Adherence to protease inhibitors.

Adherence to Protease Inhibitors TO THE EDITOR: In an excellent paper, Paterson and colleagues (1) demonstrate that adherence to protease inhibitor therapy of 95% or greater optimizes virologic outcome for patients with HIV infection. In our opinion, however, this conclusion cannot be generalized to all HIV-infected patients, especially those who were previously treated with antiretroviral agents. We performed a prospective cohort study on adherence to antiretroviral agents in 140 consecutive HIV-infected patients who were attending the outpatient service of the Department of Infectious Diseases, Catholic University of Rome, Rome, Italy, and were taking a regimen containing protease inhibitors. Median duration of antiretroviral therapy was 43 months (range, 2 to 121 months). Seventyseven percent of patients had previously received one or more other antiretroviral regimens, and 89% had been taking a protease inhibitor for at least 3 months. Nonadherence was assessed by questionnaire-based patient self-report and was defined as having missed at least one antiretroviral dose during the previous week. The validity of this simple, cheap, well-accepted, and feasible method has been shown (2). Thirty-five percent of patients were nonadherent, and nonadherence did not correlate with detectable HIV RNA levels at enrollment (odds ratio, 2.16 [95% CI, 0.99 to 4.72]) or at a median follow-up of 13 months (odds ratio, 1.14 [CI, 0.44 to 2.91]). Virologic failure is often considered a consequence of nonadherence to treatment; however, especially among pretreated patients, other factors (such as viral resistance and fitness, potency of antiretroviral agents, and HIV RNA level at baseline) can greatly influence efficacy. The lack of correlation between adherence and virologic outcome in our study is probably due to the heavy previous antiretroviral therapy in our patients. Furthermore, Paterson and colleagues did not report the distribution of completely antiretroviral-naive patients among the groups with different adherence levels. Nevertheless, we agree with the authors that, in addition to virologic reasons for treatment failure, physicians should always accurately investigate nonadherence before considering a change of therapy.