Reports

Two late-breaking trials have implications for long-term HIV management. In the multinational SMART (Strategies for Management of Anti-Retroviral Therapy) study, a CD4-guided intermittent treatment appeared inferior for HIV suppression when compared to a conventional continuous regimen, nor did it protect patients from many of the adverse events most closely associated with HAART. While trial results were disappointing, the data raise new questions about the perceived risks of HAART. In the other multinational study, unboosted and boosted atazanavir was associated with comparable safety and efficacy in treatment-naïve patients at the 48-week analysis of a 96-week trial. While data suggest that either strategy is viable, the comparison demonstrated the safety of the boosted protease inhibitor (PI) atazanavir while reaffirming the efficacy of its unboosted formulation in the treatment-naïve population.

Study Design

In the SMART study, 5472 patients at 318 participating treatment centres in 33 countries, including Canada, were randomized to receive a CD4-guided drug conservation (DC) management or to remain on treatment without interruption with a goal of continuous viral suppression (VS). All but 5% of patients were treatment-experienced at entry. The median baseline CD4+ cell count was 596 cells/mm3, and 71% of patients had a viral load <400 copies/mL. All patients had a CD4+ cell count >350 cells/mm3. The average number of years on HAART at baseline was six.

In the CD4-guided treatment arm, HAART was stopped at the time of entry and only reinitiated once CD4+ cells fell to <250 cells/mm3. Patients were then maintained on HAART until CD4+ cell count climbed >350 cells/mm3 when treatment was again discontinued. The goal was to repeat this pattern indefinitely, reducing drug exposure but preventing a decline in immune function to a level likely to produce AIDS-defining events. Patients in the VS arm received conventional continuous HAART therapy. The primary end point of the study was HIV disease progression or death.

“In early January, the data safety monitoring board of the SMART study determined that there was a significantly increased risk of HIV progression in the DC arm and enrolment was stopped,” reported Dr. Wafaa El-Sadr, Columbia University College of Physicians and Surgeons, New York, New York. She added that although the study was designed to follow patients for an average of eight years, the average follow-up was 14 months at the time the study was discontinued.

The long-awaited results of the SMART study were a surprise. Although every large study that has evaluated any type of structured treatment interruption has failed, the goal of reducing costs and avoiding adverse events has been a powerful motivation to evaluate new approaches. The SMART study was not only closed prematurely because of the greater rate of failure in the DC arm, subsequent analyses also suggested that patients receiving less antiretroviral therapy experienced more complications, including those previously attributed to antiretroviral therapy.

“The rate of myocardial infarction, stroke, coronary artery disease requiring surgery, renal and liver disease was 2.0 per 100 person-years in the DC episodic CD4-guided arm vs. 1.2 per 100 person-years in those who remained on continuous therapy,” Dr. El-Sadr reported.

At the time that the study was stopped, the primary end point of HIV disease progression or death was observed in 3.7 per 100 person-years in the DC arm vs. 1.5 per 100 person-years in the VS arm (P<0.0001). This produced a 2.15 relative increase in risk of the primary end point in the DC arm vs. the VS arm. When serious AIDS events were compared, the relative risk was 5.82 (95% CI, 1.68-20.16) for the DC vs. the VS arm. The rate of death was 1.5 per 100 patient-years in the DC arm vs. 0.9 per 100-patient years in the VS arm, producing a 1.62 relative increase in risk (P=0.02) for those randomized to the DC strategy.

Although a pilot study had predicted that the DC arm would provide similar efficacy with a reduction in drug costs and risks of complications, the SMART study demonstrated that complications were actually higher in the DC arm over the period of study, including complications involving the kidney, liver and cardiovascular system. When adverse events in these organs were evaluated separately, the VS arm was favoured consistently. When evaluated together, the relative risk of complications in any of these systems was 1.63 (95% CI, 1.02- 2.58; P=0.04) times greater in the DC group.

According to Dr. El-Sadr, who presented data showing an efficacy advantage for VS over DC for every subgroup evaluated, including those defined by viral load, baseline CD4+ cell count, nadir CD4+ cell count, age and gender, “Episodic use of HAART as defined in the SMART protocol is inferior to continuous HAART. This strategy should not be recommended.”

Boosting vs. No Boosting in Treatment-naïve Patients

In the other multinational study, presented directly after the SMART study, 200 treatment-naïve patients initiating HAART with the PI atazanavir were randomized in a 1:1 ratio to receive the PI boosted or unboosted. All patients received lamivudine and extended-release stavudine. When unboosted, atazanavir was administered in a 400 mg q.d. dose; boosted, a 300 mg q.d. dose was combined with 100 mg ritonavir. The median viral load at entry was 5.0 log10 RNA copies/mL, while the median CD4+ cell count was 200 cells/mm3. The average age was 35. About 30% of the study population was female and about half were Caucasian. The primary end point of the study was the proportion achieving a sustained suppression of viral load <400 copies/mL.

On an intent-to-treat analysis, 86% of the boosted and 85% of the unboosted patients were suppressed <400 copies/mL at 48 weeks. The secondary end point of <50 copies/mL also showed comparable suppression (75% and 70% in the two groups, respectively). In the on-treatment analyses, 93% of both groups had a viral load <400 copies/mL at the end of 48 weeks. While the proportion with a viral load <50 copies/mL in the on-treatment analyses was higher in the boosted group (87% vs. 76%), the difference was not statistically significant. The failure rates of 3% and 10% for the boosted and unboosted arms, respectively, were also not significantly different. The median increases in CD4+ cell counts over the 48 weeks were 189 cells/mm3 in the boosted group and 224 cells/mm3 in the unboosted group, which were not different statistically.

Both regimens were well tolerated. While there was a greater number of discontinuations for adverse events in the boosted arm (8% vs. 1%), the difference was largely explained by protocol-mandated discontinuations for hyperbilirubinemia, none of which were associated with clinically significant complications. Similarly, the higher number of grade 2 to 4 adverse events (43% vs. 34%) in the boosted group was largely accounted for by hyperbilirubinemia (13% vs. 5%) and jaundice (3% vs. <1%). No other adverse event differed markedly between the two groups. The mean increase in total cholesterol was 15% for the boosted arm vs. 6% for the unboosted arm (P<0.01) but <5% of either arm received lipid-lowering agents.

According to Dr. Niel Malan, Triple M Research, Port Elizabeth, South Africa, “Up until now, the clinical studies in treatment-naïve patients have been largely conducted with unboosted atazanavir. Boosted atazanavir has been primarily evaluated in treatment-experienced populations. These results demonstrate that the two strategies are comparable in treatment-naïve patients.” While the unboosted PI has already demonstrated a high degree of safety and efficacy in treatment-naïve patients, the current data “support use of boosted atazanavir in treatment-naïve patients” based on both efficacy and safety, he concluded.

Pilot Study Results

A third late-breaking study presented directly after the first two appeared to provide an intriguing link. Although only a pilot study with 36 patients, it evaluated whether a potent boosted PI could be used as a single agent to simplify therapy in patients who had previously achieved and maintained good control on a full HAART regimen. Such a strategy has the promise of reducing costs and drug exposure, much like the strategy pursued in the SMART study.

According to lead author Dr. Susan Swindells, Associate Professor of Internal Medicine (Infectious Diseases), University of Nebraska Medical Center, Omaha, the results have been favourable. Conducted with boosted atazanavir, all but three patients remained suppressed for the 24 weeks of analysis.

She noted, “In the three failures, blood samples at the time of failure in two of the three patients did not reveal any detectable drug level. None of the failures had any PI mutations at the time of failure,” suggesting that participants were either non-compliant or rapid metabolizers of the study agent. Overall, the results were characterized as promising. Dr. Swindells called for larger trials to further evaluate the potential of this strategy.

Summary

Late-breaking trials evaluating a variety of approaches to long-term HIV control have provided mixed results. While CD4-guided interruptions were not effective, a study demonstrated equivalence for a boosted vs. non-boosted PI in treatment-naïve patients on HAART. A small but promising pilot study has suggested that simplification of therapy might be an option once patients are well controlled, a possibility that will be pursued in larger studies.