Changing antiretrovirals whilst viral load <50 copies/ml and relationship with CD4 count changes

Purpose of the study

The frequency and reasons for switching antiretrovirals (ARVs) in patients on a fully suppressed cART regimen (viral load [VL] <50copies/ml) is not well described, nor is the effect of such a change on CD4 counts.

Methods

6713 patients from EuroSIDA on cART with a confirmed VL<50 copies/ml were included; a regimen change was defined as >1 ARV change (occurring on the same day) for any reason whilst VL< 50 copies/ml. Baseline was defined as the first VL<50 copies/ml on cART; Kaplan Meier methods estimated the probability of ARV change and Cox proportional hazards models, stratified by centre, identified factors associated with ARV change. Mixed models were used to model the change in CD4 count after the first ARV change.

Results

At baseline, the median CD4 was 414/mm³ (IQR 272—587). 1079 (16.1%) patients changed 1358 ARVs; 93 (8.6%) patients started an ARV they had previously taken. 224 (77.0%) of those starting an NNRTI (n=291) were previously naïve to NNRTIs, compared to 29 of 306 who started a PI or boosted-PI (9.5%). The incidence of changing ARVs was 11.8 per 100 PYFU (95% CI 11.1-12.5). At 1 year after baseline, 10.7% were estimated to have changed >1 ARV (95% CI 9.8-11.5). The most common reason for change was toxicity (n=521, 38.4%), followed by patient or physician choice (n=398, 29.3%). Table 1 shows the factors associated with changing ARVs.

Table 1

Conclusions

Changing ARVs whilst virologically suppressed was due to patient/physician choice or toxicity, increased in frequency over time and was more common in patients taking a single PI-regimen or in stavudine-containing regimens. Patients who changed ARVs had a small but statistically significant boost to CD4 count levels, and the increase in CD4 was higher in those who changed to a new class of ARV.