As the first study to determine epidemiology of MRSA nasal colonization for HIV-infected patients in mainland China, we found the prevalence (11.89%) was higher than the result of a global meta-analysis which showed the pooled worldwide prevalence was 7% (95% CI 5%-7%) in adult HIV-infected patients . Moreover, the prevalence in our study was also higher than other high-risk populations such as dialysis patients (6.2%; 95% CI 4.2%-8.5%) , men who have sex with men (1.6% ; 95% CI: 0.5–2.6%)  and injection drug users (7.4%) . This high prevalence of MRSA colonization could be explained by high rates of antibiotic use and not effective infection control programs. Further studies need to confirm the reasons.
We found HIV-infected patients with diagnosis of respiratory tract infection in the previous 6 months were a more likely to have MRSA nasal colonization, which could be associated with the more frequently respiratory episode occur in HIV-infected population and bacteria colonize in the nasal cavity as the respiratory barrier weakens. Male gender was a risk factor for MRSA colonization without MRCoNS, which was consistent with the previous results [19,20]. It may be related to the fact that 428 MSM out of 843 men recruited for this study (data are not shown), study reported that MSM was found to have a high rate of MRSA infection in other contexts regardless of HIV status. According to some studies, young people have higher risks of bacterial colonization [20,22], however, we did not find the same result in this study.
High proportions of resistance to penicillin, erythromycin, clindamycin and tetracycline were found in MRSA isolates, which is similar to observed studies [23–25]. This finding suggests medical professionals should pay more attention to the rational use of these antibiotics.
The detection rates of toxin genes for MRSA isolates in HIV-infected patients were lower than hospital patients [26–29], but similar to general populations [30–32]. Base on the results of SCCmec typing we found most MRSA isolates in this study were community-associated, which was similar to other studies [33,34]. However, hospital-associated SCCmec types were also found in this study, indicating a cross transmission between the community and hospital facilities. Moreover, the dominant STs (ST5, ST45, ST59 and ST188) in this study were consistent with other community-based populations [26,35–36], but some other STs were also previously reported in hospital patients (ST338 and ST1) [37–39] and animals ( ST398 and ST1) [40–42]. The MLST findings also indicated the cross transmission between communities and hospital settings.
Comparisons of characteristics, including antibiotic resistance and toxin genes as well as results of molecular typing, for MRSA isolates between with and without MRCoNS isolates were of no significant difference, suggesting the phenotypic and molecular characteristics of MRSA colonization were not affected by co-colonizing with MRCoNS.
To our knowledge, it is the first molecular epidemiological study of MRSA colonization with a large sample size among HIV patients in mainland China. However, some potential limitations also need to be considered. Firstly, we could not draw a cause conclusion for MRSA colonization in HIV patients because of the cross-sectional study design. Secondly, we administered the sampling at only one site, which may underestimate the prevalence of MRSA colonization. Thirdly, colonization was only assessed on the day of enrollment. Evaluation of durability of colonization may be useful to fully understand MRSA colonization dynamics.