icon-folder.gif   Conference Reports for NATAP  
  9th Conference on Retroviruses and Opportunistic Infections
Seattle, Washington, February, 2002
Back grey_arrow_rt.gif
Adherence to HIV Therapy - Helping the Medicine Go Down
David Alain Wohl, MD - University of North Carolina AIDS Clinical Treatment and Research Center
  Like resolutions to abstain from sugar, gossip and watching Sex in the City, adherence to combination antiretroviral therapies is easier to pledge than to do. Few issues related to HIV therapy are more imperative to study than adherence, yet, the research conducted in this area has largely been limited to a smattering of 'adherence is bad' studies which often use a variety of high and low tech techniques to demonstrate what we already knew, namely, that adherence is bad. More recently, we have seen nascent attempts to impact adherence. Usually small efforts that briefly apply a single intervention and rely on self-report, these first studies, while not establishing a paradigm for enhancing adherence, are at least the first wobbly steps in the field of antiretroviral adherence augmenting research. At the 9th CROI, larger and more creative attempts at increasing adherence are being presented (as posters only) - indicating the field is starting to mature and that real world strategies for tackling the leading cause of virologic failure are being developed. Certainly given the magnitude of the problem, far too little was reported here and little that anyone will be rushing home to implement at his or her clinics.
The consequence of adherence (or non-adherence) is generally obvious - less adherence risks less viral suppression and increased disease progression. The Vancouver research group examined the synergy between levels of adherence and CD4 cell count (E. Wood, et al. Abstr 465-M) to determine if non-adherence is more dangerous among those with advanced disease. Not too surprisingly it was but the study helps quantify the risk. In a cohort of over 1200 patients who initiated HIV therapy between mid 1996 and 1999, survival and viral suppression were evaluated. Looking at baseline (before treatment) CD4 counts broken down into three groups: less than 50 cells, 50-200 and greater than 200, and adherence based on pharmacy refill records the investigators calculated risk of disease progression. (It should be noted that the Canadian group have a somewhat unique and fortunate opportunity to examine pharmacy refill records and have previously demonstrated the utility of this technique.) Among those with less than 75% adherence during the first year of therapy, death occurred in an astounding 41% of the patients with a CD4 count less than 50, 12% among those 50-200 and 5% for the patients with greater than 200 cells at baseline. In contrast, among those with better than 75% adherence during the first treatment year, death rates were 11%, 7% and 2% for those with CD4 counts less than 50, 50-200 and greater than 200, respectively. These data indicate that for those who are living dangerously close to the cliff edge, the fatal shove for many is poor adherence.
Interventions to enhance adherence were evaluated in a few studies. The U.S. AIDS Clinical Trials Group (ACTG) performed one of the largest studies presented. In a substudy of ACTG 388, Seattle's own Ann Collier and colleagues (Abstr 540-T) attempted to determine whether a telephone-based intervention could increase adherence among subjects enrolled in the parent treatment trial. In several clinics where resources are available, staff (usually a pharmacist or nurse) follow-up clinic visits with phone calls to identify problems with tolerance and adherence, so this intervention has relevance. The investigators randomized 282 participants to receive standard adherence counseling as is routinely provided at the local study centers versus this same standard counseling plus serial phone calls to subjects using a standardized script designed to encourage adherence and identify the emergence of issues related to taking the study medication. A maximum of 16 phone calls over 96 weeks were scheduled. Subject adherence to study drugs was measured by response to an adherence questionnaire gathering information about medication taking during the previous 4 days. The subjects were mostly male (80%) but half were non-white and 90% were antiretroviral naive. Almost three quarters (73%) of the phone calls were completed successfully. As is often the case in studies that employ questionnaires, self-reported adherence was relatively high with approximately 60% of subjects in each substudy arms reporting 100% adherence. Study defined virologic failure was experienced by 97 (34%) of subjects and was no more common among those receiving or not receiving phone calls. This does not mean that telephone adherence counseling is not worthwhile but it may suggest that the content of the script may have to be re-evaluated and that this intervention may do best when coupled with other strategies for encouraging adherence.
An intervention based on individual adherence counseling was studied among 244 subjects receiving HAART (C. Pradier et al. Abstr 541-T). Participants were randomized to routine clinical follow-up or a series of three individual sessions with "specially trained nurses". The proportion with self-reported adherence of 100% was better in the intervention group at six months than in the control arm, 75% versus 61%. Viral load modestly improved in the intervention arm and actually increased a bit among the controls but the proportion below 40 copies/mL was not different between arms. The reasons for the discrepancy between the self-report and viral load data are obvious. Mainly, after spending several sessions learning that adherence is a good thing and is expected, what would one suppose subjects receiving the counseling would report when asked if their adherence is better, even when it is not? The absence of more objective (pill count, pharmacy refill frequency, electronic monitoring devices) or ideally, multiple measures of adherence (see Lui H, et al. A comparison study of multiple measures of adherence to protease inhibitors, Annals of Internal Medicine, 2001;134:968-977) plagues most adherence related research and makes true assessment of any intervention difficult.
One study did employ electronic monitoring devices, or MEMs caps, to assess the adherence to HIV medications following mock trials during which placebos were used prior to initiation of the actual antiretroviral regimen among 54 HIV-infected women (R. Lubelchek, et al. Abstr 545-T). The participants were aware that they were receiving placebos and that adherence to these was being monitored with the MEMs devices, which record the date and time the medication bottle is opened. Those with less than 95% adherence to the mock mediations received a second mock trial using the first mock results as feedback. Among the 28 (52%) who achieved 95% or better adherence during the first or second mock, 57% had viral loads below 400 copies by 8 weeks following initiation of antiretrovirals compared to 19% of those not reaching this level of adherence. Similarly, 32% of the successful mock adherers versus only one (4%) of the mock non-adherers had a viral load below 40 copies at 6 months. These results may provide validation to all of you out there who try to gauge how adherent subjects will be by using trials of vitamins or prophylactic medications. (I would advise not to use M&Ms as mock medication, as some advocate, as it has been my experience that adherence to these, especially the green ones, is excellent and may not be predictive of future antiretroviral success.)
Another approach to adherence that is popular in many of our state and federal correctional facilities is the observed administration of antiretrovirals (i.e. watching the patient swallow the medications). This same technique is now being studied among the free and was applied to 54 intravenous drug users in Vancouver who were starting HAART (B. Conway, et al. Abstr 545-T). Directly observed therapy (DOT) can mean different things to different people and in this study it meant watching a patient take their antiretrovirals once a day while receiving their methadone. This posed few problems for the 53% on once daily regimens but for the remainder on twice a day therapies the ingestion of the second dose was not observed. Viral suppression to less than 50 copies/mL after a median of 23 months follow-up was only achieved by 44% of the participants and was not different among once versus twice daily HAART recipients. The proportion with less than 400 copies was better at 65%. The authors note that recreational drug use (mostly cocaine) was reported to continue even among those with the lower viral loads. Determination of the utility of DOT for HIV therapy awaits a well-conducted clinical trial.
In a very interesting and important study from Johns Hopkins (G. Lucas, et al. Abstr 547-T), switching from non-substance abuse to substance abuse among almost 700 clinic patients was associated with worsening utilization of HIV therapy, poorer adherence by self-report, lower CD4 cell counts and less frequent viral suppression. Abstaining from substance abuse was found to be associated with the exact opposite. This study is a resounding endorsement for greater efforts to make substance abuse treatments available and accessible for persons with HIV infection.
Lastly, we may have be careful what we wish for when it comes to antiretroviral adherence as better adherence may not always be a good thing. In a study of 183 patients who were subject to unannounced pill counts those who were less than 50% adherent had high levels of treatment discontinuation and low risk of developing new drug resistance mutations (D. Bangsberg, et al Abstr 546-T). However, those with better but still imperfect adherence had a great chance of remaining on therapy and cultivating new antiretroviral resistance. This is not surprising as those who do not really take their medications do not expose their virus as much to the drugs and run a low risk of resistance. Conversely, if someone is more adherent, say to 60-70%, they provide much greater drug exposure to their virus and almost beg for it to become resistant. Interventions that prove to be effective at enhancing adherence will have to be mindful of this non-linear relationship between adherence and risk of resistance.