Impact of baseline protease genotype and phenotype on the response to darunavir outside clinical trials

Sir, Darunavir is the latest protease inhibitor (PI) approved for the treatment of HIV infection. It has a unique biochemical structure that confers a very high binding affinity for the HIV protease active site. These physical properties explain the good performance of the drug in clinical trials carried out in heavily antiretroviral-experienced patients, most of whom were infected with viruses showing extensive PI resistance. Current understanding of predictors of darunavir resistance has mainly been built based on data derived from the POWER studies. The following 11 protease mutations have been proposed to impair darunavir susceptibility, segregated as major (I50V, I54M, L76V and I84V) and minor (V11I, V32I, L33F, I47V, I54L, G73S and L89V) resistance mutations. It has been proposed that when 3 of these mutations are present, darunavir antiviral activity is significantly diminished. We have assessed the efficacy and durability of the response to darunavir based on genotypic/phenotypic data and clinical parameters in 21 antiretroviral-experienced HIV-1-infected individuals who began a rescue intervention based on darunavir/ ritonavir (600/100 mg twice daily) at a single HIV/AIDS reference centre located in Madrid and monitored for at least 4 weeks. All patients had previously failed at least two ritonavirboosted PIs, including in all cases lopinavir, tipranavir and/or fosamprenavir. Genetic sequence analyses of the pol and env genes were performed and drug resistance mutations were interpreted following the latest International AIDS Society-USA panel list (www. iasusa.org, October 2006). HIV-1 resistance phenotyping was carried out using a commercial phenotypic resistance assay (Antivirogram, VIRCO, Mechelen, Belgium). The level of phenotypic resistance was expressed as the fold change (FC) in IC50 compared with a wild-type reference virus. Statistical analysis was performed using SPSS v13.0 (SPSS Inc., Chicago, IL, USA). Median baseline plasma HIV-RNA was 4.49 log copies/mL and the median CD4 count was 192 cells/mm. Overall, the median number of prior PI failures was 4 (IQR, 2–4) and the median number of listed IAS-USA protease resistance mutations was 12 (IQR, 10–14). However, the median number of specific darunavir resistance mutations was 1 (range, 0–5). Genotypic and phenotypic data before beginning salvage therapy based on darunavir are recorded in Table 1. Only two patients (patient no. 4 and patient no. 8) had 3 darunavir-associated resistance mutations. In spite of the high number of total protease resistance mutations and phenotypic resistance to most other PIs, 85.7% of patients showed phenotypic susceptibility to darunavir. In agreement with the lower clinical cut-off proposed for darunavir/ ritonavir (10-fold change), only patients 2, 4 and 8 showed a darunavir-resistant phenotype at baseline (20.9, 16.4 and 66.2 FC values, respectively). All but three patients experienced a significant decline in viraemia (.0.5 log HIV-RNA copies/mL) at week 4. In fact, the median drop in plasma HIV-RNA was 1.78 log (range, 0.31– 3.05) in the whole population. More importantly, 80.9% of patients achieved and maintained viral load values below 50 HIV-RNA copies/mL during a mean follow-up of 38 weeks (range, 4–140 weeks) under the same regimen. Treatment adherence was .95% in all but one patient (no. 15), who showed a poor virological response, despite being susceptible to darunavir. Overall, darunavir was well tolerated and no serious adverse events were recorded. Two-thirds of patients had at least one additional active drug, as reported by the phenotypic assay and/or the pre-darunavir genotype, along with darunavir in the salvage regimen. It was a nucleoside reverse transcriptase inhibitor in 42.8%, raltegravir in 35.7%, enfuvirtide in 21% and a non-nucleoside reverse transcriptase inhibitor in 14.2% (Table 1). A greater viral load drop and CD4 increase was seen in patients who received darunavir along with enfuvirtide or raltegravir as the additional active drug compared with the rest of the patients (2.4+ 0.8 versus 1.6+ 0.9 HIV-RNA copies/mL, respectively, P 1⁄4 0.04; and 137.6+ 176.9 versus 65.5+ 126.9 CD4 count, respectively, P 1⁄4 0.3). Patients 1, 2 and 3 experienced viral rebounds at weeks 11, 9 and 18 of follow-up, respectively. At failure, mutation V32I was selected in patients 1 and 2, and was associated with FC increases of 9 and 30.4, respectively. Mutation I50V was selected in patient no. 3 and it was associated with a FC increase of 285.4 with respect to baseline. The concomitant use of enfuvirtide or raltegravir as additional active drug [mean difference in viral load drop (log) 1⁄4 1.1, 95% CI: 0.3–1.84; P 1⁄4 0.005] was independently associated with the viral response at week 4, in agreement with recent subanalysis of the POWER studies. When the same analysis was performed in a subset of patients who did not receive enfuvirtide or raltegravir, the total number of protease resistance mutations was the variable with the highest impact on the virological response at week 4. Herein, we confirm the very high genetic barrier for resistance to darunavir outside clinical trials in a small but welldefined group of patients. Darunavir showed antiviral activity even in patients harbouring viruses with up to 16 protease resistance mutations. It is noteworthy that 80.9% of these patients achieved undetectable viraemia, maintained during a mean of 38 weeks of follow-up. This is an unprecedented result using PIs despite the fact that raltegravir and/or enfuvirtide were altogether used by 57% of the study population. Correspondence