Background:  Previous studies have shown that the range of phenotypic fold change in didanosine (ddI) susceptibility is remarkably narrow, challenging its use as a predictor of virologic response. We investigated whether and how the fold change in ddI as measured by the ViroLogic PhenoSense assay predicts virologic response to ddI in the Jaguar study.

Methods:  Patients experiencing a virologic failure on combination therapy were randomised in the Jaguar study to receive ddI (n = 110) or ddI-placebo (n = 58). We considered 2 virologic outcomes: the magnitude of reduction in HIV-1 RNA from day 0 to week 4 and a protocol-defined virologic response—a reduction of ³ 0.5 log₁₀ or week 4 viral load < 50 copies/mL. Phenotypic susceptibility score to ddI was defined as partially continuous or continuous and association with HIV-1 RNA reduction was evaluated using linear regression. Fischer’s exact and non-parametric tests for ordered alternatives were used to quantify the strength of association between discrete categories of fold change in ddI and virologic outcomes.

Results:  We randomized 98 patients in the ddI arm had both fold change in ddI and week-4 virologic response available. The median decrease in HIV-1 RNA was 0.57 (Inter Quartile Range, 0.15 to 1.02) and the median fold change in ddI was 1.65 (IQR 1.46 to 1.97). Partially continuous phenotypic susceptibility score was poorly predictive of week 4 virologic response (p = 0.13) while continuous phenotypic susceptibility score was predictive of HIV-1 RNA reduction at week 4 (p < 0.001) with a moderate proportion of variation explained, R² = 0.12. At very low fold change (£ 1.3), 15 of 18 (83%) patients responded compared with 33 of 66 (50%) patients with 1.3 < fold change < 2.2, and 4 of 14 (29%) patients with fold change ³ 2.2 (p = 0.006). Median decrease in viral load was 1.01, 0.50, and 0.10 in patients with fold change £ 1.3, between 1.3 and 2.2, and 2.2 or higher, respectively (p < 0.0001).

Conclusions:  The relationship between phenotypic fold change to ddI and virologic response describes a continuum of susceptibility. At low fold change values (£ 1.3) in these highly experienced patients, the majority of patients responded. These patients also experienced the largest drops in viral load from baseline. Patients with intermediate fold change values (1.3 < fold change < 2.2) had approximately a 50% probability of responding while patients with fold changes ³ 2.2 had a lower probability (29%) of responding to ddI. As such, these data suggest the existence of a fold change range of intermediate probability of response not appreciated previously. 

Keywords: ddI resistance; phenotypic resistance; clinical cutoffs