Mendoza C, Perez-Elias MJ, Guardiola JM, Pedreira J, Dalmau D, et al. Drug resistance in patients experiencing early virological failure under a triple combination including indinavir. AIDS 2001, 15:1701–1706.
22. Kuritzkes D, Ickovics J, Bassett R, Hellman N, Johnson V. HIV-1 drug resistance and medication adherence in patients receiving NRTIs. HIV Resistance and Pathogenesis Meeting. Scottsdale, June 2001 [abstract 84].
23. 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.
24. Zolopa AR, Hahn JA, Gorter R, Miranda J, Wlodarczyk D, Peterson J, et al. HIV and tuberculosis infection in San Francisco's homeless adults. Prevalence and risk factors in a representative sample. [See comments.] JAMA 1994, 272:455–461.
25. Robertson M, Clark R, Charlebois E, Tulsky J, Bangsberg D, Long H, et al. HIV seroprevalence and risk factors in a representative sample of homeless and marginally housed adults in San Francisco. Am J Public Health 2003, in press.
26. Hirsch MS, Brun-Vezinet F, D'Aquila RT, Hammer SM, Johnson VA, Kuritzkes DR, et al. Antiretroviral drug resistance testing in adult HIV-1 infection: recommendations of an International AIDS Society-USA Panel. JAMA 2000, 283:2417–2426.
27. Zolopa AR, Shafer RW, Warford A, Montoya JG, Hsu P, Katzenstein D, et al. HIV-1 genotypic resistance patterns predict response to saquinavir–ritonavir therapy in patients in whom previous protease inhibitor therapy had failed. Ann Intern Med 1999, 131:813–821.
28. 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.
29. Ruff CT, Ray SC, Kwon P, Zinn R, Pendleton A, Hutton N, et al. Persistence of wild-type virus and lack of temporal structure in the latent reservoir for human immunodeficiency virus type 1 in pediatric patients with extensive antiretroviral exposure. J Virol 2002, 76:9481–9492.
30. 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.
31. Clarke JR, Braganza R, Mirza A, Stainsby C, Ait-Khaled M, Wright A, et al. Rapid development of genotypic resistance to lamivudine when combined with zidovudine in pregnancy. J Med Virol 1999, 59:364–368.
32. Condra JH. Resistance to HIV protease inhibitors. Haemophilia 1998, 4:610–615.
33. Deeks SG, Wrin T, Liegler T, Hoh R, Hayden M, Barbour JD, et al. Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viremia. N Engl J Med 2001, 344:472–480.
34. Howard A, Arnsten J, Gardner L, Rich J, Schuman P, McKenna P, et al. Lack of multi-drug resistance in nonresponders to antiretroviral therapy with poor adherence.1st International AIDS Society Conference on HIV Pathogenesis and Treatment. Buenos Aires 2002 [abstract 604].
35. UK Department of Health and Social Security. Guidelines for the use of antiretroviral agents in adults and adolescents. HIV Clin Trial 2001, 2:227–306.
36. Ledergerber B, Egger M, Opravil M, Telenti A, Hirschel B, Battegay M, et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss HIV Cohort Study. Lancet 1999, 353:863–868.
37. Deeks SG, Barbour JD, Martin JN, Swanson MS, Grant RM. Sustained CD4 T cell response after virologic failure of protease inhibitor-based regimens in patients with human immunodeficiency virus infection. J Infect Dis 2000, 181:946–953.

adherence; resistance; HIV; viral suppression; viral evolution; health policy; homeless; injection drug user

© 2003 Lippincott Williams & Wilkins, Inc.