Staccato study suggests that some CD4-guided treatment interruptions can be safe

The results of the Staccato study, published in the August 5^th edition of The Lancet, have demonstrated that CD4-guided treatment interruptions may be safe under some circumstances. In contrast to the SMART study, which was stopped early at the beginning of the year due to safety concerns, Staccato found that interrupting treatment reduced the incidence of side-effects, as well as lowering drug costs without increasing the rates of resistance.

Treatment interruptions guided by a patient’s CD4 cell count were proposed as a method to reduce exposure to antiretroviral drugs and reduce their toxicity and cost in HIV-positive patients. In contrast to interruption schedules based on fixed times on and off treatment, basing decisions on when to stop and start treatment based on a patient’s CD4 cell count were believed to avoid the risks of disease progression.

The optimism surrounding this approach was cut short by the discontinuation of the SMART study in January 2006. This large, randomised trial was comparing the effects of continuous HIV treatment to a CD4-guided strategy, finding that the patients interrupting treatment not only experienced faster disease progression and higher death rates, but also greater rates of ‘drug-related’ toxicities, such as heart attacks, strokes, liver disease and kidney disease.

Now, however, the Staccato trial has found that patients randomised to interrupt HIV treatment until their CD4 cell count fell to below 350 cells/mm³ had lower rates of treatment-related side-effects than patients taking continuous therapy, although they were more likely to suffer from thrush.

There was also no significant increase in resistance rates between the two groups, and both groups responded well to twelve weeks of continuous treatment at the end of the study.

Although it was not large enough to detect a difference in disease progression or death rates between the groups, the investigators note that treatment interruption in the Staccato study resulted in fewer deaths and instances of disease progression than seen in the SMART study. While they cannot offer a definitive explanation for the difference in the two studies’ findings, they suggest that it could be related to higher CD4 cell counts in Staccato, or possibly the lower levels of treatment experience in the trial’s participants.

Staccato vs. SMART

Although Staccato was too small to detect an effect of treatment interruption on AIDS or death rates, the investigators calculate that they would have seen around 17 AIDS-defining opportunistic infections in the treatment interruption group of Staccato if the risk of disease progression was similar in the two studies.

“No AIDS-defining event and only one death were actually observed,” the investigators write. “Such a striking discrepancy is unlikely to be due to chance and requires explanation.”

The researchers point out that most AIDS-defining events in the SMART study occurred at low CD4 cell counts. Since the Staccato study involved re-starting treatment at a CD4 cell count below 350 cells/mm³, as opposed to 250 cells/mm³ in SMART, it is possible that the maintenance of higher CD4 cell counts in Staccato contributed to the lower rates of illness.

In addition, the patients in Staccato had taken antiretroviral therapy for only 16 months on average before the trial began, in contrast to 72 months in SMART. However, “the relation between those differences and the apparently different outcome, is not obvious,” the investigators write.

The investigators also point out that the actual increased risk found in the SMART study due to treatment interruptions was low, amounting to a maximum of three events avoided per 100 patient-years of treatment. They have found that more than 60% of the Swiss participants considered this risk to be low enough that they chose to continue in the treatment interruption arm of Staccato.

“For these patients, Staccato’s results provide reassurance about the one risk that was widely feared when these trials began – development of resistance and loss of efficacy of treatment,” they conclude. “Staccato indicates that ritonavir-boosted protease-inhibitor-based HAART can be interrupted without undue harm, provided that CD4 counts are monitored.”

The Staccato study

The Staccato study included 430 patients, who had had CD4 cell counts above 350 cells/mm³ and viral loads below 50 copies/ml for at least three months. Eighty per cent of the study’s participants were Thai, all of whom took ritonavir (Norvir)-boosted saquinavir (Invirase) along with two nucleoside reverse transcriptase inhibitors (NRTIs). The remaining patients came from Switzerland (18%) and Australia (2%).

Two hundred and eighty-four of the patients were randomised to the treatment interruption arm, stopping treatment until their CD4 cell count fell to below 350 cells/mm³, when they started to take their anti-HIV combination again for at least twelve weeks. They stopped again if the CD4 cell count rose above 350 cells/mm³.

The remaining 146 patients took antiretroviral therapy continuously.

Over a median follow-up of 21.9 months, the investigators calculated that the treatment interruption strategy resulted in a drug saving of 62%, potentially reducing the cost of treatment. Half of the patients in the treatment interruption arm had re-started therapy within 18 weeks: patients with lower CD4 cell counts before starting antiretroviral treatment or at the start of the study, and those with higher viral loads before starting treatment had to re-start anti-HIV treatment earlier.

There were no AIDS-defining events during the trial, but the patients receiving continuous treatment had higher rates of diarrhoea (23 vs. 16%, p = 0.04) and neuropathy (5 vs. 2%, p = 0.03) than the interruption group. They also had higher levels of cholesterol in the blood and were more likely to report that they had lipodystrophy.

In contrast, candidiasis (thrush) was more common in the patients taking treatment intermittently, possibly due to their lower CD4 cell counts: at the end of follow-up CD4 cell counts were significantly lower in the interruption group (402 vs. 619 cells/mm³, p = 0.001).

There were two deaths during the trial – one in each group – but these were unrelated to HIV or to the drugs used in the study.

Resistance mutations in the genes for protease and reverse transcriptase were found in ten (2%) of the patients, at similar rates in the two groups (2.5 vs. 2.1%, p = 0.70), in contrast to results of previous studies. There was also a low rate of illness due to viral load increases in the treatment interruption group, characterised by fever, sore throat and skin lesions: only 17 (6%) of the patients in this group had these symptoms, of which only 6 (3%) had viral loads above 100,000 copies/ml.

At the end of the randomised portion of the study, all of the patients received twelve weeks of treatment. Similar proportions of both groups achieved viral loads below 50 copies/ml at the end of this period (91 vs. 92%, p = 0.90), suggesting that intermittent treatment did not impair the patients’ ability to suppress HIV.