590-T.

|
Head-to-Head Comparison of Widely Used HIV-1 Drug Resistance Algorithms: RetroGram and VGI-TRUGENE HIV-1
C. Torti*1 , E. Quiros-Roldan1 , C. A. B. Boucher 2 , C. Tinelli3, A. M. M. Been-Tiktak4 , S. Lo Caputo5, F. Mazzotta5, F. Castelli1, L. Tomasoni1, S. Day6, and G. Carosi1
1Univ. of Brescia, Italy; 2Univ. Med. Ctr., Utrecht, The Netherlands; 3IRCCS PoliClinic S. Matteo, Pavia, Italy; 4 Virology Networks, Utrecht, The Netherlands; 5 ASL Firenze, Italy; and 6Visible Genetics, London, UK
|
Background. As the number of available antiretroviral drugs is growing, interpretation of genotypic resistance patterns is increasingly complex due to the effect of each individual mutation on drug susceptibility as well as interactions among different mutations. Few data are available to assess concordance across the 2 genotype interpretation systems most frequently used in Italy (i.e. RetroGram and VGI-TRUGENE HIV-1).
Methods. 201 patients failing their current antiretroviral regimens, with >2-year exposure to antiretrovirals and >6-drug experience were enrolled in the Italian GenPherex-Master Study. Reverse transcriptase and protease HIV-1 genome sequences obtained at baseline from 197 patients have been interpreted with both RetroGram 1.4 and TRUGENE HIV-1 (Guidelines 3.0; v1.15). To obtain output normalisation, "no evidence of resistance" and "possible resistance" interpretations from TRUGENE HIV-1 have been compared to RetroGram 1.4 classes A and B, respectively, while "resistance" score in TRUGENE HIV-1 has been compared to classes C+D in RetroGram 1.4. The kappa-statistic measure of agreement was used as reference. kappa >0.40 has been considered significant for agreement.
Results. Amount of agreement ranged from kappa =0.51 for abacavir to kappa =0.86 for lamivudine. Agreement was greater within the non-nucleoside reverse transcriptase inhibitor class (delavirdine, kappa =0.81; nevirapine and efavirenz, kappa =0.84), while it was more varied within the other 2 classes. Intriguingly, the newest drugs showed the lowest concordance (kappa =0.51, 0.54 and 0.63 for abacavir, lopinavir and amprenavir, respectively).
Conclusions. Output interpretation did not differ significantly between RetroGram 1.4 and VGI-TRUGENE HIV-1 last release in our cohort of heavily pre-treated patients and this provides initial cross-validation. However, for some drugs, agreement was moderate (kappa <0.60) and this probably refers to limited data on resistance patterns for drugs more recently introduced into clinical practice. Comparison of genotype interpretation of both systems in relation to clinical outcome parameters is in progress.
|