| |
Episodic HAART in SMART Study Has Worse Quality of Life
|
| |
| |
The Impact of Episodic CD4 Cell Count-Guided Antiretroviral Therapy on Quality of Life
JAIDS Journal of Acquired Immune Deficiency Syndromes:Volume 47(2)1 February 2008pp 185-193
Burman, William J MD*; Grund, Birgit PhD; Roediger, Mollie P MS; Friedland, Gerald MD; Darbyshire, Janet PhD; Wu, Albert W MD ; for the SMART Study Group
From the *Denver Public Health Department and the Department of Medicine, University of Colorado Health Sciences Center, Denver, CO; School of Statistics, University of Minnesota, Minneapolis, MN; Division of Biostatistics, School of Public Health, University of Minnesota, Minneapolis, MN; AIDS Program, Yale University School of Medicine, New Haven CT; MRC Clinical Trials Unit, Medical Research Council, London, United Kingdom; and the Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore MD.
"QOL favors continuous antiretroviral therapy among patients with the characteristics of those in the SMART study...
.... The SMART study demonstrated that episodic antiretroviral therapy guided by CD4 lymphocyte counts has an increased risk of opportunistic illnesses, death, and serious cardiovascular and metabolic events compared with continuous antiretroviral therapy.14 The present analysis shows that the QOL outcomes of the SMART study also favor continuous therapy, and thus mirror the clinical outcomes."
Abstract
Objective: To evaluate the effect of episodic antiretroviral therapy on quality of life (QOL).
Design: This was a substudy of the Strategies of Management of Antiretroviral Therapy study, in which patients were randomized to continuous versus CD4 cell count-guided episodic antiretroviral therapy. QOL assessments included an analog scale for current health and the Short-Form 12 Item Survey, a standard abbreviated QOL instrument.
Results:
A total of 1225 patients had QOL assessments over a mean follow-up time of 2.4 years. Most (76%) were on antiretroviral therapy at enrollment; the median CD4 lymphocyte count was 575 (interquartile range: 455 to 784) cells/mm3; and mean current health was 75 on a scale from 0 to 100, and 50% reported very good or excellent general health.
Through follow-up, whenever QOL outcomes differed, the results were inferior among patients in the episodic therapy group compared with the continuous therapy group (current health, Physical Health Component Score [both P = 0.05], general health perceptions, physical functioning, and energy [all P = 0.03]).
HIV disease progression (opportunistic disease or death) was more common in the episodic therapy arm and was preceded by marked declines in QOL, but excluding participants with disease progression had minimal effect on QOL comparisons.
Conclusion: CD4 count-guided episodic use of antiretroviral therapy resulted in inferior QOL compared with continuous therapy.
BACKGROUND: Episodic use of antiretroviral therapy was suggested as a strategy to retain the benefits of combination antiretroviral therapy while decreasing its risks. Antiretroviral drugs have appreciable risks of effects.1,2 Common adverse effects, such as nausea, diarrhea, or changes in body appearance (lipodystrophy), may not be associated with serious toxicity but are a deterrent to adherence.3,4 Other adverse effects, such as hypersensitivity, metabolic abnormalities, and hepatotoxicity, can be severe and even life-threatening.5-7 Complete HIV suppression requires a high level of adherence,8,9 and it may be difficult to sustain such high-level adherence for years.10 These limitations of antiretroviral therapy led to the hope that episodic therapy might improve the quality of life (QOL) of persons living with HIV infection.11-13
The Strategies for Management of Antiretroviral Therapy (SMART) study was a large international randomized trial comparing CD4 count-guided episodic antiretroviral therapy versus continuous antiretroviral therapy. Enrollment in the SMART trial was stopped and the episodic therapy strategy was modified because of higher risks of opportunistic disease or death (primary endpoint: hazard ratio [HR] = 2.6, 95% confidence interval [CI]: 1.9 to 3.7) and of major cardiovascular and metabolic events (HR = 1.7, 95% CI: 1.1 to 2.5) among participants receiving episodic antiretroviral therapy.14 Here, we report QOL outcomes of participants randomized to the CD4 count-guided episodic antiretroviral therapy compared with those randomized to continuous antiretroviral therapy.
DISCUSSION
The SMART study demonstrated that episodic antiretroviral therapy guided by CD4 lymphocyte counts has an increased risk of opportunistic illnesses, death, and serious cardiovascular and metabolic events compared with continuous antiretroviral therapy.14 The present analysis shows that the QOL outcomes of the SMART study also favor continuous therapy, and thus mirror the clinical outcomes. Overall summary measures of QOL (current health by visual analog scale, PCS) and specific domains of the SF-12 instrument (general health perception, physical functioning, and energy) were inferior among patients randomized to CD4 count-guided episodic therapy. The treatment differences in QOL were apparent in the first year, often by the first follow-up visit at 4 months. Whenever the treatment difference in QOL outcomes was significant, the episodic therapy strategy was inferior to continuous therapy, and this finding was consistent across all investigated subgroups of patients. Finally, the episodic therapy strategy was not associated with decreased symptom burden.
The design of the SMART study allowed analyses of the effects on QOL of stopping and starting contemporary antiretroviral therapy, largely among antiretroviral-experienced patients. For most patients (76%) who enrolled while on antiretroviral therapy, the episodic versus continuous treatment group comparison describes the effects of stopping versus continuing combination antiretroviral therapy. Stopping antiretroviral therapy was associated with a relative decline in QOL among those who enrolled with an HIV RNA level ≦400 copies/mL on antiretroviral therapy. The inferior QOL outcomes in patients who interrupted virologically successful antiretroviral therapy (ie, HIV RNA ≦400 copies/mL) mirror the clinical outcomes of the SMART study; among patients who entered the study on antiretroviral therapy, the relative risk of opportunistic disease or death in the episodic therapy arm versus the continuous therapy arm was substantially higher among patients with baseline HIV RNA ≦400 copies/mL than among those with higher viral loads.14
For patients who enrolled while off antiretroviral therapy (24%), the episodic versus continuous therapy group comparison evaluates the effects of (re)starting antiretroviral therapy at a CD4 lymphocyte count >350 cells/mm3. Unlike earlier studies of high-dose zidovudine monotherapy or combination therapy with a ritonavir-based regimen among patients with relatively high CD4 lymphocyte counts,20,21 starting contemporary combination antiretroviral therapy in this subgroup of participants was not associated with a decline in QOL.
The large sample size of this QOL study allowed analyses of the associations between QOL and the occurrence of HIV disease progression events. In all QOL measures, there was a marked decline in QOL preceding an opportunistic disease or death, providing clinical validation of the QOL instruments used. Because there was a higher rate of opportunistic disease and death in the episodic therapy arm, we explored whether the inferior QOL outcomes in the episodic therapy arm were attributable to the higher rate of opportunistic disease and death. Excluding patients with these clinical events from the analysis resulted in only small changes in the treatment difference estimates, and episodic therapy resulted in inferior QOL outcomes even among persons who did not have HIV disease progression events.
Two prior studies assessed QOL outcomes associated with fixed-duration antiretroviral treatment interruptions. In a nonrandomized study of 12 patients interrupting therapy for 30 days (or until the viral load had increased to >3000 copies/mL), there was no improvement in overall QOL during treatment interruptions, although self-perceived health status did improve.11 In a randomized trial (n = 46) comparing continuous antiretroviral therapy with a fixed-duration cycle (4 weeks on antiretroviral therapy and 8 weeks off), there were no significant differences in QOL outcomes between the 2 arms.22 Our study used a different form of episodic therapy, CD4 count-guided interruptions, and had much greater statistical power to detect differences in QOL outcomes.
This study has several limitations. First, we evaluated only 1 type of antiretroviral treatment interruption, a specific CD4 count-guided episodic therapy strategy. In our study, this strategy resulted in long initial treatment interruptions for most patients (median time to (re)initiation of therapy of 16 months). The effects of other treatment interruption strategies on QOL may be different from those seen in our study. Second, patients in this QOL study were drawn only from study sites in the United States. Although these patients had demographic and clinical characteristics similar to those of patients in HIV care in the United States,23 QOL outcomes of treatment interruption may differ in other settings. Finally, we compared the 2 treatment strategies for each of the 11 QOL outcome measures separately, without adjustment for multiple comparisons. For each outcome measure, however, whenever the treatment difference was statistically significant, it favored the continuous therapy. The consistency of these findings across the QOL measures argues that episodic therapy is associated with inferior QOL outcomes.
The results of this randomized comparison of episodic versus continuous therapy are contrary to the hypothesized benefits of antiretroviral treatment interruption on QOL and symptom burden. The impact of antiretroviral therapy on QOL is complex and may depend on the medications used, duration of therapy, and stage of HIV disease at which therapy is started.21,24-26 Although antiretroviral medications have appreciable rates of adverse effects, HIV infection has adverse effects as well, in addition to its major complications of opportunistic disease and death. The results of this study demonstrate that the balance of these opposing effects on QOL favors continuous antiretroviral therapy among patients with the characteristics of those in the SMART study. This finding was consistent across demographic and clinically defined subgroups and was not solely attributable to the increased risk of clinical events in the episodic therapy group.
RESULTS
Baseline Characteristics
Of the 1422 patients enrolling in the SMART study at 65 sites in the United States participating in the QOL assessment, 1225 patients (86%) coenrolled in the QOL substudy between January 2002 and August 2005 (606 in the episodic therapy strategy and 619 in the continuous therapy strategy). Treatment groups were well balanced with respect to baseline characteristics, with a slightly higher prevalence of prior opportunistic diseases in the episodic therapy group (27% vs. 23%; see Table 1). Compared with all patients in the SMART study,14 participants in this QOL analysis were more often black (42% vs. 29%), less often on antiretroviral therapy at enrollment (76% vs. 84%), and less likely to have HIV RNA level ≦400 copies/mL at baseline (56% vs. 72%). At study entry, the average current state of health score was 75 (on the 0-to-100 visual analog scale), and 50% of participants assessed their general health as very good or excellent.
Follow-Up
The mean follow-up time was 29 months (range: up to 4 years); 88% of participants were followed for >1 year, and 61% were followed for >2 years. The total follow-up time was approximately 1500 person-years in each arm. QOL assessment forms were available from 94.3% of the scheduled visits (94.0% of episodic therapy visits and 94.6% of continuous therapy visits). Twelve participants (1.0%) withdrew consent or did not have study visits for more than 8 months.
Participants in the episodic therapy arm used antiretroviral therapy much less than patients in the continuous therapy arm (38% vs. 91% of follow-up time; P < 0.001), and the difference in antiretroviral therapy use was greatest in the first year after enrollment. In the episodic therapy arm, the median time to first (re)initiation of antiretroviral therapy was 16 months.
Outcome Measures
The SF-12 questionnaire items were summarized using standard techniques into 8 domains18,19 (listed in Table 1). Two scores, the Physical Health Component Summary (PCS) and Mental Health Component Summary (MCS), further summarize the 8 domains of the SF-12 into 2 principal domains; both are standardized to have a mean score of 50 and standard deviation of 10 in a US reference population.19 The PCS gives high weights to general health perception, pain, and measures of physical functioning, whereas the MCS gives higher weights to measures of mood, energy, and mental functioning. Higher values denote better QOL in each of the 8 SF-12 domains, PCS, MCS, and on the current health visual analog scale.
Changes in QOL Measures Through Follow-Up
Figure 1 displays mean change from baseline through month 36 for 4 QOL measures: current health (visual analog scale), general health perception, and the PCS and MCS scores. Through month 12, current health, general health perception, and PCS outcomes favored the continuous therapy group. The difference between the 2 treatment groups was apparent at the first visit after randomization (month 4). The difference in MCS scores between the 2 treatment groups, although in the opposite direction, was small and not significant (P > 0.4 through months 12 and 36).
Across all QOL measures, 95% CIs for the treatment differences through months 12 and 36 appeared similar (see Fig. 2), indicating that the separation of treatment groups occurred during the first year and changed little with further follow-up. Statistical tests showed no evidence for a change of the treatment effect over time (P > 0.1 for interaction), except in the SF-12 domain of pain, for which the cohort of patients with at least 2 years of follow-up reported more pain in the continuous therapy group (P = 0.06, data through month 36), whereas patients with <2 years of follow-up reported more pain in the episodic therapy group (P = 0.09).
Subgroup Analyses
The episodic therapy and continuous therapy groups were compared within subgroups of patients defined by the following characteristics: race, gender, age, prior AIDS, whether on or off antiretroviral therapy at enrollment, baseline CD4 cell count, nadir CD4 count before enrollment, baseline HIV RNA level (among patients on antiretroviral therapy), and duration of prior antiretroviral therapy.
Within subgroups, the treatment difference between episodic and continuous treatment was usually not significant. Whenever the treatment difference was statistically significant within a subgroup, the continuous therapy group had better QOL outcomes across all subgroups and all QOL measures. Selected subgroup analyses are provided in Figures 3 and 4 in this section for current health (visual analog scale) through month 12.
The subgroup of patients who were taking antiretroviral therapy at baseline discontinued therapy if randomized to the episodic therapy arm and remained on therapy if randomized to the continuous therapy arm. Current health scores by visual analog scale seemed to decline among those who discontinued therapy and to improve among those who remained on therapy (see Fig. 3A). The treatment difference in change in current health in this subgroup was estimated at -1.45 (P = 0.08). The subgroup of patients who were not on antiretroviral therapy at baseline were to start therapy if randomized to the continuous therapy arm and to remain off therapy until the CD4 count was <250 cells/mm3 in the episodic therapy arm. Current health status seemed to improve for both groups (see Fig. 3B). The treatment difference was not significant (-1.31; P = 0.37).
Subgroup analyses by HIV RNA level were restricted to patients who were on antiretroviral therapy at study entry. In the subgroup of patients with HIV RNA ≦400 copies/mL at study entry, current health outcomes were inferior in the episodic treatment group (treatment difference for change in current health of -2.63; P = 0.01). In contrast, changes in current health were similar among patients entering the study with higher viral loads. Therefore, the treatment difference (between continuous therapy and episodic therapy) in current health was larger for patients entering the study with HIV RNA ≦400 copies/mL compared with those with baseline HIV RNA >400 copies/mL (P = 0.03 for interaction, last column in Fig. 3).
Changes in QOL and Clinical Events
Among the 1225 participants, 52 in the episodic therapy group (3.6 per 100 person-years) and 24 in the continuous therapy group (1.6 per 100 person-years) experienced a new opportunistic disease or death. The HR in this substudy (HR = 2.3, 95% CI: 1.4 to 3.7; P < 0.001) is similar to that in the entire SMART study (HR = 2.6, 95% CI: 1.9 to 3.7).14
There were marked declines in all QOL measures from baseline to visits preceding an opportunistic disease or death (decreases of 2.5 to 14.1 units; P < 0.005 for all QOL measures). To evaluate whether the inferior QOL outcomes of the episodic therapy strategy can be explained by the higher incidence of opportunistic disease or death, the treatment difference in QOL was estimated with and without patients who experienced an opportunistic disease or died. For all QOL measures, excluding patients with opportunistic disease or death had little effect on the treatment difference estimate (data not shown).
The composite endpoint of serious non-AIDS events (major cardiovascular events, liver failure, renal failure, cancers not associated with HIV infection, or death not attributed to AIDS) was somewhat more common among patients in this substudy who were randomized to episodic therapy (HR = 1.3; P = 0.25). Excluding patients who had a serious non-AIDS event had little effect on the treatment difference in QOL between the 2 randomized arms, however.
Symptom Frequency and Severity
At baseline, 9.6% of patients in the episodic therapy group and 7.6% in the continuous therapy group had grade 3 symptoms, and none had grade 4 symptoms. During follow-up, 112 patients in the episodic therapy arm reported grade 3 or 4 symptoms versus 102 in the continuous therapy group (HR = 1.15, 95% CI: 0.88 to 1.51; P = 0.30). Peripheral neuropathy (grade 3) was reported for 18 and 19 patients in the episodic and continuous therapy groups, respectively.
METHODS
Study Population
Between January 2002 and January 2006, 5472 participants in 33 countries were enrolled in the SMART trial. At selected study sites in the United States, a subset of 1225 participants coenrolled in a QOL substudy and had health-related QOL assessments. Eligibility criteria included a CD4 lymphocyte count of >350 cells/mm3, age >13 years, and willingness to start or discontinue antiretroviral therapy according to randomization assignment. At baseline, participants could be antiretroviral naive, on antiretroviral therapy, or have had prior antiretroviral therapy. The study was approved by institutional review boards at each site, and written informed consent was obtained from each participant.
Study Design
Study participants were randomized at a 1:1 ratio to continuous antiretroviral therapy with the goal of maximal viral suppression (VS) (the viral suppression strategy) versus CD4 count-guided episodic therapy, with the thresholds to stop therapy at a CD4 lymphocyte count >350 cells/mm3 and to (re)start therapy at <250 cells/mm3 (the drug conservation [DC] strategy). The primary endpoint of the SMART study was the occurrence of a new opportunistic disease or death, and the main results of the study have been previously reported.14
The primary objectives of the QOL substudy were to compare drug conservation versus viral suppression groups for health-related QOL and symptom severity. The planned sample size of 1200 was calculated to detect a treatment difference in mean change in QOL of 17% of the standard deviation of QOL change (2-sided tests, ƒ¿ = 0.05, 80% power, unadjusted for multiple comparisons). We report study outcomes through January 11, 2006, when patients in the episodic therapy arm were advised to (re)initiate continuous antiretroviral therapy.
Data Collection and Follow-Up of Participants
Participants in the QOL substudy of the SMART study had health-related QOL assessments at baseline; at months 4, 8, and 12; and annually thereafter. The assessments included the self-administered Medical Outcomes Study Short-Form 12 Item Survey (SF-12, version 1),15 a visual analog scale for current health (0 to 100, with 0 being worst possible health),16 and an interviewer-administered survey of current symptoms (eg, nausea, vomiting, diarrhea, oral discomfort, fever, headache, fatigue, peripheral neuropathy). Symptoms were graded by local research staff on a scale from 1 to 4 adapted from a Division of AIDS, National Institute of Allergy and Infectious Diseases (NIAID) toxicity table; higher numbers indicate greater symptom severity.17
Statistical Methods
Analyses are by intention to treat. Treatment groups were compared for changes in QOL measures over 2 time intervals: from baseline to 12 months and from baseline to 36 months. By study design, the difference in antiretroviral therapy use between the episodic and continuous treatment strategies was expected to be greatest during the first year.
The difference in QOL outcomes between treatment groups was estimated and tested in longitudinal linear mixed effects models for change in QOL from baseline, adjusted for baseline QOL and follow-up visit. The homogeneity of treatment effect over time was assessed with likelihood-ratio tests for interaction between treatment group and follow-up visit. In subgroup analyses, differential treatment effects between subgroups were assessed through tests for interaction between treatment arm and subgroup.
Within treatment groups, the mean change in QOL was estimated using fixed effect mean estimates in longitudinal models without baseline QOL adjustment; mean change through month 12 was calculated as average of the fitted values over visits. To take into account the more frequent data collection in the first year, the mean change in QOL through month 36 was calculated as a weighted average over the 5 follow-up visits, with equal weights per year. For selected QOL outcomes, treatment group means were displayed graphically by visit. The plotted means were the fixed effect estimates in longitudinal models, which included visit, treatment group, and visit by treatment group interaction effects.
HRs comparing the episodic and continuous antiretroviral therapy arms for opportunistic disease or death and for grade 3 or 4 symptoms were estimated and tested in unadjusted Cox proportional hazards models. The changes in QOL before the primary endpoint (opportunistic disease or death) and before serious non-AIDS events (major cardiovascular events, hepatic failure, renal failure, and non-AIDS cancers) were determined by comparing, at each visit, patients who did and did not experience such an event after the visit, using longitudinal models with a time-updated indicator for events, adjusted for baseline, and including data through month 36.
Statistical analyses were performed using SAS (version 8.2) (SAS Institute, Cary, NC). Longitudinal analyses included all patients with a baseline value and at least 1 follow-up value. Missing values were not imputed. PCS and MCS scores were calculated for all complete SF-12 questionnaires. Results were considered statistically significant for P values ≦0.05.
|
|
| |
| |
|
|
|
