Also, neither was predictive of immunological IWR 1 response at week 24 (P = 0.939 and P = 0.866, respectively). Baseline RC was, as expected, low in this cohort of heavily ARV treatment-experienced patients on failing therapy. Mean RC increased by 33% following a 12-week ARDFP to values commensurate with those recently reported in a cohort of treatment-naïve patients [20], probably reflecting a return to the wild-type viral population. RC remained unchanged during that period in patients not undergoing treatment interruption (no-ARDFP). Consistent with this observation, PSSs at the start of salvage antiretroviral therapy were significantly higher among patients who had undergone a treatment interruption (ARDFP) than among the no-ARDFP

patients. Unlike findings reported in naïve patients, RC did Dabrafenib supplier not correlate with baseline viraemia in this very ARV-experienced cohort [17, 20]. There was a parallel decline in CD4 cell count and an increase in viral load during ARDFP. However, baseline RC did not correlate significantly with change in CD4 cell

count or viral load during ARDFP. In patients randomized to no-ARDFP, RC also did not predict change in CD4 cell count or viral load following 12 weeks of therapy change, despite significant improvements in both CD4 cell count and viral load. Many studies of treatment-naïve patients have demonstrated the ability of baseline RC to predict changes in both CD4 cell count and Tryptophan synthase viral load following treatment initiation [17, 19, 20]. However, RC is rarely taken into account in treatment decision-making for ARV-naïve patients in routine clinical care. It appears not to have the same predictive value in ARV-experienced patients, where it is most often obtained. There was no significant change in RC among no-ARDFP patients. These data suggest that the absence of drug pressure during treatment interruption increases RC. This increase is probably only transient and is expected to be reversed at reinitiation of therapy. Our findings of stable RC values over the 12-week course of salvage therapy in no-ARDFP patients confirm previous observations that accumulation of resistance-associated

mutations during partially suppressive HAART (as in many patients in that group) is not likely to induce further RC decline, probably because secondary ‘compensatory’ mutations appear [28, 29]. Among patients not undergoing a treatment interruption, PSS was correlated with baseline RC and was highly predictive of virologic response to salvage ARV regimen up to week 48, after controlling for other baseline variables: RC, CD4 cell count and viral load. As has been previously demonstrated, PSS can be used to identify treatment regimens containing sufficient antiviral activity to improve virological outcome. iPSS did not have a significantly better predictive value than dPSS. Findings for the ARDFP group suggest that measuring either iPSS or dPSS following treatment interruption does not appear to have any predictive value.