Gilliver et al. reply

The study by Österblad et al. confirms the importance of understanding the role of commensal bacteria, particularly in wildlife, in the ecology of antibiotic resistance. The two studies combined suggest that the gut flora of wildlife populations with very little or no contact with either humans or anthropogenic antibiotics (Österblad et al.'s study) may have negligible levels of antibiotic resistance, whereas wildlife populations living in closer proximity to humans but still with no known direct contact with anthropogenic antibiotics (our study) may have much higher levels of antibiotic resistance.

These conclusions fit with earlier findings of a higher prevalence of antibiotic resistance among baboons living close to humans than in baboons in more isolated populations1. Questions that still need to be addressed concern the extent and frequency of antibiotic exposure necessary to generate significant resistance, what determines the dynamics of decline in resistance following restrictions in antibiotic use, and the nature and extent of any reservoir of antibiotic resistance that may exist in natural environments and which could undermine future attempts to manage resistance.

These questions can only be resolved by thorough spatial and temporal mapping of antibiotic resistance in natural environments. We inferred from our study that it would be unwise to assume that resistance would decline significantly as a consequence of restricted use of antibiotics. This suggestion still holds, because resistance has been maintained for over three years at our study site, over several generations of rodents, without any obvious exposure to antibiotics.

We agree with Österblad et al. that antibiotic restrictions should still be very much on the agenda, but that agenda must include concerted attempts to understand what the consequences of restrictions are likely to be.