The Journal of Infectious Diseases APRIL 15, 2008

Inferior clinical outcome of the CD4+ cell count-guided antiretroviral treatment interruption strategy in the SMART study

The Strategies for Management of Antiretroviral Therapy (SMART) Study Group

Bottom Line

The SMART trial demonstrated that a CD4+ cell count-guided episodic antiretroviral therapy strategy significantly increased the risk of opportunistic disease or death compared to continuous therapy.

Key Findings

1. The drug conservation (DC) arm experienced a higher rate of opportunistic disease or death (3.4 per 100 person-years) compared to the viral suppression (VS) arm (1.3 per 100 person-years).
2. The hazard ratio for the primary endpoint (opportunistic disease or death) in the DC group versus the VS group was 2.6 (95% CI, 1.9–3.7; P<0.001).
3. Patients in the DC group had an increased risk of serious non-AIDS events, with hazard ratios of 1.8 (95% CI, 1.2–2.9) for death from any cause, and 1.7 (95% CI, 1.1–2.5) for major cardiovascular, renal, or hepatic disease.
4. The excess risk in the DC arm was strongly associated with more time spent with lower CD4+ counts and uncontrolled HIV viremia during the interruption periods.

Study Design

Design
RCT
Open-Label
Sample
5,472
Patients
Duration
16 mo
Median
Setting
Multicenter, International
Population HIV-infected adults with CD4+ cell counts >350 cells/mm3
Intervention Episodic antiretroviral therapy (drug conservation) deferred until CD4+ <250 cells/mm3 and restarted until CD4+ >350 cells/mm3
Comparator Continuous antiretroviral therapy (viral suppression) to maintain suppressed viral load
Outcome Composite of opportunistic disease or death from any cause

Study Limitations

Participants were primarily treatment-experienced, limiting the generalizability of these findings to ART-naive patients.
The study was prematurely halted, which may have limited the long-term assessment of secondary outcomes.
The strategy utilized episodic interruptions that may not reflect modern clinical practices for structured treatment interruptions.

Clinical Significance

This study provided definitive evidence against the routine use of CD4+ cell count-guided treatment interruptions as a drug-sparing strategy, reinforcing the necessity of continuous viral suppression to prevent both AIDS-defining and serious non-AIDS-related morbidity and mortality.

Historical Context

During the early 2000s, clinical interest in 'drug holidays' or structured treatment interruptions (STIs) was high, driven by the desire to reduce long-term toxicities, treatment fatigue, and costs associated with lifelong combination antiretroviral therapy (ART).

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

The SMART trial results were surprising because the 'drug conservation' group experienced higher rates of disease even at relatively high CD4+ counts. Pathophysiologically, why does viral rebound during treatment interruption cause systemic damage before the CD4+ count drops below the traditional 200 cells/µL threshold?

Key Response

When antiretroviral therapy (ART) is stopped, HIV replication resumes rapidly. This surge in viremia triggers intense systemic inflammation and immune activation (evidenced by biomarkers like IL-6 and D-dimer). This inflammatory state causes 'bystander' damage to the vascular endothelium and various organ systems, leading to non-AIDS-defining events like cardiovascular disease, even while the T-cell count remains objectively high.

Resident
Resident

A patient with well-controlled HIV (CD4+ count 650) requests a 'drug holiday' because they feel healthy and are tired of daily pills. Based on the SMART trial findings, how would you quantify their risk of opportunistic disease or death compared to staying on continuous therapy?

Key Response

The SMART trial showed that episodic ART (stopping until CD4+ <250 and restarting until >350) resulted in a 2.6-fold increased risk of opportunistic disease or death compared to continuous ART. Residents must communicate that 'feeling healthy' does not mitigate the increased risk of clinical progression and serious non-AIDS events that occur during treatment interruptions.

Fellow
Fellow

The SMART trial was a landmark study that shifted the HIV treatment paradigm. How did the discovery of increased 'non-AIDS' events (cardiac, renal, and hepatic) in the interruption arm change the subspecialty's understanding of ART toxicity versus HIV-induced inflammation?

Key Response

Prior to SMART, it was hypothesized that 'drug holidays' would reduce the risk of long-term ART toxicity (e.g., metabolic syndrome, lipodystrophy). SMART proved the opposite: the inflammatory consequences of uncontrolled viral replication are far more detrimental to organ systems than the toxicities of the (then-current) ART regimens, effectively ending the clinical practice of structured treatment interruptions (STIs).

Attending
Attending

Given the historical context of the SMART study, how should these findings be used to counsel 'elite controllers' or 'long-term non-progressors' regarding the initiation of ART, despite their naturally high CD4+ counts?

Key Response

The SMART study demonstrated that viremia itself is a driver of morbidity. While elite controllers maintain high CD4+ counts, many still have low-level detectable viremia and associated chronic inflammation. Attending physicians should use the SMART framework—that suppressing replication is protective against multi-organ inflammatory damage—to justify the 'Treat All' strategy, which has been further reinforced by the START trial.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The SMART study used a composite primary endpoint of 'serious opportunistic disease or death from any cause.' What are the methodological strengths and limitations of using such a composite endpoint in a trial that was ultimately terminated early by a Data and Safety Monitoring Board (DSMB)?

Key Response

Strengths include increased statistical power to detect treatment effects in populations with low individual event rates. However, early termination (as happened with SMART) can lead to an overestimation of the treatment effect (truncation bias). Furthermore, if the treatment effect is not uniform across all components of the composite (e.g., if it affects death more than opportunistic infection), the overall Hazard Ratio may be difficult to generalize to specific disease processes.

Journal Editor
Journal Editor

As a peer reviewer, if you were evaluating a modern-day sub-analysis of the SMART trial, what concerns would you raise regarding the internal validity of the study's conclusions about ART toxicity, considering the specific antiretroviral agents (e.g., d4T, ddI, early PIs) used during the 2002-2006 enrollment period?

Key Response

An editor would flag that the 'toxicity' profile of 2006-era ART is vastly different from modern integrase-inhibitor-based regimens. Because modern drugs are significantly less toxic and more durable, the 'drug conservation' rationale used to justify the SMART trial design is even less applicable today. Findings regarding metabolic complications in the continuous arm must be contextualized within the limitations of the older drug classes used at that time.

Guideline Committee
Guideline Committee

The SMART trial is cited in the DHHS HIV Guidelines as a primary reason for the recommendation against structured treatment interruptions. How does the level of evidence provided by SMART influence the 'Strength of Recommendation' for continuous ART, and does it allow for any clinical exceptions?

Key Response

The SMART trial provides Level I, Grade A evidence against treatment interruption. Current DHHS guidelines (and IAS-USA) state that ART should be continued permanently except in very specific, rare circumstances such as controlled research trials (e.g., searching for a functional cure). The findings were so definitive in their harm that any protocol suggesting an 'off-ART' period for clinical management would be considered a breach of the current standard of care.

Clinical Landscape

Noteworthy Related Trials

2003

FIRST Trial

n = 1416 · NEJM

Tested

Initial therapy with efavirenz plus two NRTIs

Population

Treatment-naive HIV-infected patients

Comparator

Initial therapy with protease inhibitor-based regimens

Endpoint

Time to treatment failure

Key result: Efavirenz-based regimens were found to be more durable and better tolerated than protease inhibitor-based regimens.
2006

DART Trial

n = 3316 · Lancet

Tested

CD4-guided structured treatment interruptions

Population

Previously untreated HIV-infected adults in Africa

Comparator

Continuous antiretroviral therapy

Endpoint

Occurrence of new WHO stage 4 events or death

Key result: CD4-guided interruption led to higher rates of disease progression and death compared to continuous therapy.
2006

STACCATO Trial

n = 447 · Lancet

Tested

CD4-guided intermittent therapy

Population

Patients with HIV with CD4 counts above 350 cells/µL

Comparator

Continuous antiretroviral therapy

Endpoint

CD4 cell count decline and viral load suppression

Key result: Intermittent therapy was associated with a higher incidence of WHO stage 2-4 events compared to continuous therapy.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis