vitro range between 22% to 58% (17, 18). Another explanation is that we were unable to detect S. aureus due to dry biofilm formation (19). Multiple studies have shown that dry biofilms can be present on most sampled surfaces. Viable bacteria were identified in biofilms, although no planktonic bacteria were present on the surfaces (20-22). Hu et al. showed that over 90% of ICU surfaces contained bacteria in biofilms, and that S. aureus was present in 50% of the cultures (22). Additionally, biofilm hampers cleaning and disinfection (22, 23). Consequently the low contamination rates we found could be an underestimation, although we found no indication for transmission to patients. The identified S. aureus population was highly diverse. While some clusters were identified, many spa types were only observed once. The low number of identified MRSA isolates was as expected, given the low prevalence rates in the Netherlands. The low observed transmission from and to the hospital environment is supported by the fact that a number of environmental isolates belonged to spa types only observed once. As the study was not set up to include all patients admitted to the sampled patient rooms, this is not unsurprising. Remarkably, 26.9% of environment S. aureus isolates belonged to a spa type not identified in nasal or clinical isolates pointing to personnel as likely source of these isolates. Strengths and limitations The main strength of this study was that we looked at the dynamics in nasal carriage, clinical samples, and the environment over a three year period. Additionally, sampling over a threeyear period enabled us to determine long-term presence of S. aureus. Our study also has several limitations. First, and most important, we did not include all patients admitted to the sampled rooms, especially around and during sampling moments. Subsequently, we were limited to determine transmissions to the environment through clinical samples. Consequently, our results most likely show an underestimation of transmissions. Second, our results likely show an underestimation of the environmental contamination rates, and consequently transmissions, due to the limitations of the sampling method and the possible presence of dry biofilm. Third, we performed spa typing instead of whole genome sequencing (WGS), which may overestimate relatedness between isolates. While the discriminatory power of WGS is higher, for the purpose of our study, we believe the discriminatory power of spa typing was sufficient since the diversity index for all sample types was well over >0.95, the criterion described by van Belkum et al. (24). Fourth, we did not determine antibiotic usage, which could partially explain dynamics within patients. Moreover, we did not include healthcare workers, who are a known reservoir of S. aureus. Finally, we only sampled patients once at admission and once at discharge. However, studies have shown that two nasal swabs taken with a week interval can classify MSSA nasal carriage accurately (25). Furthermore, they found persistent carriers did not show one positive and one negative sample in this order taken one week apart. Therefore, it is unlikely that we missed persistent carriers. 108 Chapter 2.4
RkJQdWJsaXNoZXIy MTk4NDMw