The potency and effectiveness of HAART treatment for human immunodeficiency virus (HIV) infection has been well documented [1,2]. However, medication adherence remains challenging [3–8]. Medication adherence studies show that active substance users have greater difficulty adhering to HAART regimens than non-users [9–15], and providers may hesitate to prescribe HAART to this population due to potential drug resistance.
Studies of observed therapy for HIV-infected substance users have been ongoing since 1998 [16–19]. While many of these studies show that observed therapy is feasible, acceptable, and improves viral load and CD4 cell count, they are non-randomized or report on interventions connected with drug treatment. We report on a randomized trial of community-based, modified directly observed therapy of a once-daily regimen compared with standard of care for HIV-seropositive substance users.
Patients and study design
This study was a randomized, open-label, non-blinded, single center trial. Participants were recruited from HIV primary care clinics in Rhode Island and Massachusetts. Eligibility criteria included active substance use (heroin/cocaine use in the past 6 months, other drug use on four or more of the last 7 days), or alcohol misuse (positive response on the CAGE alcohol screening questionnaire and frequency/quantity of drinks), and lack of genotypic resistance to a once-daily regimen. Eligible participants were block-randomized to either modified directly observed therapy (MDOT) or standard of care (SOC), stratified by HAART experience (defined as on HAART therapy for at least two consecutive weeks, regardless of regimen). Prior to randomization, participants were placed on a once-daily regimen. Regimens included long-acting nucleoside reverse transcriptase inhibitors such as didanosine enteric-coated (ddI-EC), lamivudine and tenofovir, in combination with a non-nucleoside reverse transcriptase inhibitor or a boosted protease inhibitor. Treatment decisions were made by the study medical director in conjunction with the patient's primary care physician. SOC participants could receive any adherence interventions offered through their clinics. MDOT participants filled their own prescriptions and a study nurse ‘pill-packed’ the medications. An outreach worker (ORW) delivered and observed patients take them at a location of the participants' choice. All MDOT participants kept a 1-week medication supply for self-administration if an outreach visit was missed.
The outreach workers attempted visits every day (7 out of 7 days) for the first 3 months and tapered them over subsequent months, up to 12 months. Participants were not removed from the study for missing visits. Assessments occurred at screening, baseline, and at 1, 3, 6, 9 and 12 months after enrollment and included a questionnaire and venipuncture. Participants received incentives only at assessments. Given our extensive pilot experience supporting the use of MDOT , we felt that SOC participants should have access to MDOT if they were failing therapy. Therefore, crossover to MDOT was allowed after 3 months if an SOC participant had less than one log10 drop in plasma viral load (PVL). Analysis is thus limited to outcomes during the first 3 months of follow-up (i.e. prior to crossover).
Primary endpoints of the study were viral load suppression and changes in PVL. Virologic suppression was defined as either undetectable virus (< 50 copies/ml) or a greater than two log10-unit reduction in PVL from baseline. We also included change in CD4 cell count as a secondary endpoint. Non-adherence was defined as participants missing at least one dose in the prior month. The Miriam Hospital provided institutional review board approval, and adverse outcomes were reported per protocol. A Data Safety and Monitoring Board provided feedback during protocol development, recruitment, and analysis phases.
Demographic and baseline characteristics of the patients in the MDOT and control arms were summarized and compared. Intention-to-treat analyses were used to compare percentage viral suppression, percentage undetectable viral load, median plasma viral load, and mean change in CD4 cell count between those randomized to MDOT or SOC. We used repeated measures logistic regression to estimate the treatment effect for viral load suppression and undetectable virus outcomes. Median regression was used for log viral loads, and linear regression was used for CD4 cell counts (fit in the square root scale to correct for skewness). For each outcome we estimated a time-averaged treatment effect because none of the fitted models showed evidence of time-by-treatment interaction. For each outcome, treatment effects were summarized for the overall sample and then separately by HAART experience. All models were fit in R using generalized estimating equations, with robust standard errors used throughout . All individuals with at least one follow-up measure were used to fit the models.
Between 2001 and 2004, 87of 153 screened participants were enrolled in the study and were followed up to January 2005. Forty-three were randomized to SOC and 44 to MDOT (Table 1). Seventeen patients were naive to HAART therapy at baseline (eight in SOC; nine in MDOT). Demographic, virologic and immunologic baseline measures were similar between both trial arms, overall and within level of stratification. Most participants had a history of incarceration and were unemployed. The reasons for and number of missing 1- and 3-month assessments were similar in both groups and unrelated to prior history of HAART therapy.
Excluding those on medication holidays (n = 7), at 1 month, a higher proportion of SOC participants missed at least one dose during the prior month, compared with MDOT participants [Fisher exact test, odds ratio (OR), 7.6; 95% confidence interval (CI), 2.0–36.9]; similar findings were true at 3-month evaluation (OR, 7.9; 95% CI, 2.4–19.5).
At 1 month following baseline, 24 participants had viral load suppression [14/37 (38%) on MDOT, and 10/37 (27%) on SOC]. At 3-months, 25 of 39 (64%) in the MDOT arm and 15 of 37 (41%) in the SOC arm had suppressed viral load (Table 2).
The OR for treatment effect, averaged over months 1 and 3, was 2.2 (95% CI, 1.0 to 4.7; P = 0.05). Treatment effects were substantially more pronounced among HAART-experienced individuals (OR, 2.9; 95% CI, 1.2 to 7.0; P = 0.02). There was no evidence for an effect among those who were HAART naive (OR, 0.8; 95% CI, 0.1 to 5.1; P = 0.79).
In the SOC arm, median log PVL was 4.8 at baseline, 3.3 at month 1, and 3.3 at month 3; for MDOT, medians at the same time points were 4.8, 3.1, and 2.5. Among those with HAART experience, median log PVL among those on SOC was 4.7 at baseline, 4.2 at month 1, and 3.8 at month 3; on MDOT, medians were 4.8, 3.2, and 2.5. Median regression analysis shows that log PVL tended to be lower among those on MDOT (difference, –0.7; 95% CI, –1.5 to 0.1; P = 0.12), with a larger treatment effect among the HAART experienced (difference, –1.2; 95% CI, –1.9 to –0.1; P = 0.01). We did not find evidence of an intervention effect among HAART-naive individuals (difference, 0.1; 95% CI, –1.1 to 0.9).
No individual had a PVL measurement less than 50 copies/ml at baseline. At month 1, two participants from each arm had undetectable viral loads. At month 3, 11 participants on SOC (seven experienced and four naive) and nine on MDOT (six experienced and three naive) had undetectable viral loads. The time-averaged odds ratio for treatment effect was 0.7 (95% CI, 0.3 to 1.9; P = 0.54), with some variation associated with HAART history.
CD4 cell count
Findings from the regression analysis of CD4 cell counts are largely consistent with those related to viral suppression and log PVL. Overall change in CD4 cell count was greater for individuals on MDOT, with a more pronounced intervention effect among HAART-experienced participants. As the model was fitted in the square root scale, the effect of MDOT on change in CD4 cell count depended on baseline CD4. Among those with a baseline CD4 cell count of 133 cells/μl (baseline mean for the group overall); mean CD4 cell count at month 3 was 241 cells/μl in the MDOT arm and 197 cells/μl in the SOC arm; the estimated difference was 45 (95% CI, 5 to 85; P = 0.03). For HAART-experienced individuals, given a baseline CD4 cell count of 136 cells/μl, 3-month mean CD4 cell count was 232 cells/μl on MDOT and 172 cells/μl on SOC, for a treatment difference of 61 (95% CI, 16 to 105, P = 0.01). Again, no evidence of an MDOT effect was observed among those who were HAART naive.
In an intent-to-treat analysis, HIV seropositive substance users randomized to MDOT were more likely to achieve HIV PVL suppression of > 2 logs than those receiving clinic-based standard-of-care. The effect was driven primarily by the enhanced PVL suppression among HAART-experienced substance users. We did not observe a substantial treatment effect among HAART-naive participants, but because of the small sample size within this subpopulation, this cannot be interpreted to mean that efficacy is different in this subpopulation.
We chose to use virologic measures as our primary outcome for the adherence intervention. Our findings of plasma viral load suppression (undetectable viral load or a greater than 2 log drop from baseline) is promising, given that the literature suggests that in HAART-experienced individuals a viral load reduction of greater than 1.5 logs is associated with a favorable long-term clinical outcome . In fact, HAART-experienced participants on MDOT were hospitalized less often and for fewer days in comparison with participants on SOC. This led to a substantial overall cost savings in the MDOT arm [22,23].
A trial of observed therapy  in an inner city clinic in LA did not find that MDOT was necessary to achieve viral load suppression. However, the difference in study findings may be in large part due to the fact that this trial did not focus on active substance users. In terms of the substance users in our study, although they were typical of many inner city clinic patients, they were not representative of the most difficult group of substance users, those not engaged in medical care. In addition, we chose to classify substance users as one population (including alcohol users, injection drug users, etc.) and so it is difficult to draw conclusions about specific risk groups .
Another limitation of this study is that we limited our primary endpoint analysis to 3-month follow-up because of the importance of adding a cross-over arm, and the expectation that patients on therapy would achieve a significant viral load reduction by this time point. The long-term effectiveness of MDOT is an important question that still needs to be addressed. However, given the social instability of this population, the significant benefit of short-term MDOT in HAART-experienced substance users is very promising.
MDOT for HAART is an effective and viable strategy to reduce viral load and increase CD4 cell counts in HAART-experienced substance users with detectable viral loads, who have failed clinic-based adherence counseling. As sustained HAART treatment is associated with reduced risk of death among marginalized populations [26,27], MDOT should be included among adherence interventions for active substance users who are failing therapy. Future studies should better define which populations are most appropriate for MDOT as well as the long-term benefits.
The authors are grateful for the financial support of the National Institute of Drug Abuse (R01DA013767 and K23DA017622), the Lifespan-Tufts-Brown Center for AIDS Research (National Institutes of Health (NIH) grant AI-42853), and the Tufts Nutrition Collaborative, a Center for Drug Abuse and AIDS Research (NIH grant DA13868). We also grateful for the dedication and hard work of all members of the AARTS team and the involvement of the participants in the trial.
1. Wood E, Hogg RS, Harrigan PR, Montaner JS. When to initiate antiretroviral therapy in HIV-1-infected adults: a review for clinicians and patients. Lancet Infect Dis 2005; 5:407–414.
2. Carpenter CC, Cooper DA, Fischl MA, Gatell JM, Gazzard BG, Hammer SM, et al
. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000; 283:381–390.
3. Kerr T, Walsh J, Lloyd-Smith E, Wood E. Measuring adherence to highly active antiretroviral therapy: implications for research and practice. Curr HIV/AIDS Rep 2005; 2:200–205.
4. Giordano TP, Suarez-Almazor ME, Grimes RM. The population effectiveness of highly active antiretroviral therapy: are good drugs good enough? Curr HIV/AIDS Rep 2005; 2:177–183.
5. Lucas GM. Antiretroviral adherence, drug resistance, viral fitness and HIV disease progression: a tangled web is woven. J Antimicrob Chemother 2005; 55:413–416.
6. Sethi AK. Adherence and HIV drug resistance. HIV Clin Trials 2004; 5:112–115.
7. Stone VE, Smith KY. Improving adherence to HAART. J Natl Med Assoc 2004; 96(2 suppl):27S–29S.
8. Simoni JM, Frick PA, Pantalone DW, Turner BJ. Antiretroviral adherence interventions: a review of current literature and ongoing studies. Top HIV Med 2003; 11:185–198.
9. Malone SB, Osborne JJ. Improving treatment adherence in drug abusers who are HIV positive. Lippincotts Case Manag 2000; 5:236–245.
10. Arnsten JH, Demas PA, Grant RW, Gourevitch MN, Farzadegan H, Howard AA, Schoenbaum EE. Impact of active drug use on antiretroviral therapy adherence and viral suppression in HIV-infected drug users. J Gen Intern Med 2002; 17:377–381.
11. Bouhnik AD, Chesney M, Carrieri P, Gallais H, Moreau J, Moatti JP. Nonadherence among HIV-infected injecting drug users: the impact of social instability. J Acquir Immune Defic Syndr 2002; 31(suppl 3):S149–S153.
12. Carrieri MP, Chesney MA, Spire B, Loundon A, Sobel A, Lepeu G, et al
. Failure to maintain adherence to HAART in a cohort of French HIV-positive injecting drug users. Int J Behav Med 2003; 10:1–14.
13. Halkitis PN, Kutnick AH, Slater S. The social realities of adherence to protease inhibitor regimens: substance use, health care and psychological states. J Health Psychol 2005; 10:545–558.
14. Kerr T, Marshall A, Walsh J, Tyndall M, Montaner J, Hogg R, Wood E. Determinants of HAART discontinuation among injection drug users. AIDS Care 2005; 17:539–549.
15. Martini M, Recchia E, Nasta P, Castanotto D, Chiaffarino F, Parazzini F, Agnoletto V. Illicit drug use: can it predict adherence to antiretroviral therapy? Eur J Epidemiol 2004; 19:585–587.
16. Mitty JA, Macalino GE, Bazerman L, Loewenthal HG, Hogan JW, MacLeod CJ, Flanigan TP. The use of community-based modified directly observed therapy for the treatment of HIV-infected persons. J Acquir Immune Defic Syndr 2005; 39:545–555.
17. Lurie MN, Carter EJ, Cohen J, Flanigan TP. Directly observed therapy for HIV/tuberculosis co-infection. Lancet Infect Dis 2004; 4:137–138.
18. Altice FL, Mezger JA, Hodges J, Bruce RD, Marinovich A, Walton M. Developing a directly administered antiretroviral therapy intervention for HIV-infected drug users: implications for program replication. Clin Infect Dis 2004; 38(suppl 5):S376–S387.
19. Lucas GM, Mullen BA, Weidle PJ, Hader S, McCaul ME, Moore RD. Directly administered antiretroviral therapy in methadone clinics is associated with improved HIV treatment outcomes, compared with outcomes among concurrent comparison groups. Clin Infect Dis 2006; 42:1628–1635.
20. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986; 42:121–130.
21. Ledergerber B, Lundgren JD, Walker AS, Sabin C, Justice A, Reiss P. Predictors of trend in CD4-positive T-cell count and mortality among HIV-1-infected individuals with virological failure to all three antiretroviral-drug classes. Lancet 2004; 364:51–62.
22. Mwamburi M, Macalino G, Griffith J, Mitty J, Wilson I, Neuman P, et al
. Immediate versus delayed modified directly observed HAART therapy (MDOT): a cost-effectiveness analysis from the adherence to antiretroviral therapy for substance abusers (AARTS) study. Presented at the XVI International AIDS Conference
, Toronto, Canada, August 2006.
23. Mwamburi M, Macalino G, Wilson I, Mitty J, Griffith J, Neuman P. Cost-effectiveness analysis of modified directly observed therapy (MDOT) for HAART for patients with poor adherence. Presented at the XVI International AIDS Conference
, Toronto, Canada, August 2006.
24. Wohl AR, Garland WH, Valencia R, Squires K, Witt MD, Kovacs A, et al
. A randomized trial of directly administered antiretroviral therapy and adherence case management intervention. Clin Infect Dis 2006; 42:1619–1627.
25. Chander G, Himelhoch S, Moore RD. Substance abuse and psychiatric disorders in HIV-positive patients: epidemiology and impact on antiretroviral therapy. Drugs 2006; 66:769–789.
26. Riley ED, Bangsberg DR, Guzman D, Perry S, Moss AR. Antiretroviral therapy, hepatitis C virus, and AIDS mortality among San Francisco's homeless and marginally housed. J Acquir Immune Defic Syndr 2005; 38:191–195.
27. Kohli R, Lo Y, Howard AA, Buono D, Floris-Moore M, Klein RS, Schoenbaum EE. Mortality in an urban cohort of HIV-infected and at-risk drug users in the era of highly active antiretroviral therapy. Clin Infect Dis 2005; 41:864–872.