Background:  CCR5 inhibitors are being developed for the management of drug-resistant HIV. These drugs have limited virologic activity in patients harboring CXCR4-using virus (X4), and incomplete viral suppression on these drugs may select for pre-existing X4. Defining the rate of emergence of X4 in treated patients with drug-resistant viremia in the absence of CCR5 inhibitors may help define the “cost” of deferring their use for future salvage regimens and clarify the degree to which CCR5 inhibitors cause expansion of X4.

Methods:  HIV-infected patients with drug-resistant viremia >1000 copies/mL on a stable ART regimen for >4 months were sampled from a clinic-based cohort study. Viral co-receptor tropism was measured with a phenotypic recombinant pseudo-virus assay every 4 months until treatment change. To distinguish co-receptor-mediated entry versus background luciferase activity, entry via CCR5- or CXCR4-expressing cells was confirmed by decreases in relative light units (rlu) with co-receptor inhibitors.

Results:  In 76 patients with drug-resistant viremia, median baseline values were:  plasma HIV RNA level, 3.8 log­­₁₀ copies/mL, and CD4 count, 241 cells/mm³. At baseline, 52 (68%) had R5-tropic virus (R5), 22 (29%) had dual/mixed-tropic viruses (DM), and 2 (3%) had pure X4. There was a median of 3 tropism observations/patient over a median of 9 months. In those with baseline R5, 12% (95%CI 6 to 26%) switched to DM by 1 year. Of 7 patients experiencing R5→DM switch, 3 had switches driven by very low-level CXCR4 entry around the assay’s limit of detection (≤255 rlu) and 1 had repeated oscillation between R5 and DM; the remaining 4 patients experienced the clear emergence of a virus that can effectively use CXCR4 in vitro. In those with baseline DM, 11% (95%CI 3 to 37%) experienced “reversion” to R5 by 1 year and 8% (95%CI 1 to 43%) “progressed” to pure X4. There was no evidence for a change in CXCR4 rlu in those without tropism switches (p = 0.39).

Conclusions:  Among stably treated patients with drug-resistant viremia, the incidence of new tropism changes is relatively low, occurs in both directions, and is often associated with small changes in CXCR4 entry, suggesting natural oscillations in X4 replication near the level of detection. Some treated patients with apparent pure R5 may have also harbored DM in the recent past. Deferring treatment change may carry a small risk of losing CCR5 inhibitors as an effective future treatment option.