February 11, 2004
Strategic or structured treatment interruptions (STIs) are one of the most controversial and interesting topics in HIV medicine. And, whether we like it or not, patients will continue to ask for them even if the data regarding them is unclear. With the life span of someone living with HIV expected to be measured in decades, it is clear that even with once-a-day medications, patients can experience burnout. However, patient burnout is only one of a multitude of reasons for an STI. In order to properly discuss this subject, it helps to differentiate between the general types of STIs. Once the kind of STI is clearly described, the data that support or refute the stratagem may be considered. For the sake of this CME, I have separated STI into the following categories, which I'll discuss at length later:
Many clinicians would not consider this to be an STI. Instead, it would be considered a temporary treatment interruption undertaken with the intention to determine if a patient's symptoms are related to HIV medications. Some medication-related side effects have an insidious onset and may be quite nonspecific. The only way to determine if troublesome symptoms are related to an antiretroviral cocktail is to interrupt therapy and see if the symptoms recede or resolve. If a question remains about the relationship between symptoms and medication, some clinicians conduct a re-challenge with the same regimen to see if the symptoms recur (however, an abacavir [ABC, Ziagen] re-challenge is strictly proscribed because of the possibility of hypersensitivity).
The prompt resolution of a patient's symptoms during this kind of treatment interruption will typically lead to the selection of an alternate medication for the most likely offending agent. Often it's quite clear which drug is causing all the problems based on the known side effect profiles of certain agents (i.e., nausea from zidovudine [ZDV, Retrovir] or diarrhea from nelfinavir [NFV, Viracept]). On the other hand, if a patient's symptoms remain unchanged off therapy, then another explanation for these symptoms can be pursued and the identical regimen resumed.
However, it may not be necessary to interrupt treatment in cases of probable medication toxicity. When one agent is highly likely to be causing symptoms (i.e., neuropathy from stavudine [d4T, Zerit]), a single agent switch with continuation of the other medications in the combination, may be the most appropriate course of action.
One of the more common HIV management errors is to continue antiretroviral therapy in the face of medication-related side effects affecting a patient's quality of life. Given the expansion of treatment options in recent years, there are few patients who must stay on a prescribed treatment course that is causing them distress.
An early pilot study, and one of the most successful in this category, was published in Nature in 2000 by Dr. Eric Rosenberg from Massachusetts General Hospital. It suggested the potential usefulness of an STI with patients experiencing acute HIV infection.1 In this study, there was evidence of increased virus-specific cytotoxic T lymphocytes and maintained T-helper-cell responses in patients who had one or two STIs after they were treated for acute infection. In this cohort, there also was a suggestion that patients were more likely to maintain lower viral load levels (below 5,000 copies/mL) off therapy and manifested a slower viral rebound after the second STI.
A study with the acronym PRIMSTOP2 was designed to examine the role of STI in primary HIV infection utilizing hydroxyurea as part of the HAART regimen. In this trial, the combination of nelfinavir, stavudine, didanosine (ddI, Videx) and hydroxyurea was given to 29 patients with primary HIV infection. After 32 weeks of continuous treatment, a series of three STIs of two, four and eight weeks (all separated by 12 weeks) was performed. HAART was stopped 12 weeks after the last STI and patients were followed for an additional 24 weeks. Overall, three patients were lost to follow up, three developed genotypic resistance to nelfinavir, and 52% had hydroxyurea discontinued due to side effects. At the 24-week endpoint, 7/26 patients (27%) had a viral load less than 1,000 copies/mL and one patient had a viral load less than 50 copies/mL. This study suggests that there may be limited benefit in early treatment of primary HIV infection with STI strategy.
Other studies of STI in acute HIV infection have not demonstrated consistent immunologic benefit.3,4 These studies utilized different treatment schemes. In one cohort3 treatment often was begun after seroconversion and in another4 a therapeutic vaccine was administered in conjunction with one treatment interruption. Significant immune control was not observed in these trials.
Recent reports have shown that even when favorable short-term outcomes were observed, virologic control eroded with longer-term follow up. In February 2004, at CROI, Bruce Walker presented data from his original Massachusetts General cohort with greater than five-year follow up. There was evidence of immune escape and rising viral loads despite previously well-controlled viremia and only one out of 14 of the patients in the study maintained a viral load below 5,000 copies/mL.5 Unfortunately, no randomized and controlled trials of STI in primary HIV infection are yet available.
The great majority of HIV-infected individuals are diagnosed beyond the window of acute infection, so studies of "autoimmunization" in chronic infection have more pragmatic implications. The landmark study that provided a number of important distinctions regarding STIs was the Swiss-Spanish-Intermittent-Treatment-Trial (SSITT). This study included 133 subjects with chronic HIV infection who received antiretroviral treatment on a schedule of eight weeks on therapy, followed by two weeks off.6 Although there was evidence of an expanded CD8+ lymphocyte response to HIV antigens as a result of the interruptions, this was not associated with viral load reductions. In fact, subjects with higher numbers of HIV-specific CD8+ lymphocytes had higher viral load increases during their interruptions. It is quite possible that during the STI and viral rebound, the destruction of CD4 cells by viral growth was the dominant effect, which overwhelmed any beneficial immune response. As expected, during treatment interruptions, CD4 counts fell. The negative results of this large, carefully performed study, makes a strong case against the "autoimmunization" theory of STI in chronic HIV infection.
Of course, a concern with any STI is the risk of developing viral resistance. However, one of the important findings from the SSITT study was that the risk of developing drug-resistant virus was quite low. In over 500 treatment interruptions, drug-resistant virus developed in only two cases. This low rate of development of viral resistance may largely have been a result of the relatively short two week time period off treatment. Other studies with longer treatment interruption times have been complicated by higher rates of viral resistance.8
Another significant finding from a sub-study of the SSITT study7 was that 14 of the subjects in the study had frequent blood sampling during the 14-day STI in order to determine the tempo and trajectory of viral rebound with the interruption. Viral load testing was performed during the STI at day 4, 8 and 14. There was minimal evidence of viral rebound seen at day 4, with less than 10% of the viral load test results greater than 100 copies/mL. However, by day 8 almost half of the subjects had viral loads above 100 copies/mL and by day 14 all but two subjects consistently had viral loads greater than 100 copies/mL. These data suggest that an STI lasting more than a few days will pose a significant risk for the development of viral resistance with repetitive cycles of replication. In fact, eight of 14 patients in this sub-study were found to have minor variants of the m184v mutation. However, the clinical relevance of these minor populations is of uncertain clinical significance. These findings are quite pertinent to the design of structured intermittent therapy studies, and suggests that with current antiretroviral agents, interruptions of greater than several days may be problematic.
The primary focus of structured intermittent therapy (SIT) is to reduce long-term exposure to antiretroviral therapy, thereby lowering a patient's risk for toxicity as well as reducing medication costs. It is also hoped that with this strategy patients may experience an improved quality of life as a result of a break from the pressure of daily medications and associated treatment fatigue. The main risk of an SIT is the premature development of drug resistance with the attendant compromise of future treatment options because of cross-resistance.
One of the primary educational messages regarding HIV management has been the importance of a patient's strict adherence to his or her antiretroviral therapy. The link between missed doses and the development of viral resistance has been well established. Although the following SIT data are interesting, this strategy of purposefully skipping doses on a schedule will require careful study in large, randomized trials.
Dr. Mark Dybul from the U.S. National Institutes of Health reported the results from a study on long-cycle SIT which was designed to compare eight weeks on therapy followed by four weeks off with a control population on continuous therapy.8 The results were disappointing; an unacceptable rate of viral resistance occurred in the SIT group (three of eight patients receiving efavirenz [EFV, Sustiva, Stocrin] developed resistance) and so the study was terminated prematurely.
Although this long-cycle SIT study was not designed to test for autoimmunization, there was neither evidence of progressive reduction in viral load during the series of interruptions nor enhancement of HIV-specific CD4 cells in the SIT group. Also, there were no significant differences in lipid levels between the SIT and control groups. As a result of this study, enthusiasm for long-cycle SIT has waned, but short-cycle SIT still beckons.
In fact, it looks as if short cycles of interruption may still have a future. The first short-cycle SIT report was a pilot study, also by Dr. Dybul, of one week on, one week off therapy with ritonavir (RTV, Norvir)-boosted indinavir (IDV, Crixivan)-based therapy in 10 subjects.9 This strategy was surprisingly successful, with controlled viral levels seen at 52 weeks of follow up. An additional benefit of this approach was the improvement in patients' cholesterol and triglyceride levels compared to baseline determinations.
Another small study of week on, week off therapy was also reported by Dr. Dybul -- in this case with eight subjects treated with an efavirenz-based regimen.10 There was no evidence of viral rebound or viral resistance at more than 52 weeks of follow up. In addition, no overall changes were observed in lipid levels. Of course, this may be due to the fact that efavirenz is known to have less of an impact on lipid levels compared to the boosted indinavir in the first trial.
A failed study of week on, week off SIT was the Staccato study. This study was different from prior studies in that it was performed in Thailand and most of the subjects received ritonavir (100 mg) together with 1,600 mg saquinavir (SQV, Invirase, Fortovase) once daily with two NRTIs. More than half of the subjects (19/36) randomized to week on, week off therapy had virologic failure, most within the first 12 weeks, so this study was terminated prematurely.11 The PI-treated patients with virologic failure had genotypic testing performed and no PI mutations were noted. All patients, except for one, were re-suppressed after resuming continuous therapy with the same regimen.
Of interest, seven of the eight subjects on the efavirenz-based therapy maintained virologic control (mean 19 weeks). The one person failing had a large body mass and a low efavirenz trough level prior to the change to intermittent therapy.
The results of the Staccato study may curb the interest for this investigational week on, week off strategy (at least for saquinavir-based therapy) for the management of HIV infection. However, as newer, more potent agents with better pharmacokinetics and longer half lives become available, this situation may change.
The first study utilizing a five-day on, two-day off SIT format was recently reported.12 In this trial from Uganda, 31 subjects were randomized to either a week on, week off regimen, a five-days on, two-days off schedule or continuous therapy. The majority of subjects received an efavirenz-based combination. There were three cases of virologic failure overall, two in the continuous-therapy group and one in the 5/2 group; all virologic failures were related to non-adherence. Additional studies of this nature are underway and the results will determine if a 5/2 strategy has merit.
The 5/2 stratagem is interesting because it fits with a typical work schedule. This may appeal to some patients who would welcome a weekend free from HIV treatment. It also might reduce one of the logistical obstacles to expanding directly observed therapy (DOT) programs. DOT for HIV infection holds some promise to help difficult-to-reach populations who are at high risk for non-adherence. Weekend coverage by medical staff for DOT is problematic; it would be beneficial if DOT could be accomplished within the confines of a standard Monday through Friday work week.
One of the pressing questions in HIV management is what to do about patients who began therapy with lower CD4 counts, but, following immunologic recovery, have CD4 counts now in the normal range. Are these patients obligated to continue life-long therapy or is it reasonable for them to stop at some point? An additional group to consider is the numerous patients who began antiretroviral therapy when the U.S. government treatment guidelines recommended therapy at higher CD4-count thresholds. Is it safe and advisable to discontinue treatment in these cases?
One of the proposed strategies to deal with this thorny issue is CD4-count-driven therapy. The notion is to start or resume antiretroviral therapy when a patient's CD4 count falls below a critical threshold and then to interrupt therapy when the patient's CD4 count rises above a specified point. Depending on the CD4-count parameters selected for starting and stopping therapy, this strategy usually translates into relatively long periods of time off therapy, often months to years.
One such trial, BASTA (meaning "stop" or "enough" in Italian), has helped clarify this issue.13 This is an ongoing, prospective, randomized and controlled trial that originally enrolled subjects with CD4 counts more than 800 cells/mm3 and plasma HIV-RNA less than 50 copies/mL on HAART. Subjects were randomly assigned (2:1) to either interrupt HAART or continue continuous therapy. In the STI arm, the goal was to maintain a CD4 count of greater than 400 cells/mm3. The primary endpoint of the study was the proportion of patients whose CD4 count remained above 400 cells/mm3. Secondary endpoints included predictors of time off therapy, metabolic changes, tolerability, costs and virologic resistance. There were 114 patients enrolled in the study (76 STI, 38 control). After a mean of 18 months of follow up, 21% of the STI group restarted therapy because their CD4 counts fell below 400 cells/mm3. The nadir CD4 count was the only predictor of the need to resume therapy; those with a nadir below 200 CD4 cells/mm3 restarted within 10 months.
|Nadir CD4||Mean Duration of First Treatment Break|
|<200 cells/mm3||6.9 months (not over 10 months)|
|200-350 cells/mm3||14.1 months|
|351-500 cells/mm3||17.8 months|
|>500 cells/mm3||all continue off therapy|
BASTA is a proof of concept for CD4-count-driven therapy. Results show that HAART may safely be discontinued in patients with more than 800 CD4 cells/mm3 who show no evidence of the development of virologic resistance. There is a caveat though: Those with nadir CD4 counts below 200 cells/mm3 will likely have to resume sooner rather than later.
Another variable-schedule STI study underway and still enrolling is the "Strategies of Management of Antiretroviral Therapy" (SMART) study.14 This large CPCRA-sponsored trial is planning to enroll 6,000 subjects and is scheduled to run for nine years. The intention is to compare the maintenance of viral suppression on continuous antiretroviral therapy versus on a CD4-driven STI strategy. Subjects with greater than 350 CD4 cells/mm3 are eligible for enrollment. The study design calls for treatment to be stopped in the STI group and then resumed when CD4 counts dip below 250 cells/mm3. The viral suppression group will be treated with HAART for the duration of the study with the goal to maintain an undetectable viral load. A number of sub-studies in SMART are planned; a partial list includes an evaluation of cardiovascular complications, risk of HIV transmission and neurologic complications.
Some patients with multidrug-resistant virus have limited or no treatment options. An STI may lead to partial or complete reversion to wild-type virus over the course of weeks or months. The hope is that by allowing this to occur, such patients may have a better response to salvage therapy, even though they have had prior documented virologic resistance to the regimen.
The French GIGAHAART Study suggested that a pre-salvage STI approach had potential merit.15 This randomized trial with 68 participants looked at patients who had drug-resistant virus and median CD4 counts of 27 cells/mm3. Those who received an eight-week STI prior to restarting an aggressive salvage regimen (with a mean of 7.3 drugs) had a significantly better response to therapy. Patients in the STI group experienced a mean viral load reduction of 1.91 log copies/mL compared to an only 0.37 log copies/mL reduction in the control group of patients who were switched straight off to the salvage regimen.
The much larger CPCRA trial 064 (270 vs. 68 participants) came to a different conclusion.16 However, the design of this trial was different in a number of ways. Although the subjects had multidrug-resistant virus, their median CD4 count of 144 cells/mm3 was substantially higher than that of the GIGAHAART study. In addition, the STI duration was longer at 16 weeks and the salvage regimen contained fewer medications (mean 3.6). The results were divergent as well; there were no differences in the viral load changes between the STI and immediate therapy groups at 12 weeks of follow up. This study was terminated prematurely due to a higher rate of adverse clinical outcomes in the STI group. There was the same number of deaths in both groups (eight each), but higher rates of opportunistic infections in the STI group. Presumably, the higher rate of infections in the STI group was related to the considerable decline in CD4 counts during the 16-week STI. We still don't understand the underlying reasons for the disparate results of GIGAHAART and CPCRA 064. The differences may be due to the shorter time off of therapy in the GIGAHAART study or its more aggressive regimen.
More pre-salvage STI studies are underway or in the planning phase in order to better understand the dynamics of viral rebound and resistance and to clarify the differing results of prior studies. However, based on the negative results of CPCRA 064, this strategy is not recommended outside of research trials unless individual patient considerations require a temporary treatment interruption.
Although the topic of this article is structured interruptions, it must be recognized that in the real world of HIV clinical care, patients are interrupting therapy on their own. These patient-mediated interruptions occur despite strenuous efforts by HIV-treating clinicians to encourage their patients to have a high level of adherence to continuous antiretroviral therapy. Many of these interruptions also happen without the medical team's knowledge and are first recognized when viral breakthrough is noted on a routine follow up of blood testing.
One of the least appreciated factors in patient-driven treatment discontinuation is the influence of medication-related side effects. Although antiretroviral therapy has improved significantly in the last several years, drug-related toxicity is still common. Even in the absence of a definite relationship between symptoms and a medication, some patients will discontinue therapy because of a perceived association. In fact, it may simply be a media report about a newly discovered problem with an antiretroviral agent that will trigger someone to stop therapy.
There are economic factors that also influence treatment interruptions. Even individuals with private health insurance are impacted by financial considerations in HIV treatment. There are considerable costs related to medication co-pays and deductible and non-covered medical fees. There are also travel expenses and time lost from work that have an adverse financial impact. For other patients who access care through the public systems, there are costs involved as well in negotiating the system and sustaining benefits.
Psychological issues also have a dramatic impact on the decision by individuals to remain on or discontinue therapy. Many people with HIV are overwhelmed by the challenges of accessing healthcare, keeping up with doctor visits, blood testing and the necessity to take medications on a daily basis.
Moreover, the warning that viral resistance will be the penalty for missing doses is a major deterrent for those patients that question their ability to stick with a program. As a result of these concerns, some patients simply opt out of treatment for a while in order to take a break from the responsibilities that are required to participate in treatment. These patients are not necessarily nonadherent, in fact some of them never miss a dose of medications while "on treatment." After a period of time off medication in which they have had a chance to enjoy a less regimented life, they are ready to marshal their personal and emotional resources and resume treatment.
There are additional considerations and precautions for STI use in clinical practice and research studies.
One important issue is hepatitis B (HBV) coinfection, which occurs in 5 to 10% of patients with HIV infection, depending on the population considered. Several commonly used antiretroviral medications have potent activity against HBV. These include lamivudine (3TC, Epivir), emtricitabine (FTC, Emtriva) and tenofovir (TDF, Viread). Many clinicians select therapy for coinfected patients with the intention to treat HBV infection as well as HIV. Data are limited, but in one small sub-study the combination of lamivudine and tenofovir in HBV coinfected patients led to an average 4.7 log decline in HBV DNA at 48-weeks follow up.17
The concern is that if a patient coinfected with HBV goes on an STI there will be a spike in HBV replication shortly after treatment is withdrawn. There have been cases of acute hepatitis, even fatal, after the discontinuation of lamivudine in coinfected patients.18 Thus, there should be careful consideration of STI in coinfected patients because of the potential for a flare up of HBV upon withdrawal of lamivudine, emtricitabine or tenofovir.
The public health ramifications represent another dimension that must be considered in the study of STIs. HIV-RNA levels are one of the strongest predictors of the rate of sexual transmission of HIV infection.19, 20 Therefore, the risk of HIV transmission would likely be enhanced during an STI, based on increased viral load levels in plasma and genital secretions.
Clearly the toxicities of antiretroviral therapy will be the driving desire behind continuing research into STI. But in addition to the potential for life-threatening toxicity from antiretroviral therapy, there are also the financial costs, especially apparent in underdeveloped parts of the world where treatment is not available, even with substantial price reductions.
This past summer at the International AIDS Society conference, Dr. Bernard Hirschel presented this theoretical model to gauge the financial impact of antiretroviral therapy on a modern industrialized country versus a developing country.21
|Total Population||7 million||12 million|
|Gross National Product||$245 billion||$6.8 billion|
|Antiretroviral Therapy Cost/Year/Patient||$12,000||$300|
|Antiretroviral Therapy Cost/Year||$144 million||$450 million|
|% Gross National Product for Antiretroviral Therapy||0.06||6.6|
Since a country typically spends about 10% of its Gross National Product (GNP) on total health expenditures, 6.6% for antiretroviral therapy alone is not within the realm of possibility based on historical precedent.
If antiretroviral costs could be reduced further through a SIT program, this would allow more patients to be treated in locales with limited financial resources. This in not only the case in underdeveloped countries; currently in the United States there are financial shortfalls in state Medicaid and ADAP programs and some patients on waiting lists do not have access to medications.
If proven effective, SIT would offer the prospect for reduced antiretroviral toxicity and expansion of treatment slots in financially-challenged regions of the world.
In general, STI or SIT should still be considered experimental and not be performed in clinical practice. One exception would include patients who began antiretroviral therapy when treatment was recommended at higher CD4 count cut-offs. These patients may safely discontinue therapy and restart based on contemporary guidelines.
Another group of patients who may consider an STI outside of a trial are those who have immunologic reserve (based on CD4 counts above 350) and who are experiencing medication-related toxicity. Especially if nadir CD4 counts were not below 200 and pre-treatment viral load levels were not high, these patients may be able to safely stop therapy for an extended period of time. The decision to withdraw treatment in this setting depends on the severity of side effects, presence of viral resistance and availability of antiretroviral options. Based on patterns of antiretroviral use from observational cohorts, it appears that more patients are on treatment interruptions.22 It is likely that many of these patients off therapy fit into one of these categories with higher CD4 counts and/or medication-related side effects.
And finally, one of the most passionate arguments for STI still will come from HIV-infected patients, many of whom are worried about the long-term toxicities of antiretroviral therapies, or from others who just need a break from daily treatment adherence. Although we do not yet have clear answers, ongoing studies promise to advance our understanding and define the appropriate place for STI in HIV practice.
For Structured Treatment Interruptions Part II: Case Studies, click here.
|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.|