Fixed-length interruptions

Many early treatment interruption studies looked at intermittent therapy using fixed-length cycles ranging from days to months. Enabling patients to stop treatment periodically on a fixed schedule may relieve ‘treatment fatigue’, lessen long-term exposure to drug toxicities and reduce cost.

The open-label FOTO study assessed one of the shortest proposed fixed-length treatment cycles – five days on followed by two days off – designed to let people take a treatment break at the weekend. A total of 30 participants with undetectable viral load were enrolled, with ten each receiving regimens containing efavirenz, nevirapine or a protease inhibitor. After 48 weeks, all ten patients taking efavirenz still had a viral load below 50 copies/ml, but one in the nevirapine group and two in the protease inhibitor group experienced viral rebound. Blood fat levels improved slightly. Participants reported that they strongly preferred to take the weekend off, and drug costs dropped by nearly a third.¹

A few studies have explored one week on, one week off (‘WOWO’) intermittent treatment schedules, based on data showing that viral load typically takes at least a week to rise above 500 copies/ml after stopping therapy. Overall, results have not been impressive. The Staccato study (described below) originally included a one week on, one week off fixed-cycle interruption arm in addition to the continuous therapy and CD4-guided arms, but this was halted in 2003 after 53% of patients experienced viral rebound. This occurred in three of four patients taking a triple NRTI-only regimen and two-thirds of those taking an unboosted protease inhibitor, but only one taking efavirenz.²

A study looking at a longer cycle – two months on, one month off – was conducted by the US National Institutes of Health. In this trial, 56 participants with full viral suppression were randomly assigned to continuous or intermittent therapy. The study was stopped after five patients in the treatment interruption arm developed drug resistance mutations and were unable to re-suppress HIV after recommencing therapy. Amongst the participants who had completed 48 weeks of follow-up, intermittent therapy did not suppress viral load, boost immune function or reduce side-effects (though blood fat levels dipped temporarily).³

In the French ANRS 106 ‘Window’ study, 403 participants with well-controlled HIV and a CD4 count above 450 cells/mm³ were randomised to continue their current therapy or begin intermittent treatment following a two months on, two months off schedule. About half received efavirenz and half a protease inhibitor. At 96 weeks, after six interruption cycles, 81% in the intermittent therapy arm and 90% in the continuous group had a viral load below 400 copies/ml. Twice as many in the intermittent group saw their CD4 count fall below 300 cells/mm³ (4 vs 2%), and the median decrease in CD4 cell count was significantly larger (155 vs 8 cells/mm³). Minor HIV-related events such as thrush and thrombocytopenia were more common in the intermittent group, but no AIDS-defining events occurred.⁴

The Italian ISS PART study enrolled 273 people with viral suppression and very high CD4 cell count above 700 cells/mm³. They were randomised to receive continuous therapy or a complex schedule of five fixed-length interruptions lasting one, two or three months, interspersed with three-month periods on therapy. After two years, more than 90% in both arms had a viral load below 400 copies/ml. But whilst 86% of those on continuous therapy still had a CD4 count greater than 500 cells/mm³, this fell to 69% in the treatment interruption arm. Although there were no cases of major illness or death, the drop-out rate in the interruption arm was high, emergence of resistance was common and there were no apparent benefits.⁵

Finally, the ongoing DART (Development of Anti-Retroviral Therapy in Africa) trial enrolled 3300 participants in Uganda and Zimbabwe, of whom about 800 took part in a treatment interruption substudy. All enrolled with a CD4 count below 200 cells/mm³ and their CD4 counts rose above 300 cells/mm³ after taking antiretroviral therapy for 12 to 18 weeks. At that time, they were randomly assigned to stay on continuous therapy or to begin three months on, three months off treatment interruption cycles.⁶

After 51 weeks, those in the intermittent arm were more than twice as likely (8 vs 3%) to experience HIV-related illnesses, including oesophageal thrush and extrapulmonary tuberculosis. The treatment interruption arm was halted in March 2006 after a safety monitoring board determined that the risk of disease progression was unacceptable. This study – one of the first to explore fixed-length treatment interruption amongst a large population with advanced HIV disease in a resource-limited setting – showed that this form of intermittent treatment cannot be recommended even though cost and drug savings would be substantial.

Another large study in Africa, Trivacan (described below), is still ongoing to compare continuous and intermittent treatment using a four months on, two months off schedule, although the original CD4-guided interruption arm was stopped due to a higher risk of illness.