Transmission of Antimicrobial-Resistant Staphylococcus Aureus Clonal Complex 9 Between Pigs and Humans, United States

Pranay R. Randad; Jesper Larsen; Hülya Kaya; Nora Pisanic; Carly Ordak; Lance B. Price; Maliha Aziz; Maya L. Nadimpalli; Sarah Rhodes; Jill R. Stewart; Dave C. Love; David Mohr; Meghan F. Davis; Lloyd S. Miller; Devon Hall; Karen C. Carroll; Trish M. Perl; Christopher D. Heaney


Emerging Infectious Diseases. 2021;27(3):740-748. 

In This Article


Our WGS analysis suggests that the clonal expansion of LA-SA CC9 in North Carolina is distinct from that in Asia and Europe and that LA-SA CC9 from IHO pigs and humans in high-density pig-producing counties of North Carolina come from a common pool. Considering the high degree of phylogenetic relatedness among intermingled IHO pig and human isolates in putative transmission clusters, the results of this study support potential transmission of antimicrobial-resistant LA-SA CC9 between IHO pigs and humans in the United States.

Our results also provide evidence of household-level transmission of LA-SA CC9 between IHO workers and minors and suggest that potential LA-SA CC9 transmission is not limited to the occupational setting. Dissemination of LA-SA CC9 into the general human population represents a public health concern for 2 reasons. Globally, communities include a higher proportion of children, the elderly, and probably immunocompromised persons, who are at higher risk of developing invasive staphylococcal infections, compared with IHO workers who are predominantly healthy adults of working age. Our analysis revealed an 11-year-old child and an IHO worker residing in the same household who were carrying identical LA-SA CC9 isolates (0 SNP differences) at the same sampling time, which provides strong evidence of household transmission of LA-SA CC9 between IHO workers and their children. Second, clinical implications might arise regarding treatment regimens for LA-MDRSA CC9 colonization and infection. Most (63.3%; 31/49) LA-SA CC9 isolates from North Carolina were multidrug-resistant and carried multiple genes conferring resistance to antimicrobial drug classes critical for human medicine.[23] Of note, the single LA-SA CC9 isolate from an IHO worker who reported a recent SSTI belonged to a putative transmission cluster, displayed an MDRSA phenotype, and previously was reported to display a high degree of pathogenicity compared with a hypervirulent community-associated MRSA strain, USA300 (GenBank accession no. CP000255), in a mouse model of SSTI.[24]

Our results support potential transmission of LA-SA CC9 between IHO pigs and humans, and between humans and other humans, in the top 10 pig-producing counties in North Carolina. These findings are consistent with previous publications on LA-SA CC9 and other lineages of LA-SA. First, a separate analysis of LA-MRSA CC9 recovered from IHO pigs in China suggested potential transmission of LA-MRSA CC9 between pigs, humans, and cows.[11] Second, an abundance of previous epidemiologic and WGS analyses support transmission of diverse lineages of LA-SA from pigs to humans, which can result in human SSTI and bloodstream infections.[8,10,25] Last, prior WGS analyses and epidemiologic studies have provided support for household transmission of LA-SA CC9 and CC398 between persons based on spatial, temporal, and genotypic overlap.[2,3,26] In our analysis, the exact transmission pathway remains unclear because we did not ascertain the direction of transmission or whether transmission occurred through direct or indirect contact.

Previous studies have suggested a S. aureus mutation rate of 5–10 SNPs per year per genome,[27–30] but our threshold of 43 SNPs was justified for 2 reasons. First, our empirically derived SNP threshold was consistent with SNP-based thresholds used by others to identify suspected transmission of MRSA in clinical settings[31] and previous measures of within-person S. aureus diversity.[32] The robustness of our findings was supported when we used the median (32 SNPs), rather than maximum (43 SNPs), pairwise SNP distance among IHO pig cluster isolates as the SNP threshold for identification of putative transmission clusters. We excluded only 1 isolate from an IHO worker from putative transmission clusters, and the excluded isolate was not the SSTI-associated isolate (data not shown). Second, the aim of this study was to clarify whether any SNP-based evidence of transmission between IHO pig and human populations in North Carolina exists, rather than provide evidence of recent or incident transmission or to identify specific pathways of transmission. Using 43 SNPs as the threshold enabled us to observe potential direct or indirect transmission that might not be observed by using epidemiologic data alone. Investigations of S. aureus transmission conventionally combine epidemiologic and strain typing data, but these methods can fail to identify transmission links in cases in which spatial and temporal overlap is lacking.[31] Using the epidemiologic data that were available to us, such as multiple S. aureus CC9 isolates from the same IHO, household, or individual, we observed SNP-based evidence of S. aureus CC9 clustering that would be expected biologically (Appendix Table 5).

Since 2016, tetracyclines have been the most heavily used antimicrobial drug class in the US pig production system, followed distantly by macrolides, lincosamides, aminoglycosides, streptogramins, and fluoroquinolones.[33,34] If antimicrobial-resistant CC9 strains were enriched through selective pressure, antimicrobial use in pig production possibly has played a role in the clonal expansion of LA-SA CC9 in North Carolina and other regions of the world. Of note, resistance to several of these antimicrobial drug classes was conferred by different AMR genes in C1, C2, and C3 (Figure 1; Appendix Table 4), highlighting different evolutionary pathways for adaptation to antimicrobial selection pressures in different regions of the world. Continued surveillance of IHO pigs and humans, including during and after regulatory and policy restrictions on antimicrobial use in animal agriculture, could provide critical insight into the potential contribution of antimicrobial use in the clonal expansion of LA-SA CC9 and its associated AMR genes in the United States.

The strengths of our study included using SNP-based analyses to examine the population structure and transmission dynamics of LA-SA CC9 among pigs and humans in a region of North Carolina with the highest density of IHOs in the United States,[35] a region in which residents and IHO workers are actively expressing concerns about IHO-related exposures.[36] Second, our study used SNP distance to classify human isolates closely related to IHO pig isolates, which is an improvement on previous studies that used spa-typing, MLST typing, absence of IEC genes (specifically scn), phenotypic AMR determination, or combinations of these techniques, to classify S. aureus isolates as livestock-associated.[2,3,12] Third, the use of a SNP-based definition for cluster analysis can capture the potential for transmission between animal and human populations that would have been missed by using more conventional epidemiologic methods alone.[31]

Limitations of our study included that we were not able to provide evidence for directionality of transmission. We rooted our high-resolution phylogenetic tree at the midpoint; therefore, we are unsure if the most ancestral clade of S. aureus CC9 is of human or animal origin. In addition, whereas the SNP-based evidence for pig-to-human transmission could have been strengthened by spatial or temporal data linking pigs and workers at the same IHO, these data were not available because of efforts to protect the privacy of participants enrolled in the epidemiologic studies and because of limited access to US IHOs in the Unites States.[37] In contrast to countries in Europe, the lack of access to IHOs prevents us from assessing the generalizability of our results in the United States. We hypothesize that we would see even closer genetic relatedness between IHO worker and IHO pig LA-SA CC9 isolates collected from the same IHO at the same time. Last, our collection of S. aureus CC9 isolates was limited. The North Carolina collection was a convenience sample that identified S. aureus CC9 isolates from only 6 IHOs, which does not represent the full population of IHOs or pigs in North Carolina. Also, we excluded many isolates selected for WGS from SNP-analysis because they did not pass our quality control criteria (Appendix), potentially introducing bias into the studied isolate sample. Additional S. aureus CC9 isolates likely are available now in the NCBI Reference Sequence Database, but publicly available LA-SA CC9 sequence data were limited when we accessed the database for this study. A more representative dataset could provide more refined estimates on frequency of transmission in North Carolina and other regions of the world.

Despite these limitations, our results show a high degree of phylogenetic relatedness between IHO pig and human LA-SA CC9 isolates in the top 10 pig-producing counties in North Carolina. The presence of a highly pathogenic SSTI-associated LA-SA CC9 isolate with an MDRSA phenotype in a putative transmission cluster warrants future investigations into the disease burden associated with these strains in the United States. Future research could further improve or build on our findings by including environmental isolates and considering WGS analysis in conjunction with spatial and temporal data analysis to investigate the frequency of transmission, environmental exposure routes, and geographic extent of LA-SA CC9. Our reference dataset might be useful in future investigations of worker and community health concerns related to LA-SA CC9 dissemination and acquisition, both in North Carolina and in other regions of the United States with high densities of IHOs.