HIV2EU: supporting standardized HIV‐2 drug resistance interpretation in Europe

Abstract

Various items are complicating the treatment of HIV‐2 infected patients. Compared to HIV‐1 there is much less treatment experience, no evidence from randomized control trials, a reduced number of effective drugs and no broadly available test for viral load monitoring. In case of treatment failure there is only limited guidance and presently no easy accessible tool for nucleic acid sequence interpretation available. To solve this problem, we initiated an expert workshop to address some of these problems. A panel of experts from four different European countries voted on a rule set for interpretation of mutations in the HIV‐2 protease, reverse transcriptase and integrase. Rules were proposed by each member and were then modified during discussion by considering data gained from HIV‐1 and accumulated experience of the follow up of HIV‐2‐infected patients. Based on the HIV‐GRADE internet‐tool an online tool was developed to make the rule set easily accessible and usable. Rules were laid down for the interpretation of HIV‐2 drug resistance to NRTIs, PIs and INIs (integrase inhibitors). Due to natural resistance of HIV‐2, usage of NNRTIs and T‐20 was not recommended as part of an antiretroviral regimen for HIV‐2. These rules were then translated in a machine interpretable format (algorithm specification interface, ASI) and the HIV‐GRADE tool was extended for usage of HIV‐2 sequences. Further consensus sequences were generated from the reference sequence data set provided by Los Alamos National Laboratories. In contrast to HIV‐1, mutations were compared to a group specific consensus sequence (Group A or Group B) and not to a consensus sequence from the most predominant HIV‐2 Group A. This change was necessary due to significant differences between the various HIV‐2 strains. We developed a rule set and an automated tool for HIV‐2 drug resistance analyses. This tool and the rules will be freely available on the internet. Access to the pre‐publication versions can be granted by each of the group members. To keep the algorithm rules up‐to‐date it will be actualized on a yearly basis.