A small group of community advocates, a divided research community, huge sums of money, and urgent questions that threaten to break the mold of conventional trial designs. All of these elements are at play in what may be one of the least publicized and most critical fields in HIV research today: long-term trial design. The National Institute of Allergy and Infectious Diseases (NIAID) has earmarked approximately $80 million dollars for studies to answer such crucial questions as when to start and when to change HIV therapy. Some trials, particularly large, international collaborations, could cost even more.

Taking on long-term management questions has, once again, brought HIV to the frontier of existing medical expertise. In January 2000, a NIAID Department of AIDS (DAIDS) meeting brought together researchers, clinicians, community advocates, statisticians and experts from cardiovascular and cancer research to discuss how best to answer long-term questions in the field of HIV disease management. The results were stimulating and daunting. "The word we got from people in cancer and heart disease was that this is at the cutting edge of clinical trials methodology in some ways," said DAIDS director Jack Killen. "There has been very little work in any field of clinical research that looks at very long-term outcomes over a very long period of time."

One of the major challenges is the sheer size of the trials. A study pool of many thousands of people is required to detect a meaningful reduction in the risk of "events" - i.e. death, severe toxicities or disease progression - associated with long-term use of particular treatment regimen or strategy. HIV treatment strategy trials of all kinds will, by necessity, enroll relatively healthy patients. It is medically and ethically indefensible not to treat those who are already sick or have CD4 counts below 200 cells/mm³. Starting with healthy individuals means it could be years before the effect of a strategy emerges. Large trials are a partial solution to this problem. The more participants there are, the more likely it is that a significant number of events will take place during the study period.

NIAID's largest, most expensive study to date, the ESPRIT trial, is enrolling 4,000 people for five years to test the effects IL-2, an immune stimulator. That trial is costing $43 million, including the establishment of a new international network of sites that could be used for other long-term strategy trials, too. At the January meeting, statisticians estimated that a when-to-start trial could require as many as 15,000 people and last a decade. Such a trial could cost over a billion dollars, according to one projection from what large-scale cardiovascular trials cost. Not all trials will have to be so big or unwieldy. But even a 6,000 person trial - the target enrollment for a "when-to-change" trial proposed by the CPCRA (see below) - handily outstrips ESPRIT in size. The network estimates that this trial will cost roughly $29 million.

Concerns about cost and feasibility have led some activists to question whether these trials should be conducted at all. They point out that big-budget "blockbuster" trials will siphon money, participants and investigators away from other research projects. Any one of the trials will also bring the networks into uncharted waters in terms of enrollment targets, accrual rates, length and, potentially, large-scale international collaborations.

Can't Get Started?

Much of the controversy concerns when-to-start trials (WTSTs). A WTST would determine whether immediate treatment, regardless of CD4 count, improves the odds of disease-free survival over waiting to start therapy until CD4 counts dip below a certain threshold. A CD4 count of 250 cells/mm³ is one widely accepted cutoff point. WTST advocates say that a large-scale, randomized controlled trial is the only way to gather data that could reduce prescription costs and boost quality of life for countless HIV-infected individuals. But even the fiercest proponents of WTSTs admit that there are significant hurdles to designing and enrolling such trials. "Everyone agrees in principle on what the need is," said Killen. "But there is a huge amount of disagreement about whether studies can be done."

One of the chief obstacles for WTSTs is persuading patients and physicians to cede highly personal choices about medication to random assignment in a trial. Current research suggests that there may be real benefit to beginning treatment during acute infection, but it is not clear whether there is any additional benefit to starting drugs at CD4 counts above 250 cells/mm³ during established infection. Short-term side effects and long-term toxicities of the drugs may prove detrimental to quality of life and survival time. As with cancer therapy, each doctor-patient pair must consider medical facts, open research questions, untested treatments, and personal factors like age, income, children and likelihood of successful adherence.

For this reason, a WTST could be difficult to enroll. Many physicians and patients think that when to start is an important question to answer - many fewer are willing to put it to the test themselves. In 1992, San Francisco researcher Don Abrams initiated COMPACT, an early WTST, and is skeptical about the prospects for enrolling such a trial today. "The problem is there isn't a huge amount of uncertainty in the patients and providers," says Abrams. "They are not willing to leave it to the flip of a coin." The Forum for Collaborative HIV Research is currently conducting feasibility studies. The Treatment Action Group (TAG) has also begun its own survey. Bill Duncan, Associate Director of Therapeutics Research at DAIDS, says that these studies will provide key guidance about how to proceed.

There are other issues, too. In order for a trial to take place, there has to be a clear, unanswered question. This elusive state of having no a priori opinion, or equipoise, is the foundation of clinical trial design. When it comes to WTSTs and equipoise, critics say it's there - to a point. Cumulative data from observational cohort studies and other trials strongly suggest that it is safe to wait until falling below a CD4 count of 350 cells/mm³ to begin treatment. British treatment guidelines use this level as the threshold for starting therapy; whereas United States guidelines set the bar at 500 CD4 cells/mm³.

Andrew Phillips, a British investigator, has presented retrospective analysis of data from the large EUROSIDA observational cohort. Phillips found that all treatment-naïve patients in the cohort had an equal likelihood of achieving maximal viral suppression as long as their pretreatment CD4 cell counts are above 200 cells/mm³ and their viral loads below 100,000 copies/mL. Since a CD4 count of 200 cells/mm³ may be the lowest acceptable limit for starting therapy, most WTSTs would likely set the threshold at 250 cells/mm³. This raises the question of whether significant time, energy and money should be invested in paring down a zone of equipoise that, by some analyses, spans just 100 CD4 cells cells/mm³.

A CD4-cell count drop of 100 cells/mm³ can take several years, and this is both a blessing and a curse. On the one hand, the wait boosts the relevance and importance of the WTST question. The interval lets people avoid the toxicities of HAART (highly active antiretroviral therapy) in the hope that less toxic drugs will be approved while they remain off treatment. By the same token, the advent of new drugs could sway individuals in the deferred arm to leave the trial and start treatment. For those who wait until the predetermined time to start therapy, improvements in drug efficacy and tolerability could obscure the effects of waiting.

Trials and Test Cases

These issues and more came up in the debate over protocol PR306, an ACTG (NIAID's AIDS Clinical Trials Group) proposal that has been abandoned but is so far the only WTST to receive serious consideration. One version of the discussion draft of the protocol called for five-year follow-up of 2,400 individuals, who would be randomized into immediate and deferred treatment arms. Lower thresholds for initiating treatment were not finalized. A CD4 count of 300 cells/mm³ was one proposed point for commencing treatment. Viral load over 100,000 copies/mL was also considered as a signal to start. Death, AIDS-defining illness, CD4 count below 200 cells/mm³ or Grade 4 drug toxicity were proposed primary endpoints. Possible secondary endpoints included undetectable viral load and complications of treatment.

In addition to the general criticisms of WTSTs, some viewed PR306 as not powerful enough statistically to answer its question. Spread out over five years, even a low rate of attrition could make it impossible to draw meaningful conclusions. Another concern was that the five years might not be long enough to measure a significant difference between the two arms. At a December, 2000 ACTG meeting, a clear majority of attendees voted against proceeding with the trial. Meeting attendees emphasized the importance of pursuing other options and approaches, including smaller, pilot studies or a trial that enrolls individuals with particularly confounding lab results, such as both high viral load and high CD4 cell count.

This no-confidence vote means that it may be some time before a U.S.-led WTST is launched. Reportedly, there have been no investigator-initiated grant proposals for WTSTs, and NIAID has not yet decided whether it will put out an RFA (request for applications). This step was supposedly part of NIAID's original plan for proceeding with long-term strategy trials.

In the meantime, other large-scale trial proposals are moving forward. NIAID is assembling review panels to consider two other large-scale trial proposals submitted by the CPCRA (NIAID's Community Program for Clinical Research on AIDS) and the VA (Veterans Affairs) network. The CPCRA SMART trial will ask whether there is a benefit to being on treatment with CD4 cell counts above 250 cells/mm³. SMART will enroll treatment-naive and treatment-experienced individuals. Those with over 250 CD4 T-cells will stop or defer treatment until their counts fall below this threshold. Individuals will stay on treatment until CD4 counts are above 350 cells/mm³ on two consecutive visits and then will go off therapy again.

The VA proposal asks a management question about "loose versus tight" viral control. In one arm of the study, antiretroviral (ARV) therapeutic changes will be minimized unless viral load rises above a specific ceiling. The other arm will apply more stringent criteria, changing ARV in response to any evidence of viral replication above the lowest level of quantification. Target enrollment for this five-year study is 4,200 participants.

These proposed trials could provide guidance for long-term treatment management questions: How much viral control is needed to continue seeing the clinical benefits of a particular regimen? Must treatment be taken continuously to be effective, or can it be used as a time-limited, strategic intervention when CD4 cell counts begin to fall? What types of resistance will emerge and when? What strategies maximize the efficacy of a given regimen? For the trials to enroll and retain target numbers of participants, these questions must remain relevant for the duration of the study. As much as possible, the study design must also be flexible enough to weather sea changes in treatment choices. Limiting participants' drug options to those approved at the start date of the trial is unrealistic and could result in high dropout rates if participants are forced to leave the study in order to access a new, promising therapy.

Losing track of participants is one of the major hazards of long-term trials. For both WTSTs and strategy trials, planners must anticipate a situation in which new treatments and findings, or interim data from the study itself cause a large number of participants to leave the trial or deviate from protocol. It is impossible to anticipate all the shifts and developments in HIV research that could occur over a five or ten year period. Still, planners can make clear decisions from the outset about how interim data will be analyzed and presented to the public.

For long-term management trials to be feasible for and of use to doctors and patients, they must also reflect, as much as possible, the real world conditions of treatment use. Short-term data from clinical and observational trials of HAART show that adherence and success rates of regimens are higher in clinical trials. Those involved in the long-term efforts stress that the highly controlled model will not provide the desired answers. "We need to make trials less restrictive, more open to a broader range of people, and to do them under, as close as we can, real world conditions. It's moving away from drug development-type trials," said Jim Neaton, a CPCRA biostatistician and principal investigator of the ESPRIT IL-2 trial.

Growing Pains

Feasibility of long-term trials depends as much on network capacity as it does on the willingness of individual participants to enroll. Here too, there are open questions. The review panel assigned to SMART will have the additional task of evaluating the overall performance of the CPCRA network. The network, which has eliminated a number of low-performing trial sites in recent years, has yet to be re-funded by NIAID. It is currently receiving "bridge funding" for one year to cover its current efforts but will not be able to launch new studies until its network funding is renewed. Meanwhile, the VA's "captive" patient population should minimize loss to follow-up in long-term studies.

In addition to specific structural issues, both the ACTG and the CPCRA will have to deal with external factors, including managed care and financial incentives, which may lead investigators to choose for to participate in industry-sponsored trials. "As more and more patients go into HMOs, they may be less accessible to research," said Neaton. "In a way, health-care settings in Canada and Europe are more conducive to these trials."

For a WTST, U.S. networks will almost certainly have to team up with international partners to reach enrollment targets. Although the ACTG has partnered successfully with some overseas groups, treatment advocates point out that it does not have a strong background in this type of multisite, international effort. One possibility is partnering with developing countries such as South Africa, Brazil or Thailand. A WTST would be much easier to do in these countries, which have both an adequate health-care infrastructure and large untreated populations with HIV. To do this, U.S. sponsors would have to navigate an ethical minefield, including issues of informed consent, relevance of the question to the study population, and availability of the drugs after completion of the trials.

Planners are also turning their attention to observational cohorts studies (OCS), which present far fewer obstacles to data collection. "We need to pursue randomized controlled trials," said Bill Duncan. "We also need to enhance existing cohorts. We need to pull them together and ensure that we are getting enough information, even with bias." OCS will be critical even after successful strategy trials are under way, said Alvaro Munoz, chief biostatistician for MACS (Multi-city AIDS Cohort Study). "Cohort studies are very well equipped to describe the treated history of individuals. They provide information on population effectiveness [of interventions] that these trials cannot provide."

As the debate continues, advocates and researchers agree that more education is needed to gain long-term commitment and public support for these trials. Particularly for a WTST, many U.S. communities will have to be swayed from the belief that it is unsafe to wait to treat, regardless of CD4 cell count. OCS data can help here, as can broader discussions about the lessons learned over the past four years of widespread use of HAART.

Four years of experience with HAART has proved insufficient to predict the long-term effects of these drugs on the heart, liver and kidney - or on HIV itself, which these regimens cannot completely suppress. As difficult as it is to conceive exactly how make long-term trials will, it is even more difficult to imagine patients and physicians making a lifetime of treatment decisions based on educated guesses. "I'm hoping that what's done here is first in a generation of management trials that will incrementally improve how we take care of patients with HIV," said Jim Neaton.

000110
AM010106

Copyright © 2001 by the American Foundation for AIDS Research (amfAR) and first displayed on amfAR's Treatment Directory web site (http://www.amfar.org/gl). They appear on AEGIS with amfAR's permission. Organizations wishing to reprint or redistribute these materials should request authorization from amfAR's Department of Treatment Information Services (212/806-1600).

AEGiS is made possible through unrestricted grants from Boehringer Ingelheim, iMetrikus, Inc., the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2001. This material is designed to support, not replace, the relationship that exists between you and your doctor.

AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.

Copyright ©1990, 2001. AEGiS & the Sisters of Saint Elizabeth of Hungary. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS and the Sisters of Saint. Elizabeth of Hungary, or the party credited as the provider of the content.