Treatment interruptions: reassuring data

Compelling evidence that structured treatment interruptions may be relatively safe even in patients with fairly advanced HIV disease was presented on Monday at the 40th ICAAC in Toronto.

Dr Michael Saag of the University of Alabama at Birmingham presented a review of outcomes in 78 patients identified as having stopped treatment for thirty days or more since January 1996, and who subsequently resumed treatment.

When they stopped treatment, individuals had a mean viral load of 8,953 copies and a mean CD4 count of 230, but these were in the main patients with relatively advanced disease as measured by their 'worst ever' scores on viral load and CD4 counts – an average viral load peak of 136,000 copies and a CD4 cell nadir of 84 cells.

Dr Saag reported that the patients had taken a mean of three antiretroviral regimens, so the criteria of successful response on resuming therapy were not set at unrealistic levels. Success was defined as regaining a viral load no more than 0.3 log below pre-interruption level, and a CD4 count no more than 10% below the pre-interruption level.

In viral load terms, 77% had a successful response, and 72% were still below their pre-STI value. In CD4 count terms, 59% had a successful response.

After interruption, the median viral load was 404 copies and the median CD4 count was 230, but Dr Saag did not report on the proportion of those who resumed therapy with a new regimen, which might partly explain the favourable virological response. However, only 11 of those who took a treatment interruption did so because of virological failure; just as many stopped due to side effects (usually GI problems).

These data are likely to make patients and doctors somewhat less apprehensive about the safety of STIs. More than one third of North American physicians have already managed a patient who has undertaken a structured treatment interruption, according to responses at an interactive ICAAC session on current dilemmas in anti-HIV treatment. Fifty-eight per cent had encountered at least one patient who wanted to try a structured treatment interruption.

Another unanswered question regarding structured treatment interruptions is the safety of interrupting treatment with drugs with long half-lives such as efavirenz. A slower decline in efavirenz levels compared to other drugs in an interrupted combination might theoretically lead to the emergence of resistance to efavirenz. Trials of another NNRTI, nevirapine, show that when administered for one or two days only as a treatment to prevent mother to baby transmission, the drug can select the emergence of NNRTI-resistant virus.

A review of all efavirenz recipients in DMP-006 who stopped treatment for 2 days or more due to adverse events showed no significant difference between interruptors and non-interruptors in rates of subsequent viral rebound, but 1 of 7 patients who stopped efavirenz after successful virological suppression after a median of 56 days of treatment harboured the K103N mutation. Those who interrupted treatment for more than 11 days were significantly less likely to regain viral suppression than those who stopped for less than 11 days.