To the Editor:
A pivotal component of HIV-1 drug resistance is the occurrence of resistance toward the class of protease inhibitors (PIs).1 Among second-generation PIs, darunavir (TMC114, DRV) was approved by the US Food and Drug Administration (FDA) in June 2006, for use in multiexperienced patients and in those resistant to more than 1 PI. The compound needs a pharmacologic booster with low-dose ritonavir. DRV is a nonpeptidic PI with excellent activity against PI-resistant HIV-1 isolates and was designed after considering the structural modifications in the protease that other PIs induced.2 DRV is chemically similar to amprenavir (APV), with the difference of a bi-THF moiety that determines adjunctive interactions with the asp29 residue in the protease, thus improving the antiviral activity. Several in vitro studies have demonstrated DRV's excellent antiviral profile3; in particular, De Meyer and colleagues4 underlined that DRV has a median effective concentration (EC50) in the nanomolar range toward multi-PI-resistant HIV-1 (defined as a median inhibitory concentration [IC50] increase ≥10 times against at least 1 PI). DRV proved to be potent against multidrug-resistant HIV, with an EC50 around 10 nM.4
The POWER 1, 2, and 3 trials that used ritonavir-boosted DRV (DRV/r) demonstrated that the compound had a large effect in drug-experienced patients.5,6 Recently, 2 different groups defined the phenotypic cutoffs for DRV/r in patients enrolled in the POWER trials,7,8 and prior utilization and resistance to APV were demonstrated to have minimal impact on the response at 48 weeks in the POWER trials.9
We investigated the genotypic resistance toward DRV among samples reported to a large database in Italy. Our aim was to identify possible determinants of genotypic resistance to DRV in a large group of treatment-experienced patients. From the Italian database Antiretroviral Resistance Cohort Analysis (ARCA; available at: www.hivarca.net), we selected the last available HIV-1 protease sequence obtained up to December 2006, from 1169 patients on a PI-based regimen and with a complete treatment history available. These subjects had been selected on the basis of having a resistance test while failing on their antiretroviral regimen (HIV-RNA >1000 copies/mL). These patients had been DRV naive and had treatment with a PI for at least 1 month. DRV resistance score was based on the previously reported set of 11 PI-related mutations: V11I, V32I, L33F, I47V, I50V, I54L/M, G73S, L76V, I84V, and L89V.10 Each mutation has been assigned a resistance “weight” based on the analysis of De Meyer and colleagues,4 and the sum of points assigned to each mutation constituted the resistance score. Virologic response to DRV should be expected to be compromised with a score ≥10. We analyzed our data using 3 score levels: 0 (reference), 1 to 3.5, and ≥4. Multinomial logistic regression analysis was carried out, and adjusted odds ratios (AORs [95% confidence interval (CI)]) were expressed. All data were adjusted for CD4 cell counts and HIV RNA levels.
The ARCA is a large ongoing database focused on drug resistance. Data are regularly collected from 58 Italian clinical centers. Presently, 15,250 sequences from 9667 followed patients are included in the database. The median (interquartile range [IQR]) PI exposure at the time of genotyping was 3.22 (1.71 to 4.90) years. Three hundred 50 (29.9%) sequences exhibited at least 1 DRV score mutation. One hundred 93 (16.5%), 96 (8.2%), 36 (3.1%), and 25 (2.1%) samples had 1, 2, 3, and >3 DRV-related mutations, respectively.
Among the sequences with a DRV score >0, the most frequently represented codon changes were I84V (43.7%, 153 samples), G73S (40%, 140 samples), and L33F (26.6%, 100 samples). Among the 11 mutations that made up the score, I50V, G73S, and I84V were most likely to be found alone; V11I, L33F, I54L, L76V, and L89V were most likely to be found with another mutation; V32I and I47V were most likely to be found with 2 other mutations; and L54M was never found alone.
Statistical analysis revealed an increased risk of developing a DRV score of 1 to 3.5 or ≥4 for each incremental HIV RNA 1 log10 value and for each extra drug used in the medical history within the nucleoside reverse transcriptase inhibitor (NRTI), nonnucleoside reverse transcriptase inhibitor (NNRTI; less so), or PI class. Statistically significant values were more striking (P > 0.05) for the variables when the score was ≥4. The higher duration of drug exposure (per each class and per incremental year) also correlated with a higher DRV score (≥4). Higher CD4 values were always protective with an AOR <1, as shown in Table 1.
We calculated the correlation between DRV score and drug exposure, as illustrated in Figure 1. Among single-drug exposures, APV showed the highest increased risk for a score of 1 to 3.5 or ≥4 (AOR = 2.4 [95% CI: 1.196 to 4.807] and AOR = 8.3 [95% CI: 4.434 to 15.456], respectively), whereas nelfinavir (NFV) seemed to be protective, especially toward a score ≥4. The role of tipranavir (TPV) could not be established with precision because of the small number of patients treated with TPV within the ARCA database.
In this analysis of the ARCA database, 819 (70.1%) of 1169 sequences were without any DRV resistance mutations, indicating full susceptibility to this compound. This was not surprising, because the list of mutations for DRV is different from those of other commonly used PIs. The potential for drug sequencing between lopinavir/ritonavir and TPV or DRV in clinical practice has been recently proposed, after genotypic and phenotypic resistance analysis.11
Determinants of a higher DRV score were high HIV RNA values and increased exposure to antiretrovirals, particularly NRTIs and PIs. Because observational cohorts are largely composed of highly drug-experienced subjects, this finding makes a lot of sense after considering the long therapeutic duration of HIV-positive subjects followed for a prolonged period.
Previous APV exposure was the greatest predictor of having greater DRV resistance, whereas NFV was a protective factor. This may be related to the structural similarity between APV and DRV2 and to the differences in the resistance pathway between NFV and the other PIs,12 respectively.
Retrospective analysis of large cohorts with available matched genotypes and treatment histories provides useful information on the expected activity of newly introduced compounds on previously treated patients. This could help in identifying new mutations not yet included in PI susceptibility scores.
Stefano Rusconi, MD*
Nicola Gianotti, MD†
Fulvio Adorni, BS‡
Enzo Boeri, PhD†
Stefano Menzo, MD§
Angela Gonnelli, MD∥
Valeria Micheli, PhD¶
Paola Meraviglia, MD¶
Michele Trezzi, MD#
Elisabetta Paolini, MD**
Andrea Giacometti, MD§
Paola Corsi, MD††
Massimo Di Pietro, MD‡‡
Laura Monno, MD§§
Grazia Punzi, PhD§§
Maurizio Zazzi, MSci∥
on Behalf of the ARCA Collaborative Group
*Sezione Malattie Infettive e Immunopatologia Dipartimento di Scienze Cliniche “Luigi Sacco” Università degli Studi Milano, Italy
†Ospedale San Raffaele Milano, Italy
‡Istituto di Tecnologie Biomediche- Consiglio Nazionale delle Ricerche Segrate (Mi), Italy
§Ancona Hospital Ancona, Italy
∥Siena Hospital Siena, Italy
¶2a Divisione Malattie Infettive Luigi Sacco Hospital Milano, Italy
#Grosseto Hospital Grosseto, Italy
**Cremona Hospital Cremona, Italy
††Careggi Hospital Firenze, Italy
‡‡SM Annunziata Hospital Antella (Fi), Italy
§§University of Bari Bari, Italy
1. Johnson VA, Brun-Vézinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1: fall 2006. Top HIV Med
. 2006;14:125-130. Available at: http://www.iasusa.gov/
. Accessed April 23, 2007.
2. Surleraux DL, Tahri A, Verscheren WG, et al. Discovery and selection of TMC114, a next generation HIV-1 protease inhibitor. J Med Chem
3. Lo Cicero M, Bulgheroni E, Zampiero A, et al. TMC114: In vitro
activity against HIV-1 primary isolates with reduced susceptibility to multiple protease inhibitors [abstract P315]. Presented at: Seventh International Congress on Drug Therapy in HIV Infection; 2004; Glasgow.
4. De Meyer S, Azijn H, Surleraux DL, et al. TMC114, a novel human immunodeficiency virus type 1 protease inhibitor active against protease-inhibitor resistant viruses, including a broad range of clinical isolates. Antimicrob Agents Chemother
5. Katlama C, Esposito R, Gatell JM, et al. Efficacy and safety of TMC114/ritonavir in treatment-experienced HIV patients: 24-week results of POWER 1. AIDS
6. Haubrich R, Berger D, Chiliade P, et al. Week 24 efficacy and safety of TMC114/ritonavir in treatment-experienced HIV patients. AIDS
7. Winters B, Vermeiren H, Van Craenenbroeck E, et al. Development of Virco®
TYPE resistance analysis, including clinical cut-offs, for TMC114. Antivir Ther
8. Coakley E, Chappey C, Benhamida J, et al. Defining the upper and lower phenotypic cut-offs for darunavir/ritonavir by the PhenoSense assay [abstract 610]. Presented at: 14th Conference on Retroviruses and Opportunistic Infections (CROI); 2007; Los Angeles.
9. Picchio G, Vangeneugden T, Van Baelen B, et al. Prior utilization or resistance to amprenavir at screening has minimal effect on the 48-week response to darunavir/r in the POWER 1, 2, and 3 studies [abstract 609]. Presented at: 14th Conference on Retroviruses and Opportunistic Infections (CROI); 2007; Los Angeles.
10. De Meyer S, Vangeneugden T, Lefebvre E, et al. Phenotypic and genotypic determination of resistance to TMC114: pooled analysis of POWER 1, 2, and 3. Antivir Ther
11. King M, Young TP, Bernstein B, et al. Phenotypic susceptibility to TMC-114 and tipranavir before and after lopinavir/ritonavir-based treatment in subjects demonstrating evolution of lopinavir resistance. Antivir Ther
12. Rusconi S, Viganò O. New HIV protease inhibitors for drug-resistant viruses. Therapy