| |
CD4+ Count-Guided Interruption of Antiretroviral Treatment, The Strategies for Management of Antiretroviral Therapy (SMART) Study Group
|
| |
| |
NEJM, November 30, 2006
ABSTRACT
Background Despite declines in morbidity and mortality with the use of combination antiretroviral therapy, its effectiveness is limited by adverse events, problems with adherence, and resistance of the human immunodeficiency virus (HIV).
Methods We randomly assigned persons infected with HIV who had a CD4+ cell count of more than 350 per cubic millimeter to the continuous use of antiretroviral therapy (the viral suppression group) or the episodic use of antiretroviral therapy (the drug conservation group). Episodic use involved the deferral of therapy until the CD4+ count decreased to less than 250 per cubic millimeter and then the use of therapy until the CD4+ count increased to more than 350 per cubic millimeter. The primary end point was the development of an opportunistic disease or death from any cause. An important secondary end point was major cardiovascular, renal, or hepatic disease.
Results
A total of 5472 participants (2720 assigned to drug conservation and 2752 to viral suppression) were followed for an average of 16 months before the protocol was modified for the drug conservation group. At baseline, the median and nadir CD4+ counts were 597 per cubic millimeter and 250 per cubic millimeter, respectively, and 71.7% of participants had plasma HIV RNA levels of 400 copies or less per milliliter. Opportunistic disease or death from any cause occurred in 120 participants (3.3 events per 100 person-years) in the drug conservation group and 47 participants (1.3 per 100 person-years) in the viral suppression group (hazard ratio for the drug conservation group vs. the viral suppression group, 2.6; 95% confidence interval [CI], 1.9 to 3.7; P<0.001). Hazard ratios for death from any cause and for major cardiovascular, renal, and hepatic disease were 1.8 (95% CI, 1.2 to 2.9; P=0.007) and 1.7 (95% CI, 1.1 to 2.5; P=0.009), respectively. Adjustment for the latest CD4+ count and HIV RNA level (as time-updated covariates) reduced the hazard ratio for the primary end point from 2.6 to 1.5 (95% CI, 1.0 to 2.1).
Conclusions Episodic antiretroviral therapy guided by the CD4+ count, as used in our study, significantly increased the risk of opportunistic disease or death from any cause, as compared with continuous antiretroviral therapy, largely as a consequence of lowering the CD4+ cell count and increasing the viral load. Episodic antiretroviral therapy does not reduce the risk of adverse events that have been associated with antiretroviral therapy.
INTRODUCTION
With the advent of potent combination antiretroviral therapy came the hope that such therapy might lead to the eradication of infection with the human immunodeficiency virus (HIV).1 It was soon recognized, however, that this goal was unlikely to be achieved owing to the existence of latent reservoirs; people infected with HIV would need to receive antiretroviral therapy for many years, if not for life.2,3 Potent antiretroviral therapy is associated with substantial benefits with regard to morbidity and mortality.4,5,6 However, the therapy is also associated with both short-term and long-term adverse events.7,8 Major metabolic and cardiovascular complications have been a particular concern.9,10 In addition, HIV can become resistant to available antiretroviral therapy, particularly if adherence is poor, which can lead to cross-resistance within a class of drugs and, ultimately, to multidrug resistance.11 Prolonged use of antiretroviral therapy is also expensive.12
The inherent risks and problems associated with lifelong antiretroviral therapy have led to the study of treatment-sparing strategies that might provide the benefits of antiretroviral therapy while minimizing the risk of adverse events and other risks associated with long-term use. We conducted the Strategies for Management of Antiretroviral Therapy (SMART) trial in order to compare a treatment strategy of episodic use of antiretroviral therapy according to the CD4+ count with the current practice of continuous antiretroviral therapy.
Methods
The SMART study was initiated by the Terry Beirn Community Programs for Clinical Research on the Acquired Immunodeficiency Syndrome (AIDS) and implemented in collaboration with regional coordinating centers in Copenhagen (the Copenhagen HIV Programme), London (the Clinical Trials Unit of the Medical Research Council), and Sydney (the National Centre in HIV Epidemiology and Clinical Research). Some members of the SMART writing group developed the study protocol with the sponsor, the Division of AIDS of the National Institute of Allergy and Infectious Diseases (NIAID). The writing group takes full responsibility for the completeness and veracity of the data, data analyses, and this article. Drugs used in the study were purchased by patients either directly or through insurance, Social Security, or public access programs.
Participants
Persons infected with HIV who were older than 13 years and were not pregnant or breast-feeding were eligible for the study if their CD4+ count exceeded 350 cells per cubic millimeter and they were willing to initiate, modify, or stop antiretroviral therapy according to study guidelines. Participants were eligible whether or not they had received or were currently receiving antiretroviral therapy. The study was approved by the institutional review board at each site, and written informed consent was obtained from all participants.
Study Design
The SMART study is a randomized trial comparing two antiretroviral treatment strategies. Investigators and participants were aware of the treatment assignments. The viral suppression strategy, which was the control strategy, was defined to be consistent with the 2003 guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents13: available antiretroviral regimens were to be used in an uninterrupted manner with the goal of maximal and continuous suppression of HIV replication. The experimental drug conservation strategy entailed the episodic use of antiretroviral therapy according to CD4+ count thresholds: the use of antiretroviral therapy was deferred until the CD4+ count decreased to less than 250 cells per cubic millimeter, at which time antiretroviral therapy was to be initiated (or reinitiated) and continued until the CD4+ count increased to more than 350 cells per cubic millimeter. The protocol also permitted antiretroviral therapy to be initiated (or reinitiated) if symptoms of disease from HIV infection (e.g., oral thrush) developed or the percentage of CD4+ lymphocytes (CD4+ percentage) was less than 15%. On confirmation that the CD4+ count was more than 350 cells per cubic millimeter, antiretroviral therapy was to be stopped and then resumed when the CD4+ count was less than 250 cells per cubic millimeter. During periods of antiretroviral therapy, the goal was to achieve maximal viral suppression. The CD4+ count thresholds for stopping and starting antiretroviral therapy were chosen on the basis of reported associations between CD4+ counts and the risks of opportunistic diseases and death.13,14,15,16
The primary end point was new or recurrent opportunistic disease or death from any cause. Qualifying clinical events included those in the revised case definition for AIDS of the Centers for Disease Control and Prevention,17 as well as additional conditions related to immunodeficiency (see the Supplementary Appendix, available with the full text of this article at www.nejm.org). Important secondary end points included death from any cause; serious opportunistic disease18; major cardiovascular, renal, or hepatic disease (see the Supplementary Appendix); and grade 4 adverse events (not including opportunistic disease) or death from any cause. Grade 4 adverse events were defined as potentially life-threatening symptomatic events requiring medical intervention, according to the toxicity table of the Division of AIDS of the NIAID. Data on lower-grade toxic effects were not collected.
Using preestablished criteria, an end-point review committee that was unaware of the treatment assignments reviewed the events classified as opportunistic disease, death from any cause, or major cardiovascular, hepatic, or renal disease. The end-point review committee classified the underlying cause of death using the Coding of Death in HIV (CoDe) project system.19 Opportunistic disease and major cardiovascular, renal, or hepatic diseases classified as confirmed or probable by the end-point review committee were considered end points, as were all deaths, irrespective of the cause.
We calculated that 6000 patients would need to be enrolled for the study to have a statistical power of 80% to detect a 17% relative reduction in the rate of opportunistic disease or death from any cause in the drug conservation group as compared with the viral suppression group, with a two-sided alpha level of 0.05. Follow-up was to continue until 910 primary end points had occurred (estimated to be at least 6 years for each participant), assuming an event rate in the viral suppression group of 1.3% in each of the first 2 years and 2.6% per year thereafter.20 Randomization was stratified according to clinical site with the use of permuted blocks of random sizes.
Data Collection and Follow-up
Before randomization, participants' antiretroviral therapy history and medical history were obtained, as were the nadir CD4+ count; the highest recorded plasma HIV RNA level; the CD4+ count, CD4+ percentage, and HIV RNA level at baseline; and the three most recent CD4+ counts, CD4+ percentages, and HIV RNA levels before baseline. Follow-up visits were scheduled at 1 month and 2 months, every 2 months thereafter for the first year, and every 4 months in the second and subsequent years. At each visit, a history was taken and an examination conducted, and the CD4+ count and HIV RNA level were measured. More frequent assessments could be carried out if clinical care was required. At baseline and at each annual visit, a 12-lead electrocardiogram was obtained; the data were electronically transmitted to a reading center for detection of any changes indicative of a silent myocardial infarction.21,22,23
Interim Monitoring of Safety and Efficacy
An independent data and safety monitoring board reviewed interim analyses from the SMART study at least annually. According to protocol guidelines, the board was to consider early termination of the study or modification of the protocol if findings concerning the primary end point (opportunistic disease or death from any cause) and the secondary end point of major cardiovascular, renal, or hepatic disease were consistent - both favoring the same treatment group - and there was clear and substantial evidence of benefit or harm. An O'Brien-Fleming boundary and the Lan-DeMets spending function were used to control the type I error with regard to the primary end point.24,25
On January 10, 2006, at its sixth meeting, the board recommended stopping enrollment in the SMART trial because of a safety risk in the drug conservation group and because it appeared to be very unlikely that superiority of the drug conservation treatment would be shown. On January 11, 2006, investigators and participants were notified of these findings, enrollment was stopped, and participants in the drug conservation group were advised to restart antiretroviral therapy. All participants continued in follow-up. This article describes findings through the closure of enrollment.
Statistical Analysis
The drug conservation and viral suppression groups were compared according to the intention-to-treat principle. Time-to-event methods (Kaplan-Meier survival curves and Cox proportional-hazards models) were used to compare the drug conservation group and the viral suppression group with respect to event rates for opportunistic disease or death from any cause; death from any cause; serious and nonserious opportunistic disease; major cardiovascular, renal, or hepatic disease; and grade 4 events.26 Follow-up data were censored either when participants were lost to follow-up before January 11, 2006, or on that date.
The hazard ratios for the comparison of the drug conservation group with the viral suppression group were estimated from Cox models with a single binary treatment group indicator. We tested the proportional-hazards assumption by including an interaction term between the treatment indicator and the log-transformed follow-up time.26
The primary end point was summarized for selected subgroups that were predefined according to baseline characteristics. The heterogeneity of hazard ratio estimates for subgroups was assessed by including an interaction term between treatment and subgroup in expanded Cox models.
Cox proportional-hazards models were used to assess the effects of CD4+ counts and HIV RNA levels as time-dependent covariates during follow-up on the hazard ratios for opportunistic disease or death from any cause. Statistical analyses were performed with the use of SAS software, version 8.2. All reported P values are two-sided and have not been adjusted for multiple examinations of the data.
Results
Baseline Characteristics
Between January 8, 2002, and January 11, 2006, 5472 participants were randomly assigned to treatment groups (2720 to the drug conservation group and 2752 to the viral suppression group). Participants were enrolled at 318 sites in 33 countries. Table 1 summarizes key baseline characteristics. The treatment groups were well balanced at entry.
Follow-up
Total follow-up time was approximately 3700 person-years in each group, with a mean follow-up time of 16 months. Approximately 26% of participants were followed for more than 2 years. Participants attended 96.5% of follow-up visits in the drug conservation group and 94.8% of follow-up visits in the viral suppression group. On January 11, 2006, the status with regard to the primary end point was unknown for 32 participants (1.2%) in the drug conservation group and 41 participants (1.5%) in the viral suppression group (Fig. I in the Supplementary Appendix).
Use of Antiretroviral Therapy and Changes in CD4+ and HIV RNA Levels
After randomization, the median duration of the first period of interruption of antiretroviral therapy for participants in the drug conservation group was 16.8 months (interquartile range, 5.7 to 42.3) (Fig. IIA in the Supplementary Appendix shows the percentage of participants who received antiretroviral therapy through follow-up). A total of 343 participants stopped antiretroviral therapy a second time, and 62 participants stopped three or more times. The average CD4+ count decreased by 87 cells per cubic millimeter per month during the first 2 months after randomization among participants in the drug conservation group (Fig. IIB in the Supplementary Appendix); thereafter, it continued to decline but at a lower rate. On average, throughout follow-up, the CD4+ count was 206 cells per cubic millimeter lower in the drug conservation group than in the viral suppression group. HIV RNA levels also changed rapidly in the drug conservation group after randomization: within 2 months, the percentage of participants with HIV RNA levels of 400 copies per milliliter or less decreased from 71.8% to 6.0% (Fig. IIC in the Supplementary Appendix). After reinitiation of antiretroviral therapy in the drug conservation group, the median time to an HIV RNA level of 400 copies per milliliter or less was 3.1 months (Fig. IIIA in the Supplementary Appendix); the CD4+ count increased by an average of 166 cells per cubic millimeter within 8 months (Fig. IIIB in the Supplementary Appendix).
On average, participants in the drug conservation group and the viral suppression group received antiretroviral therapy during 33.4% and 93.7% of the follow-up time, respectively. Participants in both groups had CD4+ counts of 350 cells per cubic millimeter or more during the majority of the follow-up time (67.9% of the time in the drug conservation group and 92.7% of the time in the viral suppression group) (Figure 1). Participants in the drug conservation group had CD4+ counts of less than 250 cells per cubic millimeter during 8.6% of the follow-up time, as compared with 1.8% of the time in the viral suppression group. The percentage of follow-up time during which participants had HIV RNA levels of 400 copies per milliliter or less was substantially greater in the viral suppression group than in the drug conservation group (72.3% and 28.8%, respectively).
Primary End Point and Its Components
There were 120 participants in the drug conservation group and 47 in the viral suppression group who had an opportunistic disease or died from any cause (Table 2). The cumulative probabilities of opportunistic disease or death from any cause after 12, 24, and 36 months were 0.030, 0.067, and 0.091 in the drug conservation group and 0.010, 0.021, and 0.042 in the viral suppression group, respectively (Figure 2A).
The estimated hazard ratio for opportunistic disease or death from any cause was 2.6 (95% confidence interval [CI], 1.9 to 3.7; P<0.001) (Table 2 and Figure 2A), and it did not vary significantly over the follow-up period (P=0.78 for proportional hazards). Among patients in the drug conservation group who reached the primary end point, death from any cause was the most common individual event (39.2%), followed by esophageal candidiasis (20.0%) and pneumonia from Pneumocystis jiroveci infection (6.7%). Among patients in the viral suppression group who reached the primary end point, the most common event was death from any cause (57.4%), followed by esophageal candidiasis (14.9%) (Table I in the Supplementary Appendix).
The estimated hazard ratio for serious opportunistic disease, a component of the primary end point, was 6.6 (95% CI, 1.5 to 29.1; P=0.01); the hazard ratio for nonserious events associated with opportunistic disease was 3.6 (95% CI, 2.1 to 6.1; P<0.001); and the hazard ratio for death from any cause was 1.8 (95% CI, 1.2 to 2.9; P=0.007) (Table 2 and Figure 2B).
Only 8% of deaths were due to opportunistic disease. The most common underlying causes of death were cancers other than those considered to be opportunistic disease (in 11 participants in the drug conservation group and 5 participants in the viral suppression group); cardiovascular disease (7 participants and 4 participants, respectively); substance abuse (3 participants and 5 participants, respectively); accident, violence, or suicide (3 participants and 4 participants, respectively); and infection other than that considered to be an opportunistic disease (3 participants and 1 participant, respectively). For 18 participants who died (15 in the drug conservation group and 3 in the viral suppression group), the underlying cause of death could not be determined (Table II in the Supplementary Appendix).
We examined the robustness of the primary end-point findings by considering three additional outcomes: opportunistic diseases (restricted to those that were new [nonrecurrent]) or death from any cause; fatal and nonfatal cases of opportunistic disease (excluding deaths from causes other than opportunistic disease); and all reported cases of opportunistic diseases and deaths irrespective of the classification after review by the end-point review committee. The estimated hazard ratios for these three outcomes in the drug conservation group versus the viral suppression group were 2.6 (95% CI, 1.8 to 3.7; P<0.001), 3.6 (95% CI, 2.2 to 5.9; P<0.001), and 2.5 (95% CI, 1.9 to 3.3; P<0.001), respectively.
Major Cardiovascular, Renal, and Hepatic Disease
Among the participants in the drug conservation group, 65 had at least one episode of major cardiovascular, renal, or hepatic disease, as did 39 participants in the viral suppression group (hazard ratio, 1.7; 95% CI, 1.1 to 2.5; P=0.009) (Table 2 and Figure 2C). Estimated hazard ratios for each type of disease were all greater than 1.0, favoring the viral suppression group.
Grade 4 Events
Grade 4 adverse symptomatic events occurred in 173 participants in the drug conservation group and 148 participants in the viral suppression group (hazard ratio, 1.2; 95% CI, 1.0 to 1.5; P=0.13) (Table 2 and Figure 2D, and Table III in the Supplementary Appendix). The hazard ratio for the composite outcome of a grade 4 event or death from any cause was 1.3 (95% CI, 1.0 to 1.6; P=0.03).
Primary End Point According to Subgroup
Estimated hazard ratios varied significantly according to race, baseline HIV RNA level, and baseline CD4+ cell count (Figure 3). Among participants who were receiving antiretroviral therapy at baseline, for those with an HIV RNA level of 400 copies per milliliter or less at baseline, the hazard ratio for opportunistic disease or death from any cause was 4.0, whereas those with levels of more than 400 copies per milliliter had a hazard ratio of 1.2 (P<0.001). This significant difference was due to different rates of opportunistic disease or death from any cause in the viral suppression group between the subgroups with HIV RNA levels of 400 copies per milliliter or less and those with HIV RNA levels of more than 400 copies per milliliter (0.8 and 2.6 events per 100 person-years, respectively), in contrast to similar rates for those subgroups in the drug conservation group (3.2 and 3.1, respectively).
Adjustment for Latest CD4+ Count and HIV RNA Level
Proportional-hazards models were used to assess the effects of time-dependent covariates corresponding to the latest CD4+ counts and latest HIV RNA levels, considered separately and together, on the hazard ratio (Figure 4). The hazard ratio for opportunistic disease or death from any cause was reduced from 2.6 to 1.7 (95% CI, 1.2 to 2.5) after adjustment for the latest CD4+ count; it was further reduced to 1.5 (95% CI, 1.0 to 2.1) after adjustment for both the latest HIV RNA level and the latest CD4+ count. Adjusted for both the latest HIV RNA level and the latest CD4+ count, the hazard ratios for opportunistic disease and for death from causes other than opportunistic disease were 1.7 (95% CI, 1.0 to 2.9) and 1.2 (95% CI, 0.7 to 2.2), respectively; unadjusted, these hazard ratios were 3.6 (95% CI, 2.2 to 5.9) for opportunistic disease and 1.8 (95% CI, 1.1 to 2.9) for death from causes other than opportunistic disease, respectively.
Discussion
Our data demonstrate that continuous use of antiretroviral therapy is superior to its episodic use as guided by the CD4+ count, with antiretroviral therapy deferred until the CD4+ count is less than 250 cells per cubic millimeter. The superiority of the viral suppression strategy, designed to achieve maximal and continuous suppression of HIV replication with the use of antiretroviral therapy, was evident with regard to the primary end point (opportunistic disease or death from any cause), as well as death from any cause, serious opportunistic disease, and an important secondary end point, major cardiovascular, renal, or hepatic disease.
Interruption of antiretroviral therapy has been advocated as a treatment strategy to enhance the quality of life, limit adverse events, and allow for the emergence of the predominant wild-type virus in patients infected with multidrug-resistant HIV.27,28 Two randomized studies have used higher CD4+ count thresholds than those used in our study for the initiation of antiretroviral therapy, but they involved only 69 patients29 and 74 patients30,31 and were therefore too small to allow for the reliable assessment of effects of treatment interruption on clinical outcomes.
More recently, the findings of two other larger, randomized trials were reported. In the Staccato study, in which the primary end point was virologic suppression and the amount of exposure to antiretroviral drugs, 284 patients were randomly assigned to receive antiretroviral therapy guided by the CD4+ count (with therapy deferred until the CD4+ count was less than 350 cells per cubic millimeter and then used until the CD4+ count was more than 350 cells per cubic millimeter) and 146 patients were randomly assigned to receive continuous antiretroviral therapy.32 After approximately 2 years of follow-up, diarrhea and neuropathy were more common among those receiving continuous therapy, and oral and vaginal candidiasis were more common among those receiving episodic therapy. In the Trivacan study, 216 patients were randomly assigned to episodic therapy with the same CD4+ count thresholds as those in our study, and 110 patients were randomly assigned to continuous antiretroviral therapy.33 An increased risk of bacterial infections and other complications were noted in the episodic treatment group.
When we began our study, data indicated that the risk of AIDS was low among patients who had never received antiretroviral therapy and among those who had received it but who also had CD4+ counts of more than 200 cells per cubic millimeter.14,15,16 Consequently, we chose to use a CD4+ count threshold of 250 cells per cubic millimeter for initiation (or reinitiation) of antiretroviral therapy in the drug conservation group, and CD4+ counts and symptoms were monitored closely. Data also indicated that complications and deaths among patients with higher CD4+ levels were largely due to grade 4 adverse events and deaths from causes other than opportunistic disease that were either associated with antiretroviral therapy or attributable to non-HIV causes.8 Thus, we expected the risk of death from causes other than opportunistic disease to decrease with the interruption of antiretroviral therapy, rather than to increase.
In our study, deaths represented a large percentage of the primary events (44.3%) and, as in other reports, deaths from causes other than opportunistic disease were common.8,34,35,36 The excess of deaths from causes other than opportunistic disease in the drug conservation group was surprising. Furthermore, contrary to available data and to the assumptions underlying our study design, participants in the drug conservation group had a higher rate of major cardiovascular, renal, or hepatic disease than did those in the viral suppression group. On the basis of prior findings, and as a consequence of less exposure to antiretroviral therapy in the drug conservation group,10,37 we expected the rate of cardiovascular disease to be 15% lower in the drug conservation group than in the viral suppression group.
There were fewer occurrences of major hepatic or renal disease than of cardiovascular disease in our study, but hepatic or renal disease was still more frequent in the drug conservation group than in the viral suppression group. Some antiretroviral drugs have been associated with adverse hepatic and renal events,7,38,39 but recent findings indicate that these events are also related to the level of immunodeficiency40,41 and that antiretroviral therapy improves these outcomes either directly by inhibiting viral replication or indirectly by improving immune function.39 Our data support these findings. Adjustment for the latest CD4+ count reduced the hazard ratio in the drug conservation group versus the viral suppression group for death from causes other than opportunistic disease, confirming that immunosuppression increases the risk of death among patients with diseases not traditionally believed to be opportunistic in nature. Much, but not all, of the difference in the rates of opportunistic disease or death from any cause between the drug conservation group and the viral suppression group was explained by differences in the CD4+ count and HIV RNA level during follow-up. The hazard ratio for opportunistic disease or death from any cause in the drug conservation group versus the viral suppression group was reduced from 2.6 to 1.5 after adjustment for the latest CD4+ count and the latest HIV RNA level. The reasons for the remaining excess risk are not clear.
Although our findings indicate that the interruption of antiretroviral therapy with the use of higher CD4+ count thresholds than those used in our study may result in lower risks of opportunistic disease or death from any cause, the lack of benefit of our interruption strategy on major adverse events associated with antiretroviral therapy suggests that such strategies should be viewed as carrying a net clinical risk unless proven otherwise in appropriately powered studies.
In summary, our findings provide clear and compelling evidence that the episodic antiretroviral strategy, guided by the CD4+ count, used in the SMART study is deleterious. Our results indicate that some of the excess risk of opportunistic disease or death from any cause in the drug conservation group appears to be attributable to the longer period during which participants had reduced CD4+ counts. Further research is needed to evaluate the effect of interrupting antiretroviral therapy on immune function, inflammation, and other markers.
|
|
| |
| |
|
|
|
