← Back to Search
Highly Active Antiretroviral Therapy In A Large Urban Clinic: Risk Factors For Virologic Failure And Adverse Drug Reactions
Published 1999 · Medicine
Reduce the time it takes to create your bibliography by a factor of 10 by using the world’s favourite reference manager
Time to take this seriously.
Highly active antiretroviral therapy (HAART) has produced dramatic reductions in morbidity and mortality rates associated with HIV-1 infection in the United States (1, 2). In clinical trials, combinations of protease inhibitors, nucleoside analogues, and non-nucleoside reverse transcriptase inhibitors have reduced plasma HIV-1 RNA levels to less than 500 copies/mL in 60% to 90% of patients (3-7). However, such therapy involves complicated dosing schedules, side effects in a substantial number of patients, and a high degree of adherence to maintain viral suppression (3, 4, 8). Given these difficulties, it has become evident that HAART may be considerably less successful outside of the clinical trial setting (9). Before HAART, reports showed lower survival rates among ethnic minority groups, women, injection drug users, and persons in lower socioeconomic groups with HIV infection (10-12). The effect of demographic differences on the success of HAART is unclear. Our hypothesis was that the virologic response to HAART would be substantially worse among unselected patients in an inner-city clinic than among patients enrolled in clinical trials. We sought to identify demographic, behavioral, and clinical features that correlated with failure to suppress viral load outside of the clinical trial setting. Methods Patients and Data Collection The Johns Hopkins HIV Clinic provides longitudinal, primary care to a large portion of HIV-infected patients in the Baltimore area. All patients enrolling in the clinic undergo a comprehensive baseline assessment, with collection of detailed demographic, social, behavioral, and clinical data. Clinicians and social workers obtain baseline information through structured interviews by using standardized forms. Trained monitors using standardized data collection instruments extract data from patients' charts and from the hospital's automated databases at baseline and every 6 months. Items in the data extraction include all baseline information and information on follow-up diagnoses, medications, laboratory values, hospitalizations, and death. The clinic-based medical record maintains a section for each visit to document prescribed therapy by treatment name, dose, and number of refills. The record is also updated when prescriptions are filled over the telephone or mailed to the patients. We analyzed data from a cohort of protease inhibitor-naive patients in whom HAART was initiated in the clinic between March 1996 and February 1998. To be included in the analysis, patients had to begin therapy with ritonavir, indinavir, or nelfinavir plus nucleoside analogues, non-nucleoside reverse transcriptase inhibitors, or saquinavir. Saquinavir was not included as a primary protease inhibitor because of the low bioavailability of the formulation in use at the time of this study (Invirase, Roche Laboratories, Nutley, New Jersey) (13). In addition to being protease inhibitor-naive, patients had to lack exposure to at least one other drug used in the regimen. All patients who were treated primarily in our clinic during the study period and met the stated criteria were included in the analysis. Data obtained from our database were reviewed by one of the authors, and clinic charts were reviewed in all cases in which length of therapy or virologic outcomes through 12 months had not yet been abstracted as part of the biannual update process. Maintenance of our database and use of its contents for retrospective analysis of patient outcomes were approved by the institutional review board of The Johns Hopkins Hospital. Informed consent was not required for our analysis. Definitions For analysis of ethnicity, patients were divided into nonwhite and white groups. Ninety-six percent of nonwhite patients in our clinic population were African American. History of injection drug use was identified when patients were enrolled in the clinic. Active and inactive injection drug use at initiation of HAART were not distinguished. In the initial analysis of the data, it became clear that patients in the highest quartile of age (>40 years) had different virologic responses than the remainder of the cohort. Patient age was subsequently categorized as the lower three quartiles or the highest quartile. For each patient, virologic response to therapy was recorded in three time frames: 1 to 90 days, 3 to 7 months, and 7 to 14 months after initiation of HAART. For each time frame, the lowest available HIV-1 RNA level was recorded (Amplicor HIV-1 Monitor, Roche Molecular Systems, Branchburg, New Jersey). Values of 500 copies/mL or less were categorized as undetectable; all other values were considered detectable. If no value was available within a given time frame, the variable was categorized as missing. Adherence to clinic appointments was recorded for each patient as the percentage of scheduled clinic appointments that were missed. Patients receiving HAART in our clinic are typically scheduled to be seen 4 to 6 weeks after initiation of therapy and at least every 3 months thereafter. Appointments may be scheduled more frequently depending on active issues. In addition, patients are often scheduled for nursing, psychiatry, dermatology, gynecology, neurology, and gastroenterology appointments at the clinic. Adherence to these appointments was included in the analysis. Analysis of adverse drug reactions focused on the protease inhibitors. The research staff reviewed the records from the time at which therapy with these drugs started and determined whether adverse drug reactions had occurred on the basis of the temporal relation to therapy initiation and known side effects of the medications. Adverse reactions included allergic reactions (hypersensitivity, rash, or pruritus), gastrointestinal symptoms (nausea, vomiting, diarrhea, or cramping), and unspecified. Nephrolithiasis was included as an adverse reaction to indinavir, and paresthesias and taste perversion were included as reactions to ritonavir. Statistical Analysis Statistical analysis was performed by using Stata 5.0 (Stata Corp., College Station, Texas). The frequencies and distributions of categorical variables were compared by using the Fisher exact test. Continuous variables were compared by using one-way analysis of variance. Multivariate analysis was performed by using logistic regression. Interaction terms between pairs of significant variables were tested by combining variables in a logistic regression analysis. Variables included in the logistic regression analysis for suppression of viral load at 7 to 14 months were ethnicity, injection drug use, age older than 40 years, baseline CD4+ lymphocyte count, baseline viral load, and percentage of clinic appointments missed. Patients with no data available on viral load at 7 to 14 months (20 patients) were omitted from the multivariate analysis to minimize potential confounding effects with the missed clinic appointments variable. In this model, the odds ratios for baseline CD4+ lymphocyte counts were expressed per additional 50 cells, those for baseline viral load were expressed per additional log10HIV-1 RNA copies per mL, and those for missed clinic visits were expressed per additional 20% missed visits. Variables included in the logistic regression analysis for adverse drug reactions were primary protease inhibitor used, use of saquinavir, ethnicity, sex, and number of new drugs added to the protease inhibitor regimen. Role of the Funding Source Our funding source had no role in the collection, analysis, or interpretation of the data for this study. Results Baseline characteristics of the 273 patients who met the criteria for inclusion are listed in Table 1. The median length of follow-up was 454 days, and 95% of patients had at least 7 months of follow-up. Eighty-seven percent of the patients included in the analysis began taking a regimen involving at least three drugs, and in 66% of the patients, the regimens involved at least three drugs to which the patient did not have previous exposure. Because the inclusion criteria were broad, a wide array of regimens was used. Table 1. Characteristics of 273 HIV-Infected Patients in Whom Protease Inhibitor Therapy Was Initiated We identified trends that reflected both the temporal availability of the individual protease inhibitors and the history of prescribing practices as they developed in our clinic during the study period. Indinavir was the first protease inhibitor to be widely used in the clinic. The median start dates for therapy with ritonavir or nelfinavir trailed the median start date for indinavir therapy by 3 months and 7 months, respectively (P<0.001 for all comparisons). Patients who began taking regimens containing indinavir were more likely to be white (36%) than were patients who began taking ritonavir (23%; P=0.05) or nelfinavir (19%; P=0.006); this probably reflects the changing epidemiology of HIV-1-infected patients treated in the clinic over this time period. Patients who began therapy with ritonavir were more likely to have saquinavir included in their regimen (68%) than were patients who began therapy with indinavir (1%) or nelfinavir (11%) (P<0.001 for both comparisons). Patients in the three protease inhibitor groups were similar for sex, age, injection drug use, baseline CD4+ lymphocyte count, and HIV-1 RNA level. Undetectable levels of HIV-1 RNA were found in 42% of the cohort at 1 to 90 days, in 44% at 3 to 7 months, and in 37% at 7 to 14 months. Twenty-three percent of the cohort had undetectable HIV-1 RNA levels at all three time points. Viral load measurements were unavailable for 8%, 7%, and 7% of the cohort in the three time frames, respectively. Univariate associations of demographic and clinical characteristics with suppression of HIV-1 RNA are shown in Table 2. Age in the highest quartile, white ethnicity, and adherence to clinic appointments were strongly correlated with undetectable HIV-1 RNA levels in all time frames. The beneficia