National Institute of Allergy and Infectious Diseases Meeting PART 1 of
Bethesda, MD January 12-13
"Long-Term Studies in HIV-1 Infection Workshop"
Accompanying this report is a report written by Jeff Schouten. It follows mine below. His report discusses more closely the details of the Workshop just held in Bethesda. Unless important I will try not to duplicate his discussion.
The purpose of these reports is: (1) to discuss the meeting held last week in Bethesda, MD by the NIH to discuss conducting a long-term randomized trial in HIV to address the question of when to start therapy, and (2) to survey our community on their opinions about this proposed initiative. After reading the enclosed report about the proposed study and the meeting, please email a response back including your opinions and thoughts. We are interested to hear the opinions from the overall community. The broader community should have the opportunity to weigh in on this important question. A decision should not be made without their opinions and without due consideration by all. Jeff Schouten, MD is a physician who treats HIV and is affiliated with STEP (Seattle Treatment Education Project); Jeff is also a community representative to the CCG of the ACTG. In San Francisco, just prior to the Retrovirus Conference, a community meeting is planned to discuss these issues. All are welcome. Hopefully, this report and possibly additional ones made available by others will help in preparing for that meeting.
The meeting in Bethesda was designed to discuss the proposal of using a large randomized clinical trial to try to answer the question of when is the best time to start HIV anti-retroviral therapy. As well, the study could address clinical long-term outcome of what to start therapy with. The main question of the day was: should a large randomized controlled study be used to evaluate whether immediate or deferred treatment is better in the long term. Would starting treatment for HIV earlier in disease progression have a better long-term outcome than deferring treatment? Would therapy started at, for example, 500 CD4s show a better long-term outcome than deferring therapy until, for example, CD4s are 350 or 200. Which outcomes should be used to evaluate treatment? Outcomes that could be assessed include quality of life, survival, sustained good health status, long-term complications (increased rates of heart disease due to elevated lipids, diabetes, fat redistribution, etc.). Other considerations in evaluating long-term outcomes include the development of new drugs and completely new classes of drugs, and does age when starting therapy effect outcome. Initiating therapy at 25 years of age may lead to different outcome than starting at 45 years old. Does starting HAART when CD4s are 500 followed by treating HIV with expected new classes of drugs lead to better outcomes, or should one defer therapy? Does initiating treatment during acute HIV infection lead to better long-term outcomes than deferring therapy? This question, I think, would be easier to answer with a long-term clinical outcome study. If new classes of drugs are developed, how would this effect such a proposed study?
Main questions include Ōis this study needed, do we need to answer this question? And, is such a study feasible or doable? One concern raised was that the availability of more effective, potent, and tolerable drugs or treatments from completely new classes may alter conditions so much that results from this study would become not interpretable. In response, it was suggested that since the development of these types of drugs are uncertain that would not be a good reason not to conduct this study. Another concern raised was if such drugs become available after a few years following starting this study individuals randomized to deferred treatment might get better results, and this could also make study uninterpretable. I think it was suggested that by breaking down comparisons by year this concern could be addressed, although IĖm not sure how their suggestion would adequately address this concern. A key question about feasibility is will results of study be meaningful. Several considerations suggest that its possible the results may not be interpretable or applicable in the future. To this, NIHer Jack Killen said even if we donĖt get definite results we can learn a lot from such a trial.
My estimate was there were about 150 attendees consisting of academic researchers, ACTG, CPCRA, and VA researchers, statisticians were noticeably well in attendance, government, and industry representatives, and about 12 community representatives. Most attendees felt the study needed to be done. They felt it was important to try to answer this question. They also felt the question could be answered. On the question of feasibility of conducting such an initiative, opinions were more split. Although they would like an answer to the question, many felt the study was not doable. Don Abrams, a CPCRA researcher, talked about his attempt to start a similar project about 6 years ago. Although there was support conceptually, doctors and their patients were unwilling to let "a flip of the coin decide their long term treatment". It appears clear that this study canĖt be done without buy-in from doctors and patients. It was suggested that a large study should be preceded by a pilot study to explore if its enrollable and doable. I think this would be necessary if the decision is to move ahead. What if the results of Abrams study had shown that it was better to wait for HAART rather than start with AZT/3TC? This information was no longer relevant a few short years later. In such a case, the results of a large randomized trial could in fact be mis-leading.
Other challenges discussed were setting up large databases across the various study groups allowing for adequate data collection , sharing and analysis. Attendees discussed how in current environment study groups such as CPCRA, ACTG, industry, etc donĖt share data and it would be big challenge to get buy-in and set up system. As expected, several virologists suggested that although we donĖt have clear evidence from large randomized trials several small studies suggest that biologically starting treatment earlier is more effective. As suggested by others at this meeting, weĖve never done a study to directly answer that question. As well, the adverse events, tolerability, adherence, and complications of todayĖs treatments raise concerns about the ultimate effectiveness of HAART. Recent studies suggest that complications of HAART such as lipodystrophy and non-adherence reduce the overall percent of virologic response. Its possible that over the long-term complications may also reduce response rates, and therefore deferring treatment may be beneficial. A proposed long-term randomized simple study might be able to answer that question. Then again, it might not be able to or it might yield misleading results. Nonetheless, many attendees still want to proceed ahead with the proposed study.
Another key consideration is how ongoing and planned HIV studies would overlap with the proposal for a large randomized simple study. Many of the questions and issues raised in Bethesda as reasons for conducting the large study, such as lipodystrophy and other such complications of HAART, are being addressed in other studies. This can create a good deal of overlap and unnecessary expenditures of resources. Will enrollment be stolen from other key smaller studies, intended to answer important questions, to enroll the large long-term trial. This is a main concern of mine. However, possibly the study could be designed very simply, intended to answer only one question the outcome of immediate versus deferred therapy. There are many unresolved questions about a study that still need to be discussed.
The PHS guidelines came under a minor momentary attack but were defended. It was suggested that the guidelines may become meaningless if this question is not answered. Researchers said the guidelines raised or will raise concerns about all these related issues when discussing when to begin therapy and with what.
Some of the outcomes suggested to monitor in proposed large study were:
Key questions, which were tentatively raised but the answers to which remain
uncertain, were the size, cost, the length of the study, and who would pay. Estimates on the length of study needed ranged from 5-10 years up to lifetime
follow-up. Various estimates ranged from several thousand participants up to 30,000. Concerns were raised about being able to enroll large numbers of
treatment- nave individuals. The costs were not addressed openly very much but estimates ranged from 10-40 million and on up, depending on the size and
length of the study. Some at the meeting referred to this trial as a great expensive not simple trial or a LNSST (large not so simple study). These
points raise a question that resources such as patients and dollars for other studies in ACTG and other venues might be limited. The ACTG may have problems
enrolling their key studies. If so, this might effect our ability to get answers to some very important questions. This point needs to be discussed
more in-depth and in additional meetings following this one. I have a particular concern about this issue. Who will pay for this study? There is an
implication that the NIH can and will pay for this study. Will the pharma industry provide support forthese studies? This wasnĖt discussed. The
organizers did not specifically address follow-up meetings.
In a summary at the end of the meeting, 6 objectives were listed for the adult scientific research agenda at NIH. They were, I think--
1. Minimize Viral Replication and Tissue Burden
2. Elucidate Pathogenesis of HIV
3. Study Immunopathogenesis of OI's
4. Preserve, Restore and Enhance HIV & OI Specific Immune Response
5. Treat and Prevent Opportunistic Complications
6. Assess Long Term Clinical Outcome (the workshop addresses this objective)
I think follow-up meetings are necessary to discuss these far-reaching concerns and questions. One 2-day meeting is not adequate to give these very difficult questions a full hearing. Personally, I have some reluctance that such a study is actually necessary or that it can or will answer the question of immediate versus deferred therapy. But I am not confident in this opinion and have mixed feelings. It would be helpful to know with more certainty that immediate therapy had or didnĖt have benefits of survival compared to deferred therapy. Many others feel strongly such a study is necessary to conduct, that we are obligated to try and ascertain answers to this question for the future. I have more real severe concerns about the doability and feasibility of such a study. As a person with HIV I donĖt think I would be willing to enter this study. As Don Abrams experienced, I am concerned that doctors and patients will have doubts when they weigh the potential benefits and negatives for enrolling their patients in such a study.
At the workshop HIV was compared to other diseases and how in other diseases these types of studies are conducted. Is HIV different?
As I stated at the start of this report, one reason for writing this paper to the community is to get feedback from others on their feelings and thoughts about the idea of conducting this long-term trial. The proposed study is massive in scope and deserves due consideration and scrutiny by the broader community.Is the study neccesary, do you want these answers? Do we owe the future to try to secure the answers to these questions? Are you willing to pay the costs in money and other resources? Is this study doable, can it be enrolled? Will the results be usable? How do doctors feel about this? Please take this opportunity to express your thoughts and opinions by sending an email back to me. Finally, as I stated above many individuals at the meeting felt this study is doable, that we are obligated to try and get answers to the question, and that the study should be implemented.
National AIDS Treatment Advocacy Project (NATAP)
Community Representative of the CCG to the ACTG
A Report on the "Long Term Clinical Studies in HIV-1 Infection Workshop" PART
2 of 2
Jeffrey T. Schouten, MD
STEP (Seattle Treatment Education Project)
[Disclaimer: This summary of the Workshop is only my perspective on the major issues discussed. It is not meant to be a complete summary of the conference, but rather to highlight the areas that I felt were the major issues, from my perspective, both as a PWA and an HIV provider. I have attempted to clearly state where my views differed from the majority of those attending the meeting. I was invited to the Workshop as a community representative, and was assigned to attend the second of the three focus groups (When to Switch) reviewed below. Please feel free to redistribute this article: e-mail it to whomever you want!]
On January 12-13, 2000, the National Institute of Allergy and Infectious Disease (NIAID) hosted a Workshop in Bethesda, Maryland to discuss the feasibility of conducting trials addressing some important strategy issues about the treatment of HIV infection. NIAID invited about 140 people, including community representatives, statisticians, and clinical researchers to address the following three major areas. When to start therapy for HIV infection. When to change therapy for HIV infection. Long-term clinical efficacy and tolerability of HIV treatment.
Dr. Anthony Fauci, the Director of NIAID, opened the meeting by reminding people that there are an estimated 40,000 new HIV infections per year in the US (one half who are under 25 years old.) Additionally, of the estimated 650,000-900,000 HIV-infected people in the US, approximately 200,000 people are unaware that they are HIV-positive. Thus, obtaining answers to the above three areas are of great importance to many people.
Dr Jack Killen, the Director of NIAIDĖs Division of AIDS (DAIDS) noted that during the recent review and awarding of funds which support the major cooperative HIV clinical research networks, the ACTG and the CPCRA, two areas of priority research still remain inadequately addressed; treatment strategy trials, and long-term trials of the effectiveness and toxicity of current regimens.
The opening speakers emphasized the lack of data to answer the question of when is the optimal time to initiate and change antiretroviral therapy. Recent survey data from the Health Care Services Utilization Study (HCSUS) report that only 24% of people in care for HIV in the U.S. have viral loads (HIV RNA) suppressed to below the limit of detection. Additionally, 70% of all people who begin on HIV therapy change their therapy within the first 9 months, and the median time on their first regimen is only 4 months!
The opening session also reviewed the results of a meeting held the day before, focusing on statistical design issues and challenges in conducting such trials. There are two major trial designs to consider in conducting such research, the randomized prospective trial, and the non-randomized cohort, or observational type, trial. The gold standard of scientific research is the randomized trial. However, there are several ongoing cohort trials, observing people on therapy, and the outcome of such therapy, which appear to be providing meaningful data. However, the majority consensus at this Workshop was that the randomized trial is the optimal design to attempt to answer the first two strategy questions, when to start, and when to change.
A limitation of most current trials is that they compare surrogate endpoints, i.e. viral load (HIV RNA) and CD4 counts, rather than clinical outcomes, i.e. opportunistic infections, major toxicity, and death. This is because trials comparing clinical endpoints need to be very large and need to be conducted over a long period of time.
The first afternoon was divided into three concurrent focus groups, which addressed issues associated with each of the three major questions noted above. The following morning, all of the participants discussed the reports of these three focus groups. This was a lively discussion, which surprisingly was characterized by significant consensus on many of the challenging problems presented in conducting such trials.
When to start therapy for HIV infection.
While there is data showing clinical benefit to highly active, antiretroviral therapy (HAART) in late stage HIV infection (i.e. people with CD4 counts below 200), there remains no data which shows long-term clinical benefits to HAART therapy in people with higher CD4 counts, because such trials have not yet been done. Such trials need to be very large, perhaps involving 10-20,000 people, and conducted over a possibly 5-10 year period. These trials would be very expensive and possibly the results at the completion of the trial would be out of date, or even incorrect, depending on changes in therapy over the time it takes to conduct such a trial. Also, many individual providers and HIV-positive people feel strongly about when they want to begin therapy: and, thus would be unwilling to participate in a trial in which when they began HAART would be determined be the flip of a coin. Dr. Donald Abrams, from UCSF, noted that when he tried to conduct such a trial in 1992-93, very few people were willing to participate in such a trial. However, many people at this Workshop thought that such a trial would be feasible, or doable.
Several people noted that current Public Health Service (PHS) Guidelines recommend that consideration for therapy for HIV occur when the CD4 counts are below 500, or the HIV RNA is above 20,000 (by PCR-RNA). These recommendations are based on expert opinions, and also include substantial caveats that the decision to begin therapy needs to be individualized, None-the-less, several people emphasized that there would have to be a major educational campaign directed both to providers and people living with HIV that there is not a study which shows a clinical benefit to starting HAART early, rather than later. Another area of consensus was that such studies would have to be conducted where people were receiving their primary health care, and that the success of such a study would be in allowing people flexibility in choosing their HAART regimen. It was also noted that the ability to conduct this study would depend on the "altruism" of people willing to participate in the trial, as they would have to provide their own medications and get very little tangible benefit from their participation in the trial. Also, the trial would need to enroll people representative of the general population infected with HIV, including factors such as race, gender, method of acquisition of HIV, age, and hepatitis C co-infection.
Feasibility was another area of major discussion; concerning whether such a trial could actually be fully enrolled, whether the question was likely to ever be answered by the trial, and thirdly, would the answer be relevant to treatment options available by the time the trial is completed.
The last area to note about such a trial is whether the answer could be obtained from a "simple" trial. One could envision a trial whereby people with CD4 counts between 200-500 are randomized to start HAART immediately, or wait until some defined point later in time to begin, such as a certain decrease in CD4 count, or development of an HIV-related illness. However, there is a significant difference in designing a trial to answer the question, "Is it better to start now, or wait?", versus designing a trial to ask, "When is the best time to start HAART?" Related to this concern, will the results of such a trial provide individuals, or groups of people, with enough information for them to make an informed decision about when to start therapy. In other words, does one size fit all? Is the optimal time to start therapy the same for people recently HIV-infected, rapid progressors, long term slow progressors, people with stronger versus weaker immunologic response to HIV-infection, etc?
Many of the Workshop participants supported conducting a "simple" trial to ask, "Is it better to start now, or wait?" I expressed a minority view that a "simple" trial, which was designed without the ability to answer the critical questions of when is the best time to start HAART, and for which specific subgroups of people who are HIV-infected is the best time to start HAART, would not be a study worth the expenditure of millions of dollars of federal funds. We need to answer the question in a way that is most applicable to individuals, and subgroups, not just a statistically significant result that applies to 10,000 people, on average.
When to change therapy for HIV infection.
The major issues discussed in this focus group included what are the most important areas of uncertainty to study. There was clear consensus that the major issue to study is that of "tight versus loose" control of viral load (HIV RNA); i.e., do you change when the viral load (HIV RNA) fails to get below, or stay below 50, or 500, or 10,000, etc? Some of the concerns expressed were that the design of such a trial, and the answer to this question, would depend on the treatment options available to the person. So, that if it is at the time of first treatment failure, the switching criteria might be different than if it was after the second, or third treatment failure. There was consensus that such a trial should be a randomized trial.
The group suggested that the study might include people who, in the view of their provider, were likely to have a significant chance of obtaining good virologic suppression with a new regimen. Another suggestion was that a study might have, as a criteria of when to switch therapy, that of change in CD4 cells counts, or change in viral load (HIV RNA) from baseline. There was quite a bit of discussion about the phenomenon of "discordant responses", or people who have increases in CD4 counts, in spite of failure to have good suppression of viral load.
There was also major consensus that there is a need for trials with clinical endpoints, rather than virologic and immunologic marker endpoints, such as CD4 counts and HIV RNA. We need to know the long-term clinical consequences of "tight versus loose" virologic control. There were some concerns expressed though that, to some extent, this might depend on the specific drug regimen.
Other issues discussed were the need to include some method of assessing adherence to regimens, both in a when to start trial, and a when to switch trial. Also, the issue of how simple a trial this should be was discussed, and the consensus was that this trial could only be conducted at sites where people received their primary care, and the simpler the trial, the more likely it would be completed.
Related to how simple a trial this would be, I expressed clearly a minority opinion that the trial needed to define certain minimum standards of care that would be provided to all participants. This is particularly relevant to a "When to Switch" trial. The new revised PHS HIV Treatment Guidelines, due out in February, will purportedly recommend using resistance tests to assist in choosing agents when therapy is changed. Most people, even the other community representatives present, did not express support for this concern; but, rather felt that after randomization to one of the arms in the trials, the participants would be managed by their primary care provider as they would be were they not in a trial. Currently, all federally funded cooperative group HIV trials provide at least the minimal standard of care. I feel very strongly that as a condition of participating in a federally funded research trial, providers should be required to follow minimal standard of care guidelines, and participants should be assured that they will be provided with such minimally acceptable care. Just because there is substandard care provided to some people outside of clinical trials does not justify allowing it within a clinical trial.
Long-term clinical efficacy and tolerability of HIV Treatment.
Unlike the above two trials, there is not a need to conduct randomized trials to assess the long-term clinical efficacy and tolerability of HAART. Issues discussed included the need for long-term data on the potential impact of HAART on cardiovascular disease, lipodystrophy, cancer risk, quality of life, etc. Other issue discussed were the need to obtain consent and approval of institutional review boards to gather genetic information that may allow for better prediction of individual risks for various complications.
Additionally, the opportunity and challenges of sharing information between large trials conducted by different federally funded trials, industry trials, CDC databases, public death records, and other cohort studies were discussed. The challenges include identifying and following people across trials and databases, while simultaneously assuring privacy and confidentiality.
Clinical Endpoints, Size, and Cost.
There were several major areas of concern that could not be resolved in this two day Workshop, including; the details of what are relevant clinical endpoints, or measures of disease progression, how to weigh toxicity as an outcome, what would the size of the above trials need to be, how long would these trials need to be, and what would the cost of such trials be. While these issues were discussed, much more work needs to be done in determining the ultimate feasibility of funding and conducting these long-term trials. An important question that was not addressed at this Workshop was what would be the cost to other HIV research, if these large expensive trials were funded?
While issues needing to be addressed in the pediatric population were discussed, there seemed to be consensus that there is not a large enough pediatric population to conduct such large scale, clinical endpoint trials. Although, there remain the same concerns discussed above applicable to the pediatric population, in addition to the concern of the effect of HAART regimens on growth and puberty. Also, it was noted that there is a real need to follow children, for life, who were born HIV-negative, but exposed to antiretroviral drugs during their development in utero to successfully prevent HIV transmission and treat their mother's HIV infection during pregnancy.
Conclusions and Summary.
This Workshop was a very useful process, and there clearly was general consensus that the answers to these questions are very important. I was somewhat disappointed that some of the more difficult issues were not addressed in more detail. These issues include the questions concerning the specific design of such trials, the number of people needed to answer the questions, the likely drop-out rates, the needed duration of these trials, the likelihood that enough clinical endpoints would occur to give a meaningful answer, and the probability that the answer obtained would still be relevant. The devil is truly in the details. I think in depth focus groups would be needed to determine if both providers and potential participants would actually sign on the dotted line and enroll in these trials.
Additionally, the suggestion was made that a pilot trial might need to be conducted to determine feasibility. I also continue to have strong reservations that simple trials will give answers that will help guide individual people and providers in deciding when to start and when to switch. Lastly, I also have serious concerns about how truly informed consent would be obtained. Will this be a trial that supporting community representatives sign up themselves to participate in?
At the conclusion of the Workshop, Dr. Jack Killen noted that, to conduct these trials, it may be necessary to think across and beyond existing cooperative HIV clinical trials groups, and that these trials need to be conducted in primary care settings. Also, these trials would be complex to conduct, but further investigation is worthwhile to purse both for HIV therapy, as well as developing the methodology that would be applicable to medical research in general. The next steps will be to assess plans currently in planning stages, and move forward with due haste, but with care and diligence. Dr. Killen also noted that more investigation is needed to study feasibility, complexity and assuring that there is enough interest and willingness to participate in such trials by both providers and potential trial participants.