To the editor—We thank D'Aquila et al.1 for their comments on our recent News and Views article2. We fully agree that both CXCR4 and CCR5 inhibitors should be evaluated in clinical trials, if suitable compounds are developed. However, we wish to clarify some aspects of the literature raised by D'Aquila et al. on the emergence of CXCR4-using syncytium-inducing (SI) HIV-1 strains in vivo. First, the study we cited3 for the lack of association of the SI viral phenotype with heterozygosity for CCR5 Δ32 was not, in fact, a study of "newly infected patients." The subcohort of 406 HIV-1-infected subjects examined in that study was seroprevalent with an average time to AIDS of 5.8 years, compared with 9–10 years as seen in seroincident cohorts3. Moreover, although D'Aquila et al. did indeed find an association of the SI viral phenotype with CCR5 Δ32 heterozygosity in their study of 95 individuals4, larger studies have either not confirmed this observation5 or have found a non-statistically significant trend6. On these last points, we look forward to the results of meta-analysis (ref. 7 and T. O'Brien, personal communication) to provide more definitive evidence as to whether there is any association of viral phenotype and chemokine receptor genotype.