SMART: background, concepts, and design.

Sharon is a 28-year-old African American whose HIV infection was diagnosed 9 months ago. Of 3 determinations, her average CD4 T cell count has been 489 cells/[mm.sup.3]; average viral load, 17,500 copies/mL. Before undertaking a regimen of highly active antiretroviral therapy (HAART), Sharon looks for study results on the long-term clinical benefits and risks of anti-HIV drugs in patients with early and mid-stage disease. She finds none.

It is called SMART.

Now that the federal government has given it pilot phase approval, SMART will be the largest, most ambitious clinical trial in the history of the HIV epidemic. In fact, it will be one of the largest trials in the history of any infectious disease. The acronym, which stands for Strategies for the Management of Anti-Retroviral Therapy, suggests that the study is an attempt to answer intelligently important questions about the long-term management of HIV disease. During the 2year pilot, investigators must enroll 500 volunteers in the first 6 months and 1000 in the first year. The target date for enrollment to begin is October 15, 2001.

Designed by investigators with the Community Programs for Clinical Research on AIDS (CPCRA), SMART seeks to determine whether one of two sharply contrasting management strategies is superior to the other in delaying AIDS-related morbidity and mortality. Specifically, the study compares a strategy of delayed, episodic anti-HIV therapy against one of immediate, uninterrupted therapy.

The National Institutes of Health has called upon the research consortiums it funds, including CPCRA, to design trials that evaluate strategic approaches to the use of anti-HIV drugs. SMART marks the federal government's first attempt to study the long-term clinical effectiveness of different strategies for managing HIV disease.

The study calls for following 6000 HIV-infected subjects for as long as 8 years. By protocol, the volunteers must have a baseline CD4 T cell count of at least 350 cells/[mm.sup.3]; investigators will randomize them in a 1:1 ratio to one of two strategies.

1) The drug conservation (DC) strategy has participants stop or defer antiretroviral therapy until their CD4 T cell counts fall below 250 cells/[mm.sup.3]. Volunteers would then undertake a regimen to return their T cell counts to 350 cells/[mm.sup.3] (or higher) for 2 consecutive visits, at which point they would stop treatment again. Assuming each round of treatment is successful in restoring the T cell count to baseline, participants following this strategy could conceivably repeat the cycle until the study ends. While on therapy, the protocol calls for volunteers to take virologically suppressive regimens. In the patient literature, this arm will be known as the Wait Group.

The rationale for the DC strategy is based on the following observations:

* HIV disease progression is rare among patients with HIV CD4 T cell counts above 250 cells/[mm.sup.3].

* Among patients on therapy, current CD4 T cell count is a better measure of short-term risk for disease progression than viral load.

* Since clinical symptoms may precede opportunistic disease in patients with higher CD4 T cell counts, regular patient monitoring further reduces the risk of actual illness.

* Even when the CD4 T cell count has fallen to a low number, anti-HIV therapy restores immune function.

* Viral load is durably suppressed with therapy among patients with CD4 T cell counts between 200 and 349 cells/[mm.sup.3] just as well as it is among patients with CD4 T cell counts above 350 cells/[mm.sup.3].

* Half or more of all patients lose control of viral load within a year of starting therapy, and thus risk drug resistance.

* Long-term use of antiretrovirals is associated with serious side effects.

* Long-term adherence to HAART is difficult.

OR

2) The virologic suppression (VS) strategy has participants immediately undertake therapy following randomization with the goal of virologic suppression, regardless of CD4 T cell count. (This includes patients with high CD4 T cell counts above 500 cells/[mm.sup.3].) Patients naive to therapy at randomization must begin a virologically suppressive first-line regimen. Those on treatment prior to randomization may continue their regimens, modifying when necessary to suppress viral load. The protocol permits volunteers in this arm to change therapies for toxicity or intolerance, but the participants will stay on regimens aimed at achieving viral suppression for the duration of trial. In the patient literature, this arm will be known as the Go Group.

The rationale for the VS strategy is based on these observations:

* Baseline viral load predicts progression to AIDS in untreated patients.

* Since changes in therapy are more likely to produce viral suppression when made at lower viral loads, HIV RNA and not CD4 T cell count is the better guide to regimen modification.

* When treatment is interrupted, the CD4 T cell count declines and viral load rises.

* Durable suppression of viral load is more likely when patients start therapy with CD4 T cell counts of 350 cells/[mm.sup.3] or higher.

* There are enough antiretrovirals on the market for most patients to find a regimen that is tolerable and convenient.

* Even though HAART does facilitate partial immune reconstitution, the CD4 T cell nadir influences risk of opportunistic disease.

The protocol excludes volunteers who are pregnant or enrolled in any other study inconsistent with participation in SMART. Regardless of the arm to which they are assigned, SMART volunteers are free to take agents of their choice while on treatment, including experimental compounds, hydroxyurea, or immunomodulators. The only caveat is that, whenever possible, the regimens suppress HIV RNA below the limits of quantification.

Volunteers must have a baseline CD4 T cell count of 350 cells/[mm.sup.3] or higher and be willing to start, modify, delay, or stop therapy, depending on the study arm to which they are randomized. Accordingly, the trial is open to volunteers who are both treatment naive and treatment experienced. This broadly inclusive design is intentional. Investigators want SMART to be acceptable to patients and clinicians, and they want the study's findings to reflect actual clinical practice and be widely generalizable.

Investigators had intended the treatment strategy of SMART's VS group to replicate the US standard of care for the management of HIV infection. (Hence, the protocol refers to the VS arm as the control group.) For years, the guidelines promulgated by a panel of the US Department of Health and Human Services set a CD4 T cell count less than 500 cells/[mm.sup.3] as the trigger to start therapy, but did not discourage therapy for patients with even higher counts. Accordingly, SMART's VS approach calls for hard-hitting treatment for every patient, including those with CD4 T cell counts above 500 cells/[mm.sup.3].

But in February 2001, in a nod to the intractable toxicities of HAART, the guidelines panel released a new blueprint for treatment. Now, the threshold for initiating therapy is 350 CD4 T cells/[mm.sup.3], with a viral load below the limits of quantification still the appropriate aim of treatment.

SMART does not question the tactical wisdom of maximally suppressing HIV RNA once a patient undertakes a HAART regimen. But it does ask, given the serious risks associated with anti-HIV drugs, whether the new, lower start-to-treat trigger is still too high. Relying on data that show clinical disease progression occurs infrequently in patients with relatively healthy CD4 T cell counts, SMART will examine the implications of delaying treatment until the T cell count has fallen to 250 cells/[mm.sup.3]. Furthermore, by allowing for episodic therapy, SMART's DC arm also questions the assumption that once therapy is undertaken it should never be stopped. Finally, SMART asks which is the better surrogate for guiding treatment decisions: viral load or CD4 T cell count?

In short, SMART pits immediate and uninterrupted antiretroviral therapy with a reliance on viral load as the guiding surrogate against deferred and episodic therapy with a reliance on CD4 T cell count as the guide.

Since SMART puts the surrogates themselves on trial, it will use clinical events (AIDS and death) as primary endpoints. Follow-up will last until investigators observe 910 clinical events, which is expected to take 6-8 years. The study's authors project that clinical disease progression will account for 70% of these events, with death accounting for the remaining 30%; however, half of these deaths will be unrelated to HIV, e.g. suicide, car accidents, etc. As the protocol notes, "[t]he ultimate goal of the treatment of HIV disease is to prevent clinical morbidity and mortality. Thus, the primary and secondary endpoints of this study are clinical outcomes rather than surrogate markers."

Ironically, even though it employs sharply contrasting approaches to HIV disease management, SMART does not anticipate a dramatic difference in clinical outcome between the groups. Even a moderate difference in treatment effect--say 17%, which is what the protocol projects--would have a significant influence on patient management. Furthermore, the trial itself would provide the missing clinical evidence for treatment guidelines.

In addition to the primary endpoint of clinical disease progression or death, SMART has several secondary endpoints, including:

* incidence of major cardiovascular and metabolic complications

* other treatment-related toxicities

* adherence

Substudies will examine differences between the groups in quality of life, HIV transmission risk behaviors, health care utilization, etc. (See page 17.)

The clinical benefits of antiretroviral therapy are well documented for patients with advanced HIV disease. However, it is not yet known whether therapy increases survival time or decreases opportunistic infections among patients with early or mid-stage disease, and if it does, whether those benefits are outweighed by the toxicities of therapy.

Based largely on strong theoretical considerations, experts have recommended the early use of uninterrupted treatment with the aim of maximal viral suppression; however, like HIV infection itself, treatment comes with its own dangers and an undetectable viral load has proven elusive for at least half of all patients on therapy. Furthermore, the prospect of a lifetime of uninterrupted therapy is not only disheartening, it is also impractical.

SMART therefore asks whether the same clinical benefits sought by standard of care treatment, as reflected in the VS arm, can be achieved through a kinder, gentler approach, as reflected in the DC arm. The long-term usefulness of HAART itself may hinge on the answer.

[ILLUSTRATION OMITTED]