Starting HAART: Should 500 Become the New 350?
A Discussion With Mari Kitahata, M.D., and Daniel Kuritzkes, M.D.
October 26, 2008
Table of Contents
One of the most talked-about studies at ICAAC/IDSA 2008 concerned the latest developments in the "when to start" debate. Mari Kitahata, M.D., unveiled intriguing data in a late-breaker slide presentation that sought to answer the question: Are HIV-infected patients better off starting therapy when their CD4+ cell count drops below 500, rather than 350?
In this press conference, Dr. Kitahata and Daniel Kuritzkes, M.D., discuss the findings and their implications. The two then take a range of questions from reporters.
The press conference opens with Dr. Kuritzkes briefly summarizing new developments in HIV drug resistance.
To try and give you, in a nutshell, an overview of what I'll be covering in my talk: In the last year and a half or so, we've seen three new drugs in novel classes approved: raltegravir [MK-0518, Isentress] as an integrase inhibitor; maraviroc [MVC, Selzentry, Celsentri] as an entry inhibitor; and then etravirine [TMC125, Intelence], which may not technically be a new class, but is the first second-generation non-nucleoside reverse transcriptase inhibitor to be approved.
These drugs, in aggregate, have revolutionized the care of patients with highly drug-resistant virus. But each of these drugs, in and of themselves, can select for drug resistance. The selection of these drugs [for inclusion in an antiretroviral treatment regimen] depends on the presence or absence of preexisting resistance in the virus population. So I will be summarizing the information about those drugs. We've already heard some new information about two of those drugs, etravirine and raltegravir, at this meeting and over the summer at the [XVII International AIDS Conference].
The other aspect of drug resistance that I'll be covering is the concern about the presence of minor variants. As you know, the HIV viral population in any given individual is highly diverse. We talk about a "quasispecies," which is a swarm or cloud of closely related viral sequences, rather than one particular sequence that's found throughout the body, so that if you took many samples from a patient and cloned and sequenced them, each sample would give you a slightly different sequence. These circulating variants can include variants that already carry the mutations that confer resistance to drugs that are being used -- the rationale for our needing combination therapy, and particularly three-drug combination therapy.
Although we have resistance tests that can detect the presence of resistant variants when they make up the bulk of the population, the current tests that are used routinely in clinical care do a poor job of detecting minor variants. There has been accumulating evidence -- some from our group, some from the [U.S.] Centers for Disease Control [and Prevention], and some from other groups -- indicating that the presence of minor variants is indeed of clinical significance, and detecting these variants with newer, more sensitive techniques can identify patients who are at higher risk of failing on specific regimens. If these tests can be developed to become more clinically useful, in terms of their throughput and practicality, then these could in fact be assays that might find a role in the clinical management of patients.
Our research suggests that patients with HIV have less risk of dying if they start sooner than currently recommended. A major challenge for any study addressing this question, whether it be a randomized trial or an observational cohort study, is the requirement for large numbers of patients to be followed over many, many years to examine clinical outcomes. We combined data from the major HIV research cohorts in North America -- in Canada and the United States -- through the International Epidemiological Databases to Evaluate AIDS Project, which is sponsored by NIAID [the U.S. National Institute for Allergy and Infectious Diseases].
The North American cohorts in this study are prospective, ongoing studies with collection of demographic and clinical data on enrolled participants. We examined patients from 1996 to 2006 who entered the study with a CD4 count between 351 and 500, to test the hypothesis that initiating therapy in that stage of disease was associated with better survival than deferring therapy [until] a lower CD4 cell count [is reached]. The total patient population was over 8,000 patients, 30% of whom immediately initiated antiretroviral therapy at a CD4 count between 351 and 500, while 70% deferred. The deferral group includes patients who initiated therapy immediately after their CD4 dropped below 350, but it also includes patients who may not have initiated at all. To address that, we used state-of-the-art methods to analyze the data while appropriately weighting those patients in the multivariate analysis.
We found a 70% improvement in survival for patients who initiate therapy between [a CD4+ cell count of] 351 and 500. The patients in this study represent the diverse characteristics of patients in routine clinical care in North America, unlike more selected populations that may be enrolled in a clinical trial, so these results are generalizable to the diverse population of patients in care where antiretroviral treatment is routinely available.
I think the important part of this is the magnitude of the effect of benefit on survival with antiretroviral treatment at this stage, and the size of our cohort. This is the largest study examining this question, comparing patients head-to-head from the [time they share the] same CD4 cell count level to [the time of their] death. These data strongly support the use of antiretroviral treatment for patients at a CD4 count of 500 and below, regardless of the presence of symptoms.
Reporter #1: Are all patients in this 350 to 500 range receiving drug therapy now? I'm also wondering if Dr. Kuritzkes could comment on this data, and what he thinks the implications are.
Mari Kitahata: Yes. Everyone in the study initiated a potent, highly active antiretroviral combination of therapy. Because the study spanned 1996 to 2006, the majority of people were starting non-boosted, PI-based regimens, but the distribution of types of regimens were similar between the two groups receiving immediate treatment and the deferral group.
Reporter #1: I asked my question inelegantly. I meant, for people who delay -- if you're following the guidelines and you delay treatment from 350 to 500 -- do they get any treatment at all? Or is it just HAART that's delayed?
Mari Kitahata: Well, HAART is the antiretroviral therapy that they would be receiving. As far as other treatments for HIV: It isn't until your CD4 count is below 200, for example, that you would be receiving prophylaxis for Pneumocystis pneumonia or, below that, prophylaxis for MAC [mycobacterium avium complex]. So at that high level of CD4 count, the primary treatment would be antiretroviral therapy.
Reporter #1: You mentioned the cohort being over 8,000 patients, but your abstract seemed to talk about 6,000. If you could clarify?
Mari Kitahata: This was a late breaker. Since that [abstract] was submitted, we added an additional cohort that increased the sample size even further.
Daniel Kuritzkes: I think these are really very important data. There have been a variety of cohort studies that have shown differences in the time to AIDS, or time to death, in patients who have initiated therapy below 200 CD4s [as well as patients who have initiated therapy with a CD4 count from] 200 to 350, which is where our current guidelines come from. But many of these cohorts, although they have suggested trends favoring earlier initiation of therapy for patients at earlier stages [of HIV disease], haven't had enough people. They have lacked the power to really be able to find meaningful differences in this higher CD4 stratum. And I think the importance of this study is, by aggregating all these North American cohorts together, they have really gotten enough patients to have the power to see this important difference, in addition to the methods that were used to correct for a number of the biases that might be intrinsic to doing a cohort study as opposed to a randomized trial.
Reporter #2: The guidelines released this summer suggested increasing the range in which people would initiate therapy. Is this data strong enough to suggest that practice should change yet again?
Mari Kitahata: The U.S. DHHS [Department of Health and Human Services] guidelines released in January of this year said that patients [with a CD4+ cell count] below 350 should receive treatment -- these are asymptomatic patients we're talking about. It's very clear, as Dan points out, that patients with CD4 counts lower than 200 or clinical AIDS should always be treated. Those guidelines stated that the evidence was insufficient to recommend treatment above 350.
The IAS [International AIDS Society]-USA guidelines that came out in August in JAMA recommend, again, treatment for all patients less than 350, and also talk about the importance of being on treatment by the time your CD4 is 350; but specify that there is limited data to create a threshold above that, and speak of individualizing therapy above a level of 350.
What our data show, I believe -- given the size and the diversity of the population -- is a recommendation that all patients who have a CD4 cell count of 500 and below receive antiretroviral treatment. So this does differ from current guidelines, and recommends treatment earlier in [the course of a patient's HIV] disease.
Daniel Kuritzkes: I think you'd have to ask the folks on the guidelines panels whether they consider these data to be sufficient to change the guidelines. I think it certainly is a strong push in the direction of changing the guidelines, [but I don't know] whether the panels would act on a single study, no matter how well done, or would [instead] look for corroborating data from additional cohorts, as they have done in the past: We had data on the 350 benchmark compared to 200 for a little while before there emerged a consensus about bumping up from 200 to 350 for starting [treatment]. I suspect the guidelines panels are going to be looking for confirmatory data from some of the European cohort studies, or others, [before they] actually change the recommendations.
Reporter #2: That's not quite the question I asked.
Daniel Kuritzkes: You asked if the data were stronger enough to change guidelines.
Reporter #2: No, I said practice.
Daniel Kuritzkes: Oh, practice. Ah, I'm sorry. Practice, OK. To change practice, I would say yes, with one caveat -- and it isn't really an issue of not following through on the data. But if you look at the recent, large, randomized trials that have been reported, including the STARTMRK study that was reported today, the average CD4 count of people who are beginning on treatment -- which represent in many centers the people who are showing up for care -- is in the mid-200s. In our center, a quarter of people get diagnosed with HIV for the first time because they were hospitalized with an AIDS-defining illness. The CDC [U.S. Centers for Disease Control and Prevention] has data showing that a third of patients have an AIDS-defining illness within a year of being diagnosed with HIV. That's because people are still being tested and identified as being HIV infected far too late.
So I would say yes; these data would make me have discussions with my patients who are in this 350 to 500 range that would be leaning much more strongly towards treatment, as opposed to saying, "Well, let's wait and see what happens." However, there are only a handful of such patients that we have in our clinic. I think that's the big challenge now: to identify patients who are infected and get them into medical care, so that these findings can be applied to them.
Reporter #3: I would just like to follow up on the question that [Reporter #2] didn't ask, that you answered: What will it take to get confirmatory studies, given that this is a big, long-term, prospective study? Who is going to be able to do something similar? And in how much time?
Mari Kitahata: I would say that's a critically important question. The Europeans in the ART-CC [ART Cohort Collaboration], for example, have addressed this issue, but their study includes only patients who have initiated HAART. So they're not able to compare, head-to-head, patients who initiate and don't initiate, which is what our study does.
There have been many studies that look at the prognosis of patients after initiating HAART at different CD4 cell counts. However, if you look at people [who start treatment] with different CD4 cell counts, and do not compare them to those at the same CD4 cell count who don't start, you introduce lead-time bias. Lead-time bias means that you are missing events that would occur during deferral of therapy. Those events, those missing data, have to be imputed. Generally, data from the pre-HAART era is used to impute that information and make these conclusions. There are a few studies that have shown that the disease progression of patients in the pre-HAART era is not the same as those in the current treatment era who aren't treated.
I think there are pluses and minuses to the different approaches to doing this study. But one of the important differences, in addition to the size and the length of follow up of our study, as well as the fact that it encompasses all of the major cohorts in North America, is the ability to compare head-to-head the two groups of patients from the same stage of disease.
Reporter #4: I know that you have to work with the statistical numbers that you have to work with -- you know, the 350 to 500 number. But when it really comes to the bottom line and treating the patient, isn't it more the trend and the slope of the trend, rather than the absolute number? How would you address that?
Mari Kitahata: Yes, I think it's very important to look at differences in the slope of the decline. Certainly, a rapidly declining CD4 cell count would make you and your patient more concerned about initiating therapy earlier. I think the point Dan made about who's getting into care is a critical one here. Because in our clinic, also, we're seeing people who are presenting much later in [their HIV] disease than we would like, and are often admitted to the hospital as their first point of care with an opportunistic infection.
Again, when the guidelines focused on [starting therapy when a patient's CD4+ cell count is in the range of] 350 to 200, what I was finding, as a consultant for the whole Northwest region for physicians calling in, was that they were interpreting [that range as a recommendation that they begin] starting people at around 200.
I think each individual person's trajectory of CD4 decline needs to be taken into account. But what this says is that we should be looking at the 500 CD4 level, and beginning to make decisions with patients at that time, rather than waiting till the lower levels.
Reporter #5: Do you feel there's a CD4 level where there is no disease progression?
Daniel Kuritzkes: I think people can be considered to be progressing in their HIV disease from the time that they become infected. I think there's ongoing damage to the immune system; there's ongoing inflammation; and there may be, as a consequence of that inflammation, additional adverse events that are not classically considered to be AIDS-defining events [but] that may be the result of ongoing virus replication.
The critical issue about whether intervening in those patients makes sense or not, or at what point it begins to make sense, really has to do with something that Dr. Kitahata alluded to in the Q-and-A during the abstract session, which is the number needed to treat -- and then, related to that, the cost effectiveness of treatment. If you treat people at very high CD4 counts, it's possible you'd be treating a very large number of people, most of whom are at minimal risk of having any meaningful progression over the next several years, to avoid a very small number of events occurring. And then, as the CD4 counts get lower, you wind up having to treat fewer people to prevent similar numbers of events. Another way of looking at it is: There are more and more events that might be occurring, and so it makes more and more sense to treat.
What I think is likely to happen is, as we continue to push the threshold for starting therapy earlier and earlier, we're also going to get better at identifying the patients at greatest risk of progressing. Because not every patient has the same risk of progressing, and there are a variety of factors -- some of them have to do with host genetics, some of them have to do with the kind of virus the patient has, and some can be determined by the CD4 slope that has been referred to -- and find some way of assessing an individual patient's risk [and thus] be able to fine tune who gets treated earliest, and who can wait a little bit longer.
Mari Kitahata: There is emerging data about the impact of chronic, uncontrolled HIV replication and organ damage, and the importance of maintaining the immune system before there is irreversible damage. I think, as Dr. Kuritzkes mentioned, the non-traditionally AIDS-related comorbidities are an important area. But clearly, antiretroviral therapy has revolutionized HIV from a very difficult disease to a chronic infection, since the introduction of active therapies. So it's important to note the great gains that have been made in helping people maintain longer and more normal lives with this infection.
Reporter #6: Dr. Kitahata, I believe in your presentation you said that analysis on the group [of patients who initiated therapy with a CD4+ cell count] over 500 was underway?
Mari Kitahata: Yes. We have a new study that is examining [this]. We have added an additional cohort, and we're looking at the survival differences in patients who initiate therapy at CD4 counts greater than 500 versus those who defer therapy to lower levels. Those results will be available shortly.
Reporter #6: By "shortly," do you mean maybe CROI [the 16th Conference on Retroviruses and Opportunistic Infections, which takes place in February 2009]?
Mari Kitahata: Perhaps.
Reporter #7: The other thing you said is that there was a 70% improvement in survival. That sounds very impressive, but what type of absolute numbers are we talking about? I mean, 70% of a very small base is still a small number.
Mari Kitahata: A relative hazard of 1.7 is significant, and substantial. We did a number of sensitivity analyses, looking at the potential for unmeasured confounding to have entered into our results, and found that a factor would have to be uncommonly large to have mitigated our results. The 70%, in this kind of analysis, is a meaningful and substantial risk of progression and death. I don't have the absolute number off the top of my head; that isn't the way that we looked at it. But it is a substantial and significant increase in the relative risk of death.
Reporter #8: Once the person starts therapy after deferral, is that risk [of mortality] maintained, or do you expect it to decline over time?
Mari Kitahata: What the study does is, it examines patients who initiate within that 350 to 500 [range], versus those who defer. The initiating people are the reference group. So the relative risk is the relative risk of increased mortality for those who defer therapy. What it says is, for those who don't initiate at that higher CD4 cell level, they have a 70% increased risk of dying [because they are] starting at lower levels. We don't want to go beyond the data we're able to show in this kind of analysis.
Daniel Kuritzkes: I think your question gets to exactly what was unique about the NA-ACCORD study here, compared to some of the other cohort studies. In previous cohort studies, you would be comparing survival in the people who started antiretroviral therapy at CD4s of 350 to 500 with people who had started therapy below 350. But the only people who can start therapy below 350 are the ones who survived to start therapy below 350; you're missing all the people who have died along the way. Here, you're really comparing the survival rates of the people who started right away versus the people who waited. So it's not quite asking, "What happens if you wait?" or, "Once you start therapy, do you still do OK?" Because generally that's tended to be true until your CD4s get very low. But you're not counting the people you have lost along the way. This study seems to suggest that you're losing people along the way.
Mari Kitahata: That is a very good description of the difference between [this study and] the studies that examine prognosis of patients from different CD4 cell count levels, which are the majority of studies that have been done prior to this. Again, this study is looking at people at the very same level.
We are, however, comparing to people who [begin treatment] at less than 350. So it is a comparison between people who are starting at 350 versus starting later; and that's, I think, again supporting the importance of the magnitude of the difference.
Reporter #9: For Dr. Kuritzkes: There is a drug in a fairly late stage of development -- PRO 140, I think it is, a monoclonal [antibody]. Can you talk a little bit about the possible place that type of approach, with its different mode of administration, might have in therapy?
Daniel Kuritzkes: I think that, as we have developed larger numbers of orally available antiretrovirals, the hurdle for injectable or infusible drugs has become greater. The bar has been raised. I think one could imagine that if several of these drugs could be developed and brought to clinical use, and they could be administered jointly, it would be much easier to provide directly observed therapy on a monthly basis, which could have advantages for a number of patients.
The other potential advantage of PRO 140, within the class of CCR5 antagonists, is that in vitro work has shown that the antibody is active against viruses that have developed resistance -- or have been selected to have resistance -- to the small molecule inhibitors like maraviroc, which is now approved, and vicriviroc. There are a relatively small number of those resistant viruses right now, however; the main reason that people seem to fail on the CCR5 inhibitors is that the other kind of virus emerges, the one that uses a different receptor. This antibody would not be helpful in that case.
I think there are a number of potential roles for it, but it's a very high bar, and I think we'll have to see how the studies really play out before we can say exactly where this would fit in.
Reporter #10: A quick follow-up for Dr. Kitahata: How many people were in each of the two groups?
Mari Kitahata: The total number was 8,374; 5,901 deferred therapy and 2,473 initiated.
Reporter #11: Do you have any thoughts about how many more people in the U.S. would be in treatment if we had a threshold of 500, as opposed to a threshold of 350?
Mari Kitahata: I don't have that number off the top of my head. Richard, do you?
Richard: It would likely be several hundred thousand, because currently, only 60% or so of people who ought to be on therapy by the old standards of 200 are on treatment, according to CDC estimates. So increasing the bar to 350 or 500 would have to mean several hundred thousand additional patients.
Reporter #11: Could you talk about what this implies for the importance of testing?
Mari Kitahata: Again, I think this points to the importance of knowing that you are HIV positive to be able to enter into care, to have access to these therapies. I think Dan has mentioned that a very important part of our being able to make a difference is to be able to work with people earlier in disease, before they become sick. So I think it's a very important part of the treatment piece.
Reporter #2: A quick question for you, Dr. Kuritzkes, and then I have a question for both of you. The question for you is: If large numbers of people are being treated at a higher CD4 level, presumably they will be treated longer? What are the implications for resistance? The question for both of you is: There was another late-breaker [presentation] today. I wonder if you could try and put that in context with the one that was just before yours.
Mari Kitahata: Dan can speak to the resistance issue better than I. But I'll start by saying that the importance of our study, in any event, was that the decade of 1996 to 2006 includes the older treatments, which were more toxic, less well tolerated and often given three times a day. And yet, with those people included in this study, we found a significant survival benefit. Newer regimens are much easier to take, more tolerable [and have] less frequent dosing. All of that speaks to the importance of being able to be adherent to the regimens, and that's the key to avoiding resistance, of course.
Fortunately, we have, for those of our patients who are in that situation, new classes of drugs and new drugs within classes, like etravirine, that are not resistant to the older drugs. So there are additional options and subsequent options. Of course, the goal is to have more options and to maintain viral suppression for as long as possible with these newer drugs.
Daniel Kuritzkes: I agree with what Mari is saying. I think one useful comparison is that Andrew Phillips from London had published data several years ago in The Lancet suggesting that over [the course of] six years or so, the risk of [failure among] patients who were starting on a first antiretroviral regimen was around 40 percent, with about 30 percent developing resistance of some sort. But those were with the older regimens. They have updated that analysis, and just last year published again in The Lancet a paper showing that, in fact, for people starting what we would consider temporary regimens, the risk over several years was only 10%. So I think what we're seeing is that the better regimens -- because they are better tolerated, because they are simpler and easier to take, with fewer side effects -- are also enhancing overall adherence to therapy, and therefore leading to less resistance.
Reporter #12: [off microphone; inaudible]
Daniel Kuritzkes: The STARTMRK study that Jeff Lennox presented? I think it goes along with the general trend we have been seeing over the last several years of continuing improvements in regimens that are available for first-line therapy. I think exactly how raltegravir will fit into first-line treatment is going to depend on a variety of factors. Generally, decisions about what drugs to initiate are based on individualizing care, at least here in the developed world. And I think the fixed-dose combination of tenofovir [TDF, Viread], FTC [emtricitabine, Emtriva] and efavirenz [EFV, Sustiva, Stocrin], which is available as a single pill, once a day, is going to remain quite popular among many people. But it's not appropriate therapy for quite a few people as well, especially women of childbearing age, people who may have acquired resistance to efavirenz by viral transmission, and so forth. For them, other regimens, either including raltegravir or including a boosted protease inhibitor, are going to be more appropriate. We'll see how this shakes out in terms of physician and patient preference.
Reporter #13: A follow-up regarding earlier intervention: Could you talk about the public health impact, in terms of reducing ongoing transmission of the virus with an earlier start of therapy?
Daniel Kuritzkes: There are theoretical arguments, based on work done by Tom Quinn and others, showing that among discordant couples, the lower the viral load in the infected partner, the lower the risk of transmission. The postulate is that if one could lower viral load -- or maybe it's a hypothesis -- if you could lower viral load in the infected individual, you would reduce the risk of transmission. That's never been tested directly, although there is an ongoing trial being done through the NIH [U.S. National Institutes of Health] by the HIV Prevention Trials Network, in collaboration with the AIDS Clinical Trials Group, principally in sub-Saharan Africa. [In this study,] discordant couples have been enrolled where the infected partner has a CD4 count higher than what would ordinarily be used to start therapy in that setting, and is [then] randomized to be treated or not, with the goal of seeing whether we reduce transmission to the partner. Incidentally, we'll also learn whether earlier treatment, in fact, reduces the risk of disease progression.
So I think that there could well be a public health impact. There are different models that argue for different impacts. Some models suggest that most transmission occurs among newly infected people, because they have the highest viral loads, and then people at later stages of disease, because they again have the higher viral loads, although they may at that point be sicker and have fewer partners. But there's clearly a lot of transmission occurring among people who are healthy and have steady-state levels of virus. It's a reasonable hypothesis that treating more people would reduce the community burden of viremia, and therefore reduce the overall risk of transmission. But I think that's not yet proven.
Moderator: OK, if there are no more questions, I would like to thank our participants for coming to talk with us today. This press conference is ended.
This transcript has been edited for clarity.
This article was provided by TheBodyPRO. It is a part of the publication The 48th Annual ICAAC/IDSA 46th Annual Meeting.
Add Your Comment:
(Please note: Your name and comment will be public, and may even show up in
Internet search results. Be careful when providing personal information! Before
adding your comment, please read TheBody.com's Comment Policy.)