The

second carriage phenotype is long-term S. aureus carriage, exemplified by the much slower loss of recruitment spa-types in recruitment-positives, and low loss rates >4–6 months after acquisition in both recruitment-positives and negatives. Our data could not fully support or refute the presence of a third “truly persistent” carriage phenotype as the proportion classified as consistent long-term carriers continued to decline as length of follow-up increased throughout the study. Further follow-up will be necessary to assess this definitively. Using our method of analysis, truly persistent carriage would be indicated by loss rates reducing to zero Selleckchem BIBW2992 some time after 24–30 months (Supplementary Fig. 1(b)) with no further change in the proportion still carrying S. aureus (Figs. 4(b) and 5(a)). Other studies have defined “persistence” using more frequent sampling over shorter timescales, 11 and 12 sometimes using quantitative culture 27; when this study was set-up resource-constraints required a compromise between less intensive long-term versus more intensive short-term follow-up. One important study limitation is clearly the lack of a sampling point earlier than one month, e.g. at one week, which would see more have enabled us to investigate “persistent” carriers defined using van Nouwen’s

rule. 12 The fact that these “persistent” carriers have been shown to differ significantly in characteristics such as clearance of a S. aureus inoculum, 19 and host genetics, 13 indicates Protein kinase N1 that at least a subgroup form a distinct sub-population. However, we did have a sampling point at one month, and Fig. 5(a) demonstrates a clear ongoing linear decline in consistent long-term carriage even in those with two initial positive cultures, suggesting that a proportion with

“persistent” carriage will not carry S. aureus long-term. In fact five of 17 “persistent” carriers (29%, 95% CI 10–56%) were not carrying S. aureus eight years later in the original study of Van den Bergh. 11 Since van Nouwen et al. found that two qualitative and two qualitative + quantitative cultures had almost identical performance for classifying “persistence” in a validation set, 12 it is unclear that doing quantitative cultures in our study would have materially altered this finding; we prioritised spa-typing all isolates over such quantitative culture. Our findings suggest that “persistence” as previously defined12 and 27 could overestimate long-term carriage at the species level, and thus that there is no quick and reliable method to identify consistent long-term S. aureus carriers. Furthermore, 15% of long-term carriers at the species level in our study did not carry the same spa-type consistently (similarly to Ref. 28). Whilst colonised with S.