The Body Covers: The 6th Conference on Retroviruses and Opportunistic Infections
Poster Session 15: Antiretroviral Therapy: Adherence and Quality of Life
February 1, 1999
Abstract 92: How Much Adherence is Enough? A Perspective Study of Adherence to Protease Inhibitor Therapy Using MEMSCaps
Authored by D. Paterson, S. Swindells, J. Mohr, M. Brester, E. Vergis, C. Squier, M. Wagener, N. Singh
The current mantra of all clinical HIV treaters and the recurrent theme of patient visits is compliance, compliance, compliance. Available antiretroviral therapy has the ability to achieve and sustain suppression of viral replication in many individuals. These authors begin with the premise that adherence is a critical factor in virological suppression. Conversely, adherence that is suboptimal may promote the development of viral resistance and virologic failure. The implications of resistance include the inability to use these same therapeutic options in the future, and the potential of drug resistant virus and development of cross-resistance. But how much adherence is enough? In this study, the authors attempted to quantitate in a prospective clinical trial, the relationship between adherence and clinical virologic outcome, the risk factors for poor adherence and patient assessment of adherence.
Adherence was measured electronically in eighty-eight patients using MEMS caps. This device is a microelectronic implant in the bottle cap that records the date and time that the bottle is opened. It does not, of course, guarantee that the medication is consumed. Adherence was calculated as the number of doses taken as a percentage of the number of doses prescribed. Virologic failure was defined as HIV RNA >400 copies/mL at the latest measurement. Risk factors variables that were looked included demographics, medications, CD4 cell counts, viral loads, psychiatric morbidity, alcohol/substance abuse, health, physician adherence assessment, and patient self-report. There were ninety-four males and six females; seventy-eight caucasians, twenty-two African-Americans, with a mean age of 41. Fourteen were IVDU's.
Sixty-seven were men who had sex with men. Nineteen were men who had sex with females. Twelve of the patients had CD4 counts <50, 25 had CD4 counts >500, with the remainder scattered in-between. HIV RNA copies were undetectable in twenty-two (<400 level), eighteen had >100,000, with the remainder again scattered in-between. The authors graphically represented the correlation between adherence and virologic failure. In patients with 95% compliance, the virologic failure rate was approximately 20%. At 90-95% adherence, the failure rate rose to 36%. At 80-90% adherence, the failure rate jumped to 50%, and at 70-80% adherence it jumped up to 75% viral failure. In patients with less than 70% compliance, there was a 94% failure rate. The values show a highly statistically significant relationship between suboptimal adherence and the risk of virologic failure (P = .0001). Factors significant for predicting poor adherence included: 1) active depression, 2) alcoholism, 3) less education (although 31% of the college educated patients had greater than 95% adherence compared with 17% of high school educated patients), 4) African-American race, 5) high baseline HIV RNA, and 6) low baseline CD4 count. Factors that did not seem to show an association with adherence included: 1) sex, 2) risk factor for HIV infection, 3) employment status, 4) income, 5) active illegal drug abuse, 6) prior psychiatric therapy, or 7) history of opportunistic infection.
Patient/physician assessments were also done. Patient self-report adherence was looked at, and of patients who denied being non-adherent, 29% were less than 80% adherent and 45% of patients were 80-95% adherent. On looking at physician judgement of adherence, of the patients perceived as having greater than 80% compliance, 32% actually had less than 80% adherence. In patients perceived as less than 80% adherent, 20% had greater than 95% adherence and 34% had 80-95% adherence. When looking at the possible variability of adherence, it was actually found that no particular protease inhibitor regimen was associated with better adherence than any other. Interestingly, the mean adherence was 73% for b.i.d. regimens and 71% for t.i.d. regimens, not statistically significant.
The authors concluded that: 1) >95% of adherence to protease inhibitor therapy was associated with virologic success, and failure rates increased relative to decreasing levels of adherence. 2) Patient self-report and physician judgment are unreliable measures of adherence. 3) Modifiable variables associated with poor adherence include active depression and active alcoholism. Efforts to improve adherence should focus on untreated depression and alcohol abuse.
This study, along with several others, should be humbling to HIV health providers. It is quite clear, even when we have established relationships with our patients, we may not actually be able to judge their level of compliance. This has a tremendous impact on outcome, as was clearly demonstrated in this study, with almost linear relationship, or lack thereof, between compliance and virologic failure. This should lend further impetus to the need for development of easier regimens with respect to numbers of doses and pill burden, as well as side-effect profiles.
Authored by A. Kaplan, et al.
Authored by L. Miller, et al
These two abstracts again sought to quantify, and characterize, the relationship between adherence, the ability of a clinician to determine compliance and virologic outcome. In the study by Kaplan, electronic MEMS caps were used and compared with monthly pill counts. The viral load was measured monthly as well. Clinicians were asked to estimate protease inhibitor adherence. In the first study, 71 subjects were enrolled. Mean adherence varied dramatically from 33% to 100%, and decreased in many subjects over time. Viral RNA levels, however, became undetectable or decreased in more than one log in 43 subjects. A comparison of adherence between initial responders in those subjects whose viral load did not decrease revealed greater adherence, not surprisingly, among the responders (86% versus 75%). However, there was significant variability in adherence observed among those subjects with good initial responses (33% to 100% compliance). When looking at which one of their measures was able to predict with any sensitivity, they found that MEMS was close to 100%. Pill count, interestingly, only had a sensitivity of 70%, and patient interview only correlated 9% of the time with compliance. The authors also looked at the effect of alcohol use on adherence. Those who drank nearly everyday had about a 30% compliance, those who drank 1-4/week had about 70% compliance, and those who drank less than once a week had approximately 90+% compliance.
The conclusions again were that interviews as a measurement of adherence usually led to substantial overestimates of the proportion of medications taken. This study also concluded that interventions to improve adherence should be aimed at enhancing the patient/provider relationship and helping patients to fit the regimen into their daily routine. Determining compliance is always one of the thorniest problems, perhaps more difficult than interpreting genotypes and designing appropriate regimens. Providers must re-double their efforts to at least correct or address those issues that have been identified, such as what to do if a patient is asleep when a dose is due -- should you or should you not take it? Clinicians should inform patients that they can call for medication refills. There is also counseling on alcohol abuse and drug abuse, although clearly this is one of the more difficult aspects. Reducing pill counts and modifying regimens can be attempted if it is clear that the patient won't take the pills (even if it is a wonderful regimen, if the medications are not taken, then perhaps a less optimal regimen should be designed that is easier to take). Confusion about directions in 11% of patients is very alarming, and indicates that printed instructions should be available for all patients.
This article was provided by The Body PRO. Copyright © Body Health Resources Corporation. All rights reserved.