Marked decrease in acquired resistance to antiretrovirals in latest years in Italy

Objectives:The aim of this study was to evaluate acquired drug resistance in Italy in the 2009e2018period.Methods:We analysed 3094 patients from the Italian ARCA database who had failed antiretroviraltreatment and who had received a genotypic test after 6 months of treatment. Drug resistance mutationswere identified using International AIDS Society (IAS)-USA tables and the Stanford HIVdb algorithm. Theglobal burden of acquired resistance was calculated among all subjects with antiretroviral failure. Timetrends and correlates of resistance were analysed using standard statistical tests.Results:Patients of non-European origin and non-B subtypes increased significantly from 11.5% (103/896) to 19.2% (33/172) and from 13.1% (141/1079) to 23.8% (53/223), respectively, over time. Overall,14.5% (448/3094), 12.1% (374/3094) and 37.8% (1169/3094) of patients failedfirst, second and later lines,respectively. According to both IAS and HIVdb, in the study period resistance to any class, nucleosidereverse inhibitor, non-nucleoside reverse inhibitor, and protease inhibitors (PIs) declined significantly.Integrase strand transfer inhibitor (INSTI) resistance declined significantly from 31% (36/116) to 20.8%(41/197) according to HIVdb but not to IAS. Divergent data were highlighted regarding the proportion ofnon-European patients carrying any, PI and INSTI resistance using IAS tables compared with the StanfordHIVdb algorithm, as the former failed to detect a decrease in resistance while the latter indicates areduction of 1.6-, 5- and 1.8-fold resistance for such drug classes. In the multivariate analysis, the risk ofresistance increased in patients with a larger number of treatment lines and higher viraemia anddecreased in those starting therapy in the last biennium of the study.Discussion:A marked reduction in drug resistance was observed over 10 years, compatible with highergenetic barrier and potency of new antiretrovirals. Nonetheless, concerns remain for subjects with non-Bsubtypes when using mutation lists instead of interpretation systems because of the extensive poly-morphism of the protease region.