August 17, 2006
While the value of highly active antiretroviral therapy (HAART) for the management of HIV infection and AIDS has been proven in many clinical trials in all parts of the world, one of the most basic strategic questions of HIV management has still not been adequately addressed: Once HAART has been started, should it be continued indefinitely or should therapy be interrupted at some point and then later resumed? Also, if continuous treatment is not the best strategy, in what manner and under what circumstances should treatment be administered?
It was to answer these questions that the study Strategies for Management of Anti-Retroviral Therapy, also known as the SMART trial, was conceived. The first set of analyses of this important CD4+ cell count-driven treatment interruption study was presented at the 13th Conference on Retroviruses and Opportunistic Infections (CROI 2006)1 and further results were presented at the XVI International AIDS Conference (AIDS 2006) in Toronto, Canada. But, before we discuss SMART, let's first look at why we should even consider stopping therapy. After all, not that long ago in the developed world, a diagnosis of AIDS carried a wretched 90% two-year mortality rate. (This is still true in most underdeveloped countries.) Thanks to HAART, HIV is now considered a manageable disease with good prospects for long-term survivability. In fact, the "diabetes paradigm" is often used nowadays when discussing HIV therapeutics. We don't readily envision stopping diabetes medications, so why does this issue even enter into the debate on HIV therapeutics?
The Impetus for Structured Treatment Interruptions
The primary motivation for an HIV treatment interruption is to limit potential medication-related toxicity. Short-term side effects such as nausea, vomiting, diarrhea, fatigue and malaise are not uncommon, particularly with some of the earlier regimens (many of which are still utilized today). Beyond short-term toxicity, it has been established that metabolic, lipodystrophic and cardiovascular complications are associated with antiretroviral therapy. These complications appear to be cumulative and progressive, although to varying degrees they are reversible after a change in therapy or a discontinuation of therapy.
Another significant issue is the potential of waning adherence or treatment fatigue, which can lead to viral rebound and the accumulation of resistance mutations that can limit future treatment options. It is growing increasingly common to see multi-drug-resistant virus isolated from HIV-infected patients as a result of HAART exposure, particularly in the setting of sub-optimal adherence.
A third reason for considering a treatment interruption strategy is the long-term financial costs of therapy. The yearly drug acquisition costs of antiretroviral therapy in the United States ranges from $10,000 to upwards of $50,000 for more intensive salvage regimens. With AIDS Drug Assistance Programs in the United States straining to cover current medications and many not covering the most expensive agents, cost has increasingly become a problem in the United States. This is true even for people with health insurance, who can have high deductibles.
The financial costs of therapy assume even greater significance in resource poor parts of the world and are often the limiting factor in the implementation of antiretroviral treatment programs. Even with the availability of generic combination tablets at a small fraction of the cost of branded antiretroviral medications, HIV medications are still not affordable in most of the poorest countries of the world. Hence, a strategy that might achieve immunologic stabilization of HIV infection at a substantially lower financial cost could potentially enable greater access to life-sustaining treatment.
An additional reason why treatment interruptions are an important area of study, and why SMART was conceived, is because of the simple fact that patients commonly embark on treatment interruptions for a wide variety of reasons on their own, and thus more data are needed to better understand the potential risks and benefits of their actions.
Who Conceived of and Carried Out the SMART Study?
The researchers for the SMART trial were from Community Programs for Clinical Research on AIDS (CPCRA) -- the National Institutes of Health-funded HIV research association that was organized to conduct community-based trials. CPCRA is separate and distinct from the AIDS Clinical Trials Group (ACTG) sites, which are predominantly based in academic institutions. When the SMART trial was designed and promoted, there were some conservative voices in the HIV research establishment opposed to the notion of a treatment interruption trial of this nature. It was not unusual to hear statements to the effect that the SMART study was "anything but smart." This seems extremely unfair since the SMART study represented a significant departure from the "business as usual" thinking in the somewhat entrenched and slow-moving HIV research establishment.
Clearly, important strategic treatment studies that involve limiting therapy will only be performed in the context of government-supported research. Industry is keen to fund studies to determine if perhaps more intensive drug therapy might be better, but quickly loses interest when scientific questions of lower dosages or intermittent therapy are raised. Certainly their shareholders do not blame them for such reticence.
What Was the Design and Main Goals of the SMART Study?
SMART is a prototypical "large simple trial" with very straightforward inclusion and exclusion criteria. To be included, a participant had to have HIV and a CD4+ cell count above 350 cells/mm3. Participants could be treatment naive or treatment experienced, and on or off HAART at study entry. The SMART study design involved the random assignment of participants into two very different treatment groups.
The first group, the drug conservation (DC) group, was treated based on a strategy of CD4+ cell count-driven structured treatment interruptions in which antiretroviral therapy was started (or restarted) below a relatively low 250 CD4+ cell count threshold. HAART was then stopped when the CD4+ cell count rose above a 350 higher limit threshold. The next cycle of HAART would again commence when the CD4+ cell count dropped below 250 cells/mm3. Two consecutive CD4+ cell counts beyond the 250 CD4+ cell lower limit or the 350 CD4+ cell upper limit were required to trigger starting or stopping medications, respectively. The second confirmatory test was intended to limit the potential for overly rapid changes in therapy.
Participants in the second group, the viral suppression (VS) group, were simply started or continued on HAART with the goal of suppressing viral levels to below the limit of detection. It should be noted that as a consequence of this strategy some of the participants in this group initiated HAART at CD4+ cell counts considerably higher than the usual 350 CD4+ cell count mark.
The ultimate goal of SMART was to determine whether a CD4+ cell count-driven intermittent treatment strategy would provide sufficient immunologic control to prevent opportunistic complications of HIV infection while at the same time limit exposure to HAART with its attendant side effects. Hence, the primary study endpoint was the risk of HIV progression events and death between the two groups. The trial was also targeted to assess the risk of the emergence of viral resistance with treatment interruptions versus continuous HAART.
During the first year of the trial, participants were followed closely; office visits and CD4+ cell count/HIV-RNA testing were scheduled every two months. But, from month 12 onward, the frequency of the visits and blood testing was reduced to every four months. The relatively low frequency of follow-up evaluation and testing after month 12 meant that it was possible for trial participants to go as long as eight months with a CD4+ cell count below 250 cells/mm3 before resuming therapy. This aspect of the trial design may have influenced the results, since the structured treatment interruption group may have had a heightened risk of opportunistic infections during these time periods while awaiting the confirmatory CD4+ cell count that would have prompted resumption of HAART.
Patient Selection and Study Enrollment
Although the trial began as a U.S. effort, it was widely expanded. Before study enrollment was prematurely halted, a total of 318 study sites in 33 different countries on five continents were included. The majority of study participants (57%) were from the United States, followed by Europe with 26% of the total enrollees.
As mentioned previously, one of the challenges with enrollment into SMART had to do with the pre-conceived notions and biases of some investigators who thought the study design was questionable in the first place. Clinical research is often performed in group practice settings with multiple treating clinicians on the investigative team. Interestingly, some clinicians were not willing to mention the study to their patients, because they believed that treatment interruption in general was simply too dangerous. Other clinicians thought the continuous treatment arm, to which participants who were randomized would be required to start HAART even if they had a high CD4+ cell count, was too risky an approach. So, to some extent, opinions were aligned on both ends of the treatment spectrum.
There was also a similar divide amongst HIV-infected patients: Some were adamantly opposed to the notion of discontinuing therapy in spite of high CD4+ cell counts while others wanted to stop therapy or hold off on medications until absolutely necessary. Some of the inquiries regarding the SMART study were from individuals solely interested in a supervised treatment interruption and the chance that they might be randomized to continue therapy led to their choice not to participate in the trial. Participants best suited for the SMART trial were those who were not committed to one strategy or the other, but who were at ease with the possibility of being randomly assigned to either approach.
One of the most helpful elements of the study, for those in the continuous treatment arm, was the freedom of the investigators to construct the best regimen possible, using even investigational agents, in order to maintain viral suppression.
Overall, 5,472 people joined SMART before new participant enrollment was suspended on Jan. 10, 2006 due to the recommendations of the Data Safety Monitoring Board (DSMB). At study entry, the average age of the participants was 46 years. Women made up 27% of the participants and blacks accounted for 30%. There were no significant differences between the baseline measures in the two groups, as shown in the table below.
(Slides by Wafaa El-Sadr, reprinted with permission.)
|Median Baseline CD4+ (IQR)||596||599||598 (range: 466 to 792)|
|Median Nadir CD4+ (IQR)||250||252||251 (range: 154 to 360)|
|HIV RNA <400 copies/mL (%)||71.0||70.8||70.9|
|Prior Clinical AIDS (%)||24.7||23.4||24.1|
|ART Naive (%)||4.5||4.8||4.7|
|Years of Prior ART (IQR)||6||6||6 (range: 3 to 8)|
DC = drug conservation arm, simply another way of describing the group of participants randomized to treatment interruption and cycling of therapy|
VS = an abbreviation for viral suppression and represents the continuous therapy control arm of the SMART study
IQR = interquartile range
ART = antiretroviral therapy
One of the first questions to be answered was whether the study design actually resulted in significant differences in HAART exposure between the two groups. As the graph below illustrates, the continuous treatment (known alternatively as the viral suppression or VS) group spent 93% of the study time on antiretroviral therapy while the treatment interruption (known alternatively as the drug conservation or DC) group was on therapy only 33% of the time.
The primary endpoint of this study was the occurrence of HIV clinical disease progression or death. The definition of HIV clinical disease progression comes directly from the U.S. Centers for Disease Control and Prevention (CDC)'s 1993 Revised Case Definition for AIDS. An HIV progression event is the occurrence of one of the defined opportunistic complications of AIDS. So, for example, if a participant developed Candida esophagitis, this would qualify as an HIV progression event.
A secondary endpoint was the development of a serious AIDS event. Examples of this included: CNS toxoplasmosis, progressive multifocal leukoencephalopathy and AIDS dementia complex. But serious or not, any of these CDC-defined AIDS-related conditions would signify as an HIV progression event and would be counted as a clinical endpoint.
The slide below shows the primary finding of the SMART trial: a 2.5-fold increased risk of a progression event or death in participants in the treatment interruption (DC) group as compared with the continuous treatment (VS) cohort. These differences were highly statistically significant and were the basis for halting the trial.
Death is the hardest of the "hard endpoints." The slide below combines the relative risk of death and HIV progression events in the treatment interruption group compared with the continuous treatment arm. The actual breakdown of the causes of death was not presented, but when this information becomes available it will help improve our understanding of HIV-associated mortality in this trial.
The slide below depicts the time course of the endpoints and shows that most of the events, as well as differences in event rates, occurred within the first two years of study enrollment.
The greater risk of opportunistic complications and disease progression was not totally unexpected since participants on the treatment interruption arm were off antiretroviral therapy 67% of the time while in the trial. However, the higher overall death rate in the treatment interruption group was disappointing, and not expected. The hope and belief was that a lower risk of HAART-related metabolic and organ-specific complications would mitigate some of the immunologic risk in the treatment interruption group and perhaps shift the cost-benefit equation in favor of the less intensive treatment scheme. But one of the most surprising findings of this trial was that the risk of cardiovascular, renal and liver complications was higher in the treatment interruption group, as shown in the slide below.
The explanation for the greater organ-specific risk in the treatment interruption group is not known, but further analysis of the specific events may lead to a better understanding.
The following slides explore the impact of baseline CD4+ cell count and nadir CD4+ cell count on the primary outcomes.
No major differences in outcomes were seen on the basis of baseline CD4+ cell count. These findings are somewhat surprising since, intuitively, one would expect there to be less risk from a treatment interruption in participants with a higher baseline CD4+ cell count. These individuals with a higher CD4+ cell count have more of a cushion between their baseline count and the CD4+ cell count in the range that is commonly associated with disease progression or death.
Another surprising finding was the apparent indifference to outcome associated with a patient's nadir CD4+ cell count. Previous structured treatment interruption studies have shown that the nadir CD4+ cell count was an important predictor of the length of time that HAART could be withheld before a patient's CD4+ cell count dropped below a predetermined threshold. Hence, the presumption was that trial participants with a low nadir CD4+ cell count would not fare so well on a treatment interruption as patients with a higher nadir CD4+ cell count.
In yet another analysis, there was a comparison between participants who entered the study with an HIV-RNA level above 400 copies/mL versus below this number. There were no differences in risk of events between those randomized to interrupted or continuous HAART in the group with detectable viral loads over 400 copies/mL. Therefore, the benefit associated with continuous treatment was concentrated in participants who had well-controlled viral replication.
As expected, the participants in the treatment interruption group were much more likely to have a CD4+ cell count in the range associated with disease progression, as shown in this graph below.
What about the influence of proximate CD4+ cell count and viral load on the outcomes in SMART? The data in the table below were presented at AIDS 2006 and were quite instructive. The risk of disease progression and death was closely linked with the proximate CD4+ cell count and viral load. Since participants randomized to the structured treatment interruption arm were more likely to drop to lower CD4+ strata, this translated into a higher rate of disease progression events. In fact, after controlling for the effect of proximate CD4+ cell count and viral load, the relative risk of progression events and death associated with the structured treatment interruption arm of SMART was reduced from 2.5 (P < .001) to 1.4 (P = .12).
|Rates (/100 PY) According to Proximal CD4 and HIV-RNA Levels (% of Total Follow-Up Time in Arm)|
|Proximal CD4 Count (cells/µL)||Proximal HIV-RNA (log10 copies/mL)|
|DC||12.4 (8%)||4.1 (23%)||2.6 (34%)||2.5 (35%)||4.3 (21%)||2.6 (16%)||2.6 (35%)||5.2 (28%)|
|VS||13.7 (2%)||3.9 (6%)||1.1 (22%)||1.1 (70%)||0.8 (68%)||1.8 (17%)||3.0 (10%)||5.7 (5%)|
As indicated in the graph below, the risk of opportunistic disease or death is progressively greater at lower CD4+ cell counts. This is true for both groups of patients in SMART. However, even with CD4+ cell counts above 350 cells/mm3, there was increased risk associated with treatment discontinuation. At higher CD4+ strata, there are statistically significant differences observed favoring the viral suppression group, but it should be noted that the absolute risk is low in both groups.
Another analysis of the SMART data at AIDS 2006 by Cohn et al2 examined the rate of serious progression of disease events in the study. Examples of serious progression events include mycobacterium avium complex, toxoplasmosis, lymphoma, visceral Kaposi's sarcoma and AIDS dementia. In contrast, examples of non-serious events would include esophageal candidiasis and chronic herpes simplex infection. The incidence of both serious and non-serious progression events was greater in the structured treatment interruption group compared to the continuous treatment group. Of note, however, was the finding that in patients with a proximate CD4+ cell count above 350 cells/mm3 the proportion of serious progression events was similar in the structured treatment interruption arm (13%) and the continuous treatment arm (15%).
The most important conclusion from the SMART study was that an HIV treatment strategy that employed cycled treatment interruptions at CD4+ cell counts above 350 cells/mm3 and then resumed therapy below a 250 CD4+ cell count threshold was inferior to continuous treatment. In addition, and more critically, treatment interruption was associated with a greater risk of HIV disease progression events; serious hepatic, renal and cardiovascular events; and overall death.
The immediate clinical implication for physicians and patients is that anyone currently pursuing a "SMART-type treatment interruption strategy" should be aware of the potential risks of such an approach.
Another possible immediate clinical implication of the SMART results has to do with the timing of treatment initiation in treatment-naive patients. Some clinicians may look at the results of this study and now consider a 250 CD4+ cell count as the lowest limit at which observation off therapy is acceptable. However, these data cannot be easily extrapolated to the treatment-naive population. It helps to keep in mind that only 4.7% of the participants in SMART were treatment naive and it is likely that almost all had viral loads greater than 400 copies/mL. In this subgroup no differences were seen between the continuous and interrupted treatment groups.
So does the failure of SMART to show benefit with cycled treatment interruption represent the end of the line for CD4+ cell count-driven structured treatment interruption research? For several reasons, this is unlikely.
The first has to do specifically with the original design of SMART, which, in retrospect, utilized a CD4+ cell count starting threshold that was too low. This is not meant as a criticism, since it is not possible to predict the optimal parameters of a new therapeutic strategy and SMART represented an innovative trial that was intended to explore new terrain. Nonetheless, the latest data presented at AIDS 2006 clearly showed that the proximate CD4+ cell count and viral load accounted for most of the outcome differences between the groups. Since the major explanation for the poorer outcomes in the structured treatment interruption group could be attributed to the under 350 CD4+ cell count strata, this begs the question as to whether a higher CD4+ cell count starting threshold might have led to very different results.
Previous trials of CD4+ cell count-driven therapy -- notably the BASTA trial from Italy -- showed that therapy could safely be stopped in participants with a CD4+ cell count greater than 800 cells/mm3.3 By design, in the BASTA trial, therapy was restarted when the CD4+ cell count fell below 400 cells/mm3. Following treatment interruptions, HIV-RNA levels rebounded and CD4+ cell counts declined at variable rates. One of the most important findings from that study was that the higher the nadir CD4+ cell count prior to HAART initiation, the longer it took before the participant reached the 400 cells/mm3 CD4+ cell count threshold for resumption of antiretroviral therapy.
At AIDS 2006, the BASTA trial was updated with long-term follow-up data.4 There were 114 participants randomized in a 2:1 ratio to structured treatment interruptions (76 patients) or continuous treatment control (38 patients) and the 56-month results were presented. In the structured treatment interruption arm there were 108 interruptions of therapy (between one and five per patient). This strategy translated into HAART being withheld 70.4% of the time in the structured treatment interruption group. Not surprisingly, a decrement was observed in the CD4+ cell counts in the structured treatment interruption group compared with the continuous treatment group. But CD4+ cell counts rebounded in response to restarting treatment. Viral resistance developed in 4/76 (5.2%) of the patients in the structured treatment interruption group and 1/38 (2.6%) in the control group (P = ns). Clinical progression or death was observed in three patients in the control group and in none in the structured treatment interruption group. So this relatively small CD4+ cell count-guided strategy trial continued to show positive results in the structured treatment interruption group. These investigators have recently started enrollment in a much larger randomized trial in order to test this strategy with a study powered to detect smaller differences.
Other studies simply looking at stopping therapy have also suggested that HIV medications could be safely withdrawn in patients who start therapy with higher CD4+ cell counts than are currently recommended (i.e., higher than 350 cells/mm3). The ACTG 5170 study that was presented at CROI 20065 demonstrated that HAART could be safely stopped in a cohort of patients with high CD4+ cell counts and who had originally gone on treatment with CD4+ cell counts above current guidelines.
In fact, the Staccato study, which was also presented at AIDS 2006 and published in the Lancet,6 showed no major differences between treatment interruption and continuous therapy groups when a 350 CD4+ cell count start and stop point were utilized. The caveat is that all of these studies had much smaller numbers of participants and were not powered to show differences in disease progression and death. But the point remains that a higher CD4+ cell count trigger for starting HAART may be a crucial determinant of success.
Other reasons that structured treatment interruption research will continue to be pursued have to do with the practical realities of clinical practice. Patients continue to take treatment interruptions without either asking or telling their providers and a better understanding of the risks and benefits would be useful. Similarly, some clinicians recommend treatment interruptions under different circumstances. Finally, the multiple compelling reasons that led to the undertaking of the SMART study still exist and will stimulate future research into innovative treatment approaches.
The Staccato Study
The results of the following study, known as the "Staccato" trial, contradict the results of the SMART study, but the key may be the different treatment initiation thresholds. Staccato is an important structured treatment interruption trial because of its innovative design and ambitious scope. It was carried out by a collaboration of investigators from Switzerland, Australia and Thailand, although the majority of participants were enrolled in Thailand. Support for Staccato was provided by the Swiss National Science Foundation and several other private foundations. Most of the antiretroviral medications were donated by Gilead Sciences, Inc., Roche and Abbott Laboratories.
Four hundred thirty patients were randomized 2:1 to scheduled treatment interruptions (n = 284) or continued therapy (n = 146). The median time on randomized treatment was 21.9 months (range 16.4 to 25.3). Most patients were enrolled in Thailand and they were treated with stavudine (d4T, Zerit), didanosine (ddI, Videx) and ritonavir (RTV, Norvir)-boosted saquinavir (SQV, Fortovase, Invirase) from 2001 until 2003 when the nucleosides were switched to tenofovir (TDF, Viread) and lamivudine (3TC, Epivir).
Patients were randomized into one of three quite different treatment strategies. The first group received continuous antiretroviral therapy and served as the standard treatment control. The second group received HAART on a week on, week off (WOWO) schedule, as a test of short cycle structured treatment interruptions. The third group received CD4+ cell count-guided therapy, with antiretroviral therapy started when the CD4+ cell count dipped below 350 cells/mm3 and stopped when the CD4+ cell count rose above this same 350 CD4+ cell count threshold. Prior to study completion, patients on the structured treatment interruption arm had antiretroviral therapy restarted and then had repeat testing performed 12 and 24 weeks later to see if viral re-suppression -- the primary study endpoint -- could be achieved.
The short-cycle structured treatment interruption, week on, week off arm of the study was halted prematurely because of the occurrence of unacceptably high rates of treatment failure and development of viral resistance. These results were previously published.7
The remainder of the trial was continued in order to determine if the other investigational arm -- CD4+ cell count-guided structured treatment interruption strategy -- would perform as well as continuous HAART. The expectation was that CD4+ cell count-guided therapy would limit exposure to antiretroviral therapy, decrease the rate of medication-related adverse events, reduce the overall financial cost of therapy and yet be sufficient to control HIV replication.
One of the co-primary endpoints of the study was the proportion of patients with plasma HIV RNA below 50 copies/mL at the end of the trial, 24 weeks following the scheduled restarting of HAART in the structured treatment interruption group. The other co-primary endpoint was the difference in exposure to antiretroviral medication between the two groups. Secondary endpoints of the study included rate of opportunistic infections, changes in CD4+ cell counts and adverse events.
To be included in the study, candidates had to have a CD4+ cell count greater than 350 cells/mm3 and plasma HIV RNA below 50 copies/mL. They also had to be on a stable HAART regimen for at least three months and have no evidence of pre-existing drug resistance.
Staccato Study Results
One co-primary endpoint of the trial -- achieving a viral load of below 50 copies/mL 24 weeks following structured treatment interruption completion -- was observed in 257 of 284 (90.5%) patients in the scheduled treatment interruption group compared with 134 of 146 (91.8%) patients in the continued treatment group (difference 1.3%, 95% CI?4.3 to 6.9, P = .90). The other co-primary endpoint -- time on HAART during the trial -- was 37.5% for the structured treatment interruption group compared with 99% for the control group.
In reviewing the secondary endpoints, there were no AIDS-defining events in the study, but a higher rate of diarrhea and neuropathy was seen in the continuous treatment control group. On the other hand, oral and vaginal candidiasis were more frequent with scheduled treatment interruption. Also, CD4+ cell counts were lower in the structured treatment interruption group -- 484 versus 655 in the control group. Overall, 10 patients (2.3%) had evidence of resistance mutations, with no significant differences between groups. Drug cost savings in the scheduled treatment interruption group amounted to 61.5%.
Although this study was not powered to detect changes in clinical events, if the 3% rate of opportunistic infections and deaths observed in SMART had occurred in the structured treatment interruption arm of Staccato, this would have translated into 17 events or deaths. In fact, only one death (due to colon cancer) was seen in the structured treatment interruption group. The most plausible reason for the contrast in outcomes is the different starting threshold in Staccato versus SMART.
The rate of the development of resistance mutations is an important consideration in any structured treatment interruption trial, and interestingly, there was a low rate of resistance observed in the Staccato study. The most common antiretroviral regimen in this study was ritonavir-boosted saquinavir. Like other boosted protease inhibitors, this combination has a relatively high barrier to the development of drug resistance and may explain the favorable resistance findings.
The short-cycle structured treatment interruption portion of the Staccato study demonstrated unacceptable clinical and virologic failure rates with a week on, week off strategy. However, a pilot study that is being conducted by Cal Cohen et al of a five day on treatment, two day off treatment (FOTO)8 that was presented at CROI showed promising results. Larger studies of this strategy are planned for the future.
The lower CD4+ cell counts in the structured treatment interruption group compared with the control are not unexpected since these patients had roughly one third the exposure to HAART during the almost two years of study follow-up.
The Staccato study shows that stopping HAART when CD4+ cell counts were above 350 cells/mm3 and restarting below this same level appeared to be safe and resulted in a very significant reduction in exposure to antiretroviral therapy. One caveat is that in the scheduled treatment interruption group there was a greater chance of low CD4+ cell counts and minor manifestations of HIV infection. The results of this trial contradict the SMART study, but the different HAART starting threshold -- 350 in Staccato versus 250 in SMART -- may be the crucial factor that led to the disparate results. Undoubtedly, further study of this CD4+ cell count-guided strategy utilizing a higher HAART starting mark will be needed to better understand the role, if any, of structured treatment interruption in HIV management.
|Please note: Knowledge about HIV changes rapidly. Note the date of this summary's publication, and before treating patients or employing any therapies described in these materials, verify all information independently. If you are a patient, please consult a doctor or other medical professional before acting on any of the information presented in this summary. For a complete listing of our most recent conference coverage, click here.|