Abstract
Objective
To assess the relationship between viral susceptibility at baseline and virological response in human immunodeficiency virus (HIV)–infected patients treated with multi-drug salvage regimens after multiple previous treatment failures.
Design
Retrospective analysis of 50 patients from the Frankfurt HIV cohort who had received treatment with a minimum of six drugs, and for whom a sample for baseline viral phenotyping was available. Methods: Viral drug susceptibility was measured retrospectively from stored samples using the Antivirogram, a recombinant virus assay based method. Virological response was defined as a viral load of <400 copies/ml at week 24. For analysis of treatment response, drop-outs were dealt with in two ways, either as failures (DAF) or censored (DAC). Several logistical regression models were applied to identify predictors of response, including baseline virus load, number of new drugs and phenotypic sensitivity scores.
Results
At baseline, drug resistance was extensive: 96% of patients had viruses resistant to at least one drug class and 32% had viruses resistant to all three drug classes. In the DAF analysis, 39 patients experienced virological failure. In the DAC analysis, eight were censored and 31 patients experienced virological failure. In multivariate models that adjust for baseline viral load, the number of new drugs and total phenotypic sensitivity scores, the baseline viral load and phenotypic sensitivity score remained significantly associated with virological outcome, whereas in those adjusted for baseline viral load, the number of new drugs, NRTI phenotypic sensitivity score and PI phenotypic sensitivity score, only the latter remained significantly associated with virological outcome. Both the DAF and DAC analyses produced similar results. In all models used, virological failure was shown to be significantly associated with baseline viral load and phenotypic sensitivity score.
Conclusions
In this retrospective analysis based on a small number of patients, viral drug susceptibility at baseline was strongly associated with virological outcome at 24 weeks, independent of covariates such as baseline viral load and treatment history. Baseline viral load also maintained a significant, independent association with virological outcome in most models.