This study included all MDR-MRSA strains (n=452) detected in Norway during the study period 2008-2020, from a total of 23,412 MRSA strains. This project was made possible by the continued Norwegian MRSA surveillance effort that has been ongoing since 2008, with the aim of keeping multidrug-resistant pathogens from becoming endemic in healthcare institutions in Norway. Although the overall and yearly proportion of MDR-MRSA was relatively low, we observed an increase in the number of MDR-MRSA strains over time, as well as major changes in molecular epidemiology, during the study period. Specifically, we observed temporal shifts and predominace of certain genotypes (spa-types t1476, t127, t188 and t030/t037), indicating that these were successfully established and disseminating MDR-MRSA clones.

Although the information on place of acquisition for the MDR-MRSA strains is inherently limited by some uncertainty, it is interesting to note that only 13.1 % of strains were registered as aquired in Norway, while a majority were registered as aquired abroad. Compared to the overall MRSA population [12], this suggests import to be responsible for a comparatively larger proportion of MDR-MRSA strains, and similarly that Asia and Africa are the most frequent routes of transmission. This is also supported by the fact that we detect specific clones that have previously been reported as endemic or prevalent in these specific regions and countries [17, 18]. These findings highlight the importance of global cooperation, surveillance, antibiotic stewardship and infection control efforts that can limit both the emergence and global dissemination of these important pathogens.

A high proportion of the MDR-MRSA strains were healthcare-associated, mostly isolated from patients during hospital admission. The proportion of healthcare-associated strains was markedly higher among the MDR-MRSA strain collection than that reported in our surveillance study of the whole Norwegian MRSA population [12]. This suggests that the most highly resistant MRSA clones are more often hospital-associated than less resistant MRSA clones, although they have become more common in the community setting as well. This may among other factors be due to the high selection pressures and competition between MRSA clones provided by high antibiotic exposure [19]. MRSA are also endemic in many hospital environments wordwide [20], thus providing reservoirs for efficient spread into a vulnerable populations, especially in resource-limited settings with inadequate infection control [21]. A limitation to this study is the lack of temporal data on hospital and nursing home admissions, and thus that we had to rely on a broad definition of healthcare-associated MRSA. With this in mind, it is possible that the number of HA- and specifically hospital-associated cases are overestimated. However, the same definition of HA-MRSA was used in the surveillance study of the whole Norwegian MRSA population, in support of the relative differences observed in this study.

Interestingly, there was a larger proportion of MDR-MRSA strains from carriage than from clinical infections, compared to the corresponding numbers for the overall MRSA population in Norway [12]. This finding may reflect that a majority of these strains have been acquired abroad, and are thus (in Norway) mainly detected from asymptomatic carriers, e.g. due to screening upon contact with the healthcare system and differing screening practices. While it is possible that some MDR-MRSA strains have reduced virulence due to e.g. the physiological costs of maintaining resistance mechanisms, this can vary significantly between different strains and settings [22–24]. The prevalence of PVL-positive strains was 20.4 %, which was much lower than for the whole Norwegian MRSA strain population [12]. This also likely reflects the low number of infection compared to carriage strains, as PVL positive strains have typically been linked to clinical infections, whereas PVL-negative strains are more commonly associated with asymptomatic carriage [25].

In the MDR-MRSA strain collection, we observed almost universal resistance to antibiotics such as erythromycin and ciprofloxacin/norfloxacin in addition to cefoxitine. High levels of resistance were also observed for tetracycline, gentamicin, and clindamycin. Thus, in this group of strains, the choice of antibiotic treatment for potential infections is alarmingly and severely restricted. Resistance to mupirocin is however low, meaning that this is still an option for eradication/decolonization of a majority of these strains, while vancomycin and linezolid remain as treament options for all strains.

We observed a very large heteogeneity of spa-types and clonal complexes that were resistant towards five or six groups of antibiotics, indicating that multi-drug resistance is indeed a major challenge within the general MRSA population, not only in a few specific clones. The limited number of genotypes that were resistant to seven or eight groups of antibiotics however included well-known clones like MRSA CC239, that have been evolving antimicrobial resistance in high selection pressure environments for many decades, given their continous presence and global spread since the large hospital outbreaks of the 1970s [26, 27]. This underscores the critical importance of preventing the introduction of MRSA into hospitals and other healthcare facilities, especially in low-prevalence countries, and furthermore the importance of antibiotic stewardship and infection control in mitigating the risk of developing highly resistant strains within these environments.

Overall, we observed significant concordance between phenotypic (AST) and genotypic antimicrobial resistance profiles. In almost all cases, previously described AMR genes or mutations were detected, that could explain the observed phenotypes. Additionally, we noticed high clonality of resistance profiles, which corresponds well with the apparent clonality of specific MGEs carrying these AMR genes. Consequently, one can reasonably anticipate an antimicrobial profile based on genotyping results, although we do not suggest that antimicrobial susceptibility testing should be bypassed. On the other hand, certain genes and mutations which do not appear to confer phenotypic resistance in S. aureus, at least not to an extent which provides resistance according to clinical breakpoints, continue to be reported as resistance genes in the major AMR databases. Furthermore, especially for some groups of antibiotics, the genetics underlying the resistance phenotypes are still not well enough understood and/or are diffucult to interpret. This underscores the necessity for further investigation into the mechanisms underlying bacterial resistance as well as the importance of continuous curation and updating of AMR databases.

Nanopore sequencing technology was used to investigate the specific MGEs associated with AMR genes and mutations. Long-read sequencing techologies such as Oxford Nanopore Techologies (8) have facilitated a much more comprehensive investigation of MGEs, which can often be problematic to resolve by short-read sequencing (9). Consequently, our investigation revealed that AMR genes in MDR-MRSA strains are predominantly located within plasmids, as well as within staphylococcal cassette chromosome mec (SCCmec) elements and transposons, with a limited presence on prophages. The most common AMR genes were associated with wide range of MGEs, including different SCCmec types and multiple transposons and plasmids. Some MGEs were clonal, while others were widely distributed, indicating high potential for spread within the MRSA population. A majority of the plasmids were furthermore closely related, although differing in AMR traits, thus indiciating high plasticity. The most prevalent AMR phenotypes were furthermore caused by multiple AMR genes and/or mutations. As an example, we observed eight different genes or gene combinations in strains with MLS resistance. This large variation is likely a consequence of high selection pressure and thus multiple adaptations of MRSA clones over a long period of time, which now serves as an arsenal of genes available for providing resistance against different MLS antibiotics. This high diversity both in AMR genes and in acquisition mechanisms thus provide a significant advantage for dissemination of antibiotic resistance within the MRSA population. These findings highlight the complexity of resistance gene dissemination and the adaptive strategies employed by S. aureus in response to antimicrobial pressures.