Background:  In patients failing HAART, significant antiviral activity of didanosine (ddI) can be retained in spite of resistance mutations in RT. We have compared the ability of RT genotype and of phenotypic resistance to ddI to predict this activity.

Methods:  In a randomized, placebo-controlled study, patients under stable ARV therapy and plasma HIV RNA >1000 copies/mL (median = 3.8 log/mL) were randomized to add ddI or placebo to their failing regimen for 4 weeks. Patients (168) received either ddI  (n = 110) or placebo (n = 58). Plasma HIV RNA change from baseline was evaluated at week 4. RT genotype examined the presence of thymidine analogue mutations (TAM) and of all nucleoside analogue mutations (NAM). Phenotypic resistance to ddI was measured using the Phenoscript assay (Viralliance, Paris) and expressed as a resistance index (RI = fold-change in IC₅₀ relative to NL4-3). The number of mutations and the RI values were divided into quartiles and their relationship with W4 HIV RNA change was evaluated with the Kruskal-Wallis test. The correlation between genotype and phenotype was analyzed using the Cochran-Mantel Haenszel test.

Results:  At baseline, the median nb of TAM was 3 (IQR: 1 to 3) and the median nb of NAM was 4 (IQR:  2 to 5). The median baseline RI for ddI was 1.60 (IQR:  1.2 to 2.0). In spite of the very narrow range of RI values for ddI, there was a significant correlation between the RI for ddI and the number of TAM or of NAM (p = 0.0007 and p = 0.0006, respectively). The RI for ddI was also affected by the presence of a L74V mutation (n= 12, p = 0.0008). At week 4, median HIV RNA change was -0.6 log/mL in the ddI arm and +0.07 log/mL in the placebo arm, (p <0.0001). In the ddI arm, a strong correlation between W4 HIV RNA change and baseline nb of TAM or of NAM was observed (p = 0.0004 and p = 0.0003, respectively), while no correlation was seen in the placebo arm (p = 0.54 and p = 0.94 respectively).  By contrast, the relationship between W4 HIV RNA change in the ddI arm and baseline phenotypic resistance to ddI was weaker and non linear (p = 0.027).

Conclusions:  Significant antiviral activity of ddI can be retained in nucleoside analogue-experienced patients. This activity can be predicted by the number of TAM or NAM in HIV RT, but the predictivity of phenotypic resistance to ddI, as measured in tissue culture assays, is less accurate.

Keywords: didanosine; resistance; phenotype