Response to antiretroviral therapy among patients exposed to three classes of antiretrovirals: results from the EuroSIDA study.

There is an increasing proportion of HIV-positive patients exposed to all licensed classes of antiretrovirals, and the response to salvage regimens may be poor. Among over 8500 patients in EuroSIDA, the proportion of treated patients exposed to nucleosides, protease inhibitors (PIs) and non-nucleoside reverse transcriptase inhibitor (NNRTI) increased from 0% in 1996 to 47% in 2001. Four-hundred-and-thirteen patients, who had failed virologically two highly active antiretroviral therapy (HAART) regimens and experienced all three main drug classes, started a salvage regimen of at least three drugs, in which at least one new PI or NNRTI was included. Median viral load was 4.7 log copies/ml [Interquartile range (IQR) 4.2-5.2], CD4 lymphocyte count 150/mm3 (IQR 60-274/mm3) and follow-up 14 months. Of these patients, 283 (69%) subsequently experienced at least a 1 log decline in viral load and 202 (49%) achieved a viral load < 500 copies/ml. Conversely, the CD4 count halved from the baseline value in 88 (21%), and 45 (11%) experienced a new AIDS-defining disease. In multivariable analyses, a 1 log viral load reduction was related to baseline viral load [relative hazard (RH) 1.27 per 1 log higher; P = 0.008], a previous viral load of less than 500 copies/ml (RH 1.69; P = 0.002), more recent initiation of the regimen (RH 1.36 per year more recent; P = 0.02), number of new drugs in the regimen (RH 1.20 per drug; P = 0.02), time since start of antiretroviral therapy (RH 0.94 per extra year; P = 0.035) and time spent on HAART with viral load > 1000 copies/ml (RH 0.96 per extra month; P = 0.0001). Analysis of factors associated with CD4 count decline and new AIDS disease also indicated improved outcomes in more recent times and a tendency for a better response in those starting more new drugs, but no relationship with the total number of drugs. Outcomes in people starting salvage regimens appear to depend on the number of new drugs started but not on the total number of drugs being used.