Effective treatment of patients in a deep salvage situation with "non-active HAART": experiences with the expert advice system RADATA.

BACKGROUND

Clinical studies suggest expert recommendations as a possibility to optimize highly active antiretroviral therapy (HAART) in patients with multi-drug resistant virus strains. An online system (RADATA) has been developed to provide expert advice for the drug therapy of HIV-infected patients.

OBJECTIVE

To evaluate the efficacy of expert-advice-guided HAART switches in patients with triple-class failure.

METHODS

Virological and immunological outcome of patients having undergone at least three prior ART regimens, including nucleoside inhibitor (NRTI), nonnucleoside reverse transcriptase inhibitor (NNRTI), and protease inhibitor (PI) use, were analyzed. Changes in HIV-RNA and CD4-cell count were evaluated every 3 months.

RESULTS

183 patients with a median baseline viral load of 3.90 log copies/ml (1.88-6.54 log) and a CD4-cell count of 298 c/ll (5-910 c/ll) were eligible for analysis. The patients had a median of seven prior ART regimens and a treatment duration of 83 months. A median of three (range 0-8) NRTI-, two (0-7) thymidine-associated (TA), one (0-4) NNRTI-, and three (0-13) PI-associated resistance mutations were present at baseline. Despite available resistance analyses and expert recommendations, 66% (n = 119) of the patients started a new ART regimen without any active drugs according to the resistance analysis. The HIV-RNA declined by a median of 0.61 log and 0.92 log after 12 and 24 months, respectively, while the CD4-cell count rose by a median of +9 c/microl and +25 c/microl during this period. No significant differences related to number of prior regimens or number of active substances used could be found.

CONCLUSION

Despite extensive pre-treatment and multiple resistances against prescribed HAART, our patients demonstrated a decline in viral load and a stable CD4-cell count over the observation period. We conclude that the activity of antiretroviral regimens is not exclusively explained by the current algorithms used for estimating antiretroviral drug activity.