4%). The most common FTC resistance mutation was M184V (86.7%). The PrEP drug resistance levels estimated in UK HIV-infectious MSM of 1.6, 0.9 or 4.1%, depending on the definition used, were within the range of values used Estrogen antagonist in simulation studies, which have suggested that circulating PrEP drug resistance will have negligible impact on PrEP efficacy [18]. The decline in PrEP resistance occurred despite an increase in the use of TDF (from 43.4 to 55.9%) and FTC/lamivudine (from 70.3 to 78.1%) between 2005 and 2008 in UK MSM on treatment. Conversely, zidovudine (ZDV) usage, the major driver for the development of TAMs, was found to have decreased

from 31.4 to 11.0%. Our study has a number of limitations. First, all mutations have been regarded as reducing susceptibility to PrEP commensurate with their impact on the efficacy of ART for treatment. However, the impact of mutations on PrEP efficacy is unknown, and Cong et al. [5] speculate that

TDF resistance may have a greater impact than FTC resistance. Furthermore, our TDF-FTC resistance definitions represent a worst-case scenario for PrEP resistance, as it is unlikely that exposure to HIV with only FTC mutations, such as M184V, would result in infection because of the increased sensitivity of TDF [5, 9] and because viruses with both K65R and M184V mutations have been shown [19] to have increased susceptibility to TDF compared with HIV with the K65R mutation alone, so true GBA3 PrEP resistance is likely to be lower selleck chemical than the calculated prevalence.

Secondly, although the methodology used in this paper avoids the overestimation of resistance that is known to occur if only data from ART-experienced patients with resistance tests are used [14], there may be unrecorded covariates (e.g. clinician’s assessment of adherence) which influence which patients are selected for resistance testing and introduce selection bias which cannot be controlled for. Thirdly, despite, in our methodology, the calculated PrEP resistance being adjusted for the reversion of TDR mutations between infection and resistance test, this is still likely to be an underestimate of true PrEP drug resistance. Our methodology assumes that diagnosis occurs 2 years after infection, but the time gap is likely to be larger. Fourthly, transmission risk has been found to be linked to the level of viral load [12], although a meta-analysis [20] found large variations between studies, precluding reliable estimation of a per-act transmission probability for MSM. Therefore, the plasma viral load measurements in this analysis were used to classify individuals as infectious or not infectious and the actual level of viral load has not been taken into account. Finally, simplistic weighting based on estimated population size was used to combine the various diagnosis/treatment groups. Ideally, this should consider the difference in sexual risk behaviours known to exist based on diagnosis status [10, 11].