• Predictors of Long-Term Response to Therapy Resumption After Treatment Interruption in HIV-1 Infected Patients Failing Antiretroviral Therapy
  Authored by Nicola Gianotti, Alessandro Soria, Laura Galli, Daniela Vacchini, Massimo Cernuschi, Adriano Lazzarin
  Abstract PL6.3

One of the more controversial strategies that has been suggested for people who have experienced virological failure to more than one HAART regimen is to interrupt therapy. The main rationale is the observation that levels of resistant virus in plasma are often observed to be greatly reduced over the period of a few months off therapy. This is replaced by non-resistant viral strains (sometimes referred to as "wild type"). Most virologists would argue that resistant strains of virus have not disappeared completely (i.e., become extinct) but have merely reached levels too low to be detected by most resistance tests. Nonetheless -- while perhaps unlikely -- it cannot be ruled out that some strains do actually become extinct.

Trials randomizing people to interrupt or not before starting a "salvage" regimen have been performed with some indication that viral load responses to the salvage regimen are improved in those who have interrupted. However, long term outcomes are going to be very important as they may give a clue as to whether there is a long term advantage (perhaps due to extinction of some strains, or their restriction to being present only in long-lived latently infected cells).

This analysis of data from a clinic cohort in Milan added to the body of data on this issue by presenting data on longer follow-up and evaluating the factors associated with a better response to the post-interruption regimen. Gianotti et al. selected patients who had interrupted therapy for at least 60 days having had a viral load of more than 400 on their last regimen and for whom at least 48 weeks of follow-up on the post-interruption regimen was available. Fifty-six patients met this criterion. They had been exposed to a median of eight drugs from three classes. The mean viral load at interruption was 4.4 log copies/mL. The median duration of interruption was 112 days. During interruption the mean viral load increased to 5.18 log copies/mL and the mean CD4 count decreased from 344/mm³ to 266/mm³.

Nineteen (34 percent) of the 56 patients had a viral load below 400 copies/mL at 48 weeks after interruption. The mean viral load overall was 3.8 log copies/mL. There was a strong tendency for those with better viral load responses to also experience better CD4 count responses. Many factors were analyzed to try to identify which were independently associated with viral load below 400 copies/mL. The only factors that emerged were the maximum viral load before interruption (higher viral load, lower chance of viral load below 400 at 48 weeks) and, interestingly, those with longer duration of interruption were also more likely to have viral load below 400 copies/mL.

The obvious concern with treatment interruptions is the possible increased risk of AIDS events -- not only due to the CD4 count decline but also because there appears to be a beneficial effect of HAART over and above what can be seen from the CD4 count. This analysis is not the best placed to examine this issue as those who died either during interruption or within 48 weeks of starting the post-interruption regimen were excluded. However, the authors reported that there were only four AIDS diseases during the 48 weeks of therapy. All occurred in those patients interrupting who had CD4 counts below 200/mm³.

Treatment interruption in people with a CD4 nadir below 200/mm³ remains a risky strategy. It is as yet far from clear whether any virological benefits will ultimately transfer into sufficiently large clinical benefits to make the risks of interrupting worth taking.