Section by David Alain Wohl, MD, Clinical Assistant Professor at University of North Carolina, with selected contributions by Harvey S. Bartnof, MD

Over the last four years since the first PI drug became available, it has become increasing clear that anti-HIV drugs are only effective if they are taken regularly.

At the 7th CROI, a study found that unless at least 95% of doses are consistently taken, drug resistance might develop and the drugs are likely to fail. At the 8th Annual Retrovirus Conference, 11 studies specifically addressed the important topic of adherence to antiretroviral therapy (ART) dosing.

Directly Observed Therapy
Margaret Fischl, MD, Director of the Miami AIDS Clinical Trials Unit (ACTU) in Florida, presented an update of her study that compared the outcomes of treatment-naïve (no previous treatment) patients at the Miami ACTU with those in the state prison (abstract 528). Within the prison system, inmate patients are administered antiretroviral therapy (ART) by a method called directly observation therapy ("DOT"). (DOT has been used successfully for years in the treatment of tuberculosis or "TB"). For the DOT patients, prison health practitioners observe the swallowing of each medication dose. Dr. Fischl examined HIV viral load responses among 50 prisoners receiving their study medications via DOT and 50 ACTU clinic outpatients who were receiving study medication in the conventional, unobserved way. The two groups were different demographically, with more of the incarcerated patients likely to be African-American, Latino, male and have a history of injection drug use (IDU). Further, the patients in prison at baseline had lower CD4 cell counts and higher viral loads. The results were that after 24 weeks, 90% of the prisoners had viral loads that were undetectable (limit 50 copies per milliliter), compared to 77% of those who were not in prison. These differences in response rates persisted up to 90 weeks of follow-up and were highly statistically significant. In general, simpler regimens of 3 drugs had better response rates than more complex 4-drug combinations.

There are many confounding factors that could have contributed to the inmates’ success. Those include the regimented structure of prison life, the influence of Department of Corrections medical staff and possibly limited access to illicit drugs including crack cocaine. Also, it is unclear how the patients in each group were chosen. The differences between the two groups of patients extend beyond the presence or absence of DOT and, therefore, DOT alone should not be regarded as the crucial determinant of the observed results.

Other studies have not shown such stellar viral load results with DOT in incarcerated patients. Dr. S. L. Hader of the Centers for Disease Control and Prevention (CDC) reported that DOT among residential treatment facilities in New York City led to an adherence rate of 99% (abstract 476). Yet, one-third never achieved an undetectable viral load. (In this study, 75% had taken ART previously.) Having a treatment interruption of at least 8 days was significantly associated with VL failure. In another example, David Wohl, MD reported at last year’s annual meeting of the Infectious Disease Society of America that DOT in the North Carolina Department of Corrections did not increase the rate of viral undetectability, when compared to standard self-administration (abstract 357). Moreover, many of the inmates complained that DOT renders them conspicuous as being HIV-infected, since they stand on line for medication. Consequently, many of them chose not to present for DOT. Under such conditions, DOT may present an obstacle rather than a path to improved adherence.

Support and Education to Increase Adherence
Dr. Alan L. Gifford and colleagues at the University of California at San Diego studied the effect of a group patient education program for HIV infected persons on adherence to multiple antiretroviral regimens (abstract 479). These sessions met for 2 hours on 6 occasions and were run by a nurse and an HIV positive peer counselor. This program was compared to two other interventions: a more standard support group without the educational component and the provision of printed materials related to adherence. A total of 168 patients were enrolled. Adherence was measured by self-report and drug levels at study visits. The group education program had better adherence compared to those only receiving printed materials but was not significantly better than the more conventional support group intervention. Adherence at immediate post-intervention evaluation was predictive of on-study adherence. At no point was adherence reported as above 90% for any of the groups. At 6 months there was no significant difference in adherence rates between the three arms. The results suggest that group support can be helpful in augmenting adherence with or without a structured educational component. Additionally, the benefit of such interventions is likely lost unless reinforced by ongoing interaction. These studies suggest that adherence and treatment education groups help adherence, but they also suggest that longer-term educational groups may be crucial to maintain continued success for many years.

Mental Health Matters, Other Risk Factors
Factors associated with non-adherence in the Multi Center AIDS Cohort Study (MACS) were presented (abstract 484). In an examination of the self-reported adherence to ART among 478 men (84% Caucasian), 16% reported incomplete adherence. Independent significant predictors of self-reported non-adherence included Black race, 14 or more alcoholic drinks per week and scoring high on surveys of depression and HIV burnout (e.g. "I am tired of always having to monitor my sexual behavior"). Other studies have not found an association between adherence and race. (It is possible that race might be a surrogate for lack of education, IVDU, or other variables such as behavioral and/or attitudinal factors.) Patients with depression or burnout were approximately 4 times more likely to report non-adherence. These data emphasize the need for clinicians to assess the potential root causes of non-adherence – and be vigilant in identifying and working through depression and treatment fatigue.

Regarding illicit drug use (IDU), Dr. K.A. Gebo of Johns Hopkins University in Baltimore, Maryland reported that patients with a history of IDU were not associated with non-adherence (abstract 477). However, among those with an IDU history, two factors were significant associated with non-adherence: heroin use in the previous six months and the patient’s belief that stress would impede perfect adherence. Among those without a history of IDU, the following factors were significantly associated with non-adherence: eating fewer than 2 meals daily and less than "strongly" believing that it is important to take medications as prescribed. Factors not significantly associated with adherence included race, gender (sex), age, using alcohol, tobacco or marijuana, and CD4 counts.

Adherence studies do have limitations. For example, after spending hours advising patients to take their medications, it may be difficult for them to admit when they have not and lead them to take the therapies on days when they know levels will be checked (so called ‘white coat adherence’). A perfect method to measure adherence does not exist, but a combination of measures such as medication event (electronic) monitoring system ("MEMS") devices, pharmacy records, self-report and/or unannounced drug levels and pill counts can provide more information.

Other adherence findings presented were: