Livestock-Associated, Antibiotic-Resistant Staphylococcus aureus Nasal Carriage and Recent Skin and Soft Tissue Infection among Industrial Hog Operation Workers

Swine production work is a risk factor for nasal carriage of livestock-associated (LA-) Staphylococcus aureus and also for skin and soft tissue infection (SSTI). However, whether LA-S. aureus nasal carriage is associated with increased risk of SSTI remains unclear. We aimed to examine S. aureus nasal carriage and recent (≤3 months prior to enrollment) SSTI symptoms among industrial hog operation (IHO) workers and their household contacts. IHO workers and their household contacts provided a nasal swab and responded to a questionnaire assessing self-reported personal and occupational exposures and recent SSTI symptoms. Nasal swabs were analyzed for S. aureus, including methicillin-resistant S. aureus (MRSA), multidrug-resistant-S. aureus (MDRSA), absence of scn (livestock association), and spa type. S. aureus with at least one indicator of LA was observed among 19% of 103 IHO workers and 6% of 80 household members. Prevalence of recent SSTI was 6% among IHO workers and 11% among 54 minor household members (0/26 adult household members reported SSTI). Among IHO workers, nasal carriers of MDRSA and scn-negative S. aureus were 8.8 (95% CI: 1.8, 43.9) and 5.1 (95% CI: 1.2, 22.2) times as likely to report recent SSTI as non-carriers, respectively. In one household, both an IHO worker and child reported recent SSTI and carried the same S. aureus spa type (t4976) intranasally. Prevalence of scn-negative S. aureus (PR: 5.0, 95% CI: 1.2, 21.4) was elevated among IHO workers who reported never versus always wearing a face mask at work. Although few SSTI were reported, this study of IHO workers and their household contacts is the first to characterize a relation between nasal carriage of antibiotic-resistant LA-S. aureus and SSTI. The direction and temporality of this relation and IHO workers’ use of face masks to prevent nasal carriage of these bacteria warrant further investigation.


Data Collection
Data were collected between October 2013 and February 2014 by community organizers from the Rural Empowerment Association for Community Help (REACH) with researchers from the Johns Hopkins Bloomberg School of Public Health (JHSPH) and the Gillings School of Global Public Health at the University of North Carolina at Chapel Hill (UNC). Community organizers recruited volunteer livestock workers who fit the following inclusion criteria: currently worked at an industrial hog operation (IHO), resided in North Carolina, could speak English or Spanish, and were at least 18 years of age. Only individuals employed in production of live pigs were considered IHO workers for the purposes of this study; individuals employed in other types of food animal production or meat processing (e.g. slaughterhouses, processing plants, rendering facilities) were not included in the study. Up to three individuals living in the same household as an IHO worker were eligible to participate if they were at least seven years old and spoke English or Spanish. Because enrolled participants were encouraged to refer IHO workers in their social networks to participate, it was not possible to determine the total number of IHO workers who learned about this study, or the percent who were not interested in contacting community organizers for further information about participating. Before participating, adults provided written informed consent. Written parental permission and informed assent were collected for participants seven to 17 years of age (hereafter referred to as "minors"). The JHSPH Institutional Review Board (IRB) approved this study (IRB00004608). The UNC Non-Biomedical IRB approved reliance on the JHSPH IRB.
Each participant attended a baseline enrollment session at the REACH office or in a community meeting space (e.g. a local church). Each enrollment session lasted between 2-3 hours. Participants responded to a baseline questionnaire which assessed demographic information, household characteristics, work activities, risk factors for exposure to S. aureus, and symptoms of SSTI and doctor-diagnosed S. aureus infection during the three months prior to the enrollment session. Participants also provided a self-collected BBL CultureSwab (BD, Sparks, MD) from both of their anterior nares, under supervision and using illustrated diagrams and verbal instruction. Nasal swabs were self-collected rather than collected by investigators based on prior experience in a repeated-measures S. aureus nasal carriage study in which participants demonstrated consistent and repeatable nasal swabbing technique [6]. Self-collected nasal swabs can be as effective in recovering pathogens from the anterior nares as investigator-collected swabs in a supervised setting [6,[26][27][28]. Participants were provided with a monetary incentive after completing study tasks (approved by the JHSPH IRB).

Detection of S. aureus and MRSA
Baseline swabs were transported to UNC at 4°C within 12 hours of collection. A trip blank was included with each shipment to confirm lack of contamination during transport. Within three days of arrival, each swab was clipped into one ml of phosphate-buffered saline and vortexed for 60 seconds. One hundred μL of this eluate was plate spread on CHROMagar™ Staph aureus (BD, Franklin Lakes, NJ) (CSA) while the remaining eluate and swab were refrigerated at 4°C. After incubation at 37°C for 24 hours, at least two colonies with morphology characteristic of S. aureus were streaked on quadrants of a new CSA plate. If direct plating did not yield colonies with S. aureus morphology, swabs and remaining PBS were enriched overnight at 37°C in 10 mL Mueller-Hinton broth containing 6.5% NaCl. A loopful (~10 μL) of Mueller-Hinton broth was streaked onto both Baird Parker and CSA plates to increase sensitivity of detection [29], and incubated at 37°C for 24 hours. Up to two colonies with morphology characteristic of S. aureus on either media were then streaked on quadrants of a new CSA plate. In total, presumptive S. aureus colonies were isolated within two to four days of nasal swab collection. Presumptive colonies were archived at -80°C in brain heart infusion broth with 15% (w/v) glycerol.
A crude DNA extraction was performed on each isolate using a protocol adapted from Reischl et al. [30]. Multiplex PCR was used to amplify the spa, scn, mecA, and mecC genes [31]. Strain LGA251 was used as an extraction and PCR control for spa and mecC, while a clinical MRSA isolate was used as an extraction and PCR control for spa, scn, and mecA. Sterile water was used as a negative control. PCR products were visualized on 2% agarose gels stained with ethidium bromide. Colonies positive for spa, a S. aureus-specific gene, were classified as S. aureus. Singleplex PCR was used to evaluate the presence of pvl-encoding genes among all S. aureus isolates [32].
All isolates were characterized by spa typing using the Ridom StaphType software and the Ridom SpaServer (http://spa.ridom.de/index.shtml) [33]. S. aureus CCs, including CC398 and CC9, are traditionally inferred through multi-locus sequence typing (MLST), which we did not perform. However, spa types associated with CC398 and CC9 have been described by other studies in which both MLST and spa typing were performed [15,34,35]. Thus, we assigned putative CC398 or CC9 to isolates in this study based on spa type, the literature, and the Ridom SpaServer, as others have done [6,36].

Assessment of antibiotic susceptibility
One isolate from each S. aureus-positive nasal swab was assessed for susceptibility to a panel of antibiotics (S1 Table), using the Phoenix Automated Microbiology System (BD Diagnostic Systems, Sparks, MD). Testing was completed by the Clinical Microbiology Laboratory at the Johns Hopkins Hospital. Isolates for testing were shipped from UNC to JHU between January and February, 2014. MRSA were defined as S. aureus isolates resistant to cefoxitin and positive for either mecA or mecC [37]. MDRSA were defined as S. aureus isolates resistant to three or more classes of antibiotics. MRSA isolates meeting the definition of MDRSA were classified as multidrug-resistant MRSA. However, MRSA and MDRSA did not necessarily overlap by these definitions.

Indicators of livestock association
There is currently no consensus marker for LA-S. aureus. We examined four indicators of livestock association among S. aureus isolates: strain type CC398, strain type CC9, absence of scn, and tetracycline resistance. We have used CC398, absence of scn, and tetracycline resistance as indicators of livestock association in previous work [6,38]. CC9 has been identified as a marker of LA-S. aureus among livestock workers in Asia [39], and has been observed among livestock herds and people in contact with these herds in the Midwestern US [7], and in the same geographic region as the present study [6,38].

Definition of recent SSTI outcome
We created the main infection outcome variable, "any symptoms of S. aureus SSTI in the past three months, " as "Yes", "No", or "missing" based on participants' responses to the baseline questionnaire. This variable was coded "Yes" for participants who reported "Yes, in the past three months" to any of the following: doctor-diagnosed S. aureus infection; skin boil; pus-filled abscess; red, painful, swollen skin bump or pimple; or spider bite that was itchy and oozes liquid; "No" for participants who reported "No" to all of the above; or as "Missing. " Prior to responding, participants were shown pictures of S. aureus SSTI with each of these manifestations (courtesy of the Centers for Disease Control and Prevention [CDC] and Dr. Tara Smith, Kent State University). Hereafter, we refer to this outcome as "recent SSTI. "

Statistical analysis
We compared the distributions of potential individual (e.g. use of antibiotics, participation in contact sports) and household risk factors (e.g. number of household members, pets, home located on a hog operation) for S. aureus nasal carriage between workers and household members. For analyses in which prevalence of S. aureus nasal carriage was evaluated as an outcome, results for adult and minor household members were combined for sample size considerations. We calculated the prevalence of nasal carriage of S. aureus, MRSA, and MDRSA among workers and their household members, including livestock-associated classifications of the above.
We used log-binomial regression models to estimate crude prevalence ratios (PR) and 95% confidence intervals (CI) for associations between participant type (worker versus household member) and S. aureus nasal carriage outcomes, S. aureus nasal carriage outcomes and recent SSTI, and occupational exposures and S. aureus nasal carriage outcomes. We used generalized estimating equations with an exchangeable correlation matrix to account for the non-independence of observations within-household [40]. Individual and household-level risk factors that were associated with both participant type and nasal carriage patterns were included in log binomial regression models where sample size allowed. However, adjusted PRs are not presented because inclusion of these covariates did not change the magnitude or precision of effect estimates. All analyses were performed using SAS 9.3 (Cary, NC). An anonymized (de-identified) dataset has been included in the Supplemental Materials showing sufficient variables to reproduce the analysis (S1 Appendix).

Participant and household characteristics
One hundred eighty-three participants, comprising 103 workers (56%), 26 adult household members (14%), and 54 minor children (30%) completed a baseline questionnaire and provided a baseline nasal swab (Table 1). All but two participants identified as Hispanic or African-American. Traditional risk factors for nasal carriage of antibiotic-resistant S. aureus were uncommon among IHO workers and adult household members (e.g. hospitalization, use of antibiotics, use of a gym or workout facility, and/or playing contact sports within the past three months; among adult household members, working in a medical facility was also uncommon). Among minors, recently playing contact sports (36%) and using a gym or workout facility (55%) were more common, but using antibiotics was uncommon (6%) ( Table 1).
In total, the 183 participants comprised 81 households (S2 Table). Most households comprised three to five individuals (73%) and 45% reported having children under seven in the home. Over half reported having no health insurance (58%). Care from a private doctor was most common (44%) but 32% reported seeking care at an urgent care or emergency department for medical care (S2 Table).

Prevalence of S. aureus, MRSA, and MDRSA
Forty-five of 103 workers (44%) and 31 of 80 household members (39%) carried S. aureus at baseline and MRSA was observed in only one worker (Table 2). Twenty-one workers (20%) and eight household members (10%) carried MDRSA (Table 2). S. aureus carried by IHO workers exhibited a larger diversity of antibiotic resistance patterns (10 distinct patterns observed) than S. aureus carried by their household members (five patterns observed) (S3 Table).  Table 2). Putative S. aureus CC9 was only observed among workers (9/103) but putative S. aureus CC398, including CC398 with additional indicators of livestock association, was observed among both workers and their household members ( Table 2). We did not observe pvl-encoding genes among any S. aureus isolates in this study.

Prevalence of and risk factors for recent SSTI
Six of 103 IHO workers (6%) reported recent SSTI (Table 3). Three IHO workers reported a swollen skin bump, two reported a skin boil, and one reported a spider bite that was itchy. One of the <High school b 107 (60) 48 (48)  7 (27)  IHO workers with a swollen skin bump reported this symptom was a doctor-diagnosed S. aureus infection. Six of 80 household members (8%) reported recent SSTI, and all of these household members were minors (6/54; 11%) ( Table 3). Three minors (10, 11, and 12 years old) reported a swollen skin bump and two minors (7 and 8 years old) reported a spider bite that was itchy. One minor reported a doctor-diagnosed S. aureus infection, but did not describe its presentation. Among all participants, we observed that those who carried S. aureus (PR: 4.5, 95% CI: 1.4, 14.9) and MDRSA (PR: 3.1, 95% CI: 1.2, 7.8) ( Table 3) intranasally were more likely to report a recent SSTI. Among IHO workers, we observed a higher prevalence of recent SSTI among individuals carrying versus not carrying MDRSA (PR: 8.8, 95% CI: 1.8, 43.9) and scn-negative S. aureus (PR: 5.1, 95% CI: 1.2, 22.2); associations between recent SSTI and nasal carriage of S. aureus and tetracycline-resistant S. aureus were positive but not statistically significant (Table 3). Characteristics of S. aureus isolates recovered from the anterior nares of workers and minors who reported recent SSTI are described in Table 4. One IHO worker-minor pair in the same household who reported recent SSTI were S. aureus nasal carriage positive and carried the same S. aureus spa type (t4976), although the minor's isolate was tetracycline-susceptible (Table 4). One additional IHO worker-minor pair in the same household who reported a doctor-diagnosed S. aureus infection were S. aureus-nasal carriage negative. None of the six IHO workers who reported recent SSTI carried S. aureus CC398 but two carried scn-negative S. aureus CC9, one carried scn-negative, tetracycline-resistant t002, and one carried tetracyclineresistant t4976 (Table 4).

Occupational risk factors for nasal carriage
Among the 103 participating IHO workers, working between 51-60 hours per week was most common (40/99) (S4 Table). Almost all workers (96%) reported always wearing boots or other foot protection while at work, but always wearing gloves (86%), and always wearing long sleeves and pants or coveralls (86%) were somewhat less common. Only 37% of IHO workers reported always wearing a face mask at work. A higher nasal carriage prevalence of MDRSA (PR: 3.9, 95% CI: 1.0, 15.2), scn-negative S. aureus (PR: 5.0, 95% CI: 1.2, 21.4), and tetracyclineresistant S. aureus (PR: 4.1, 95% CI: 0.9, 17.7) was observed among workers who reported Table 3 never wearing a face mask versus those who reported always wearing a face mask at work ( Table 5). Examination of statistical associations between IHO workers' occupational exposures and recent SSTI was limited due to small numbers. But we did observe increasing prevalence of recent SSTI as IHO workers' reported frequency of face mask usage decreased (S4 Table).

Discussion
In this study population of IHO workers and their household contacts, we found that nasal carriage of S. aureus was associated with recent symptoms of SSTI. Among IHO workers specifically, we observed positive associations between nasal carriage of each S. aureus-related outcome we examined (S. aureus, MDRSA, scn-negative S. aureus, tetracycline-resistant S. aureus) and recent SSTI. Only associations with MDRSA and scn-negative S. aureus were statistically significant; however, the direction and similar magnitude of these associations indicates that nasal carriage may be associated with SSTI outcomes in this study population. Overall, the twelve participants who reported recent SSTI included two IHO worker-minor pairs; one of these worker-minor pairs reported that a doctor had confirmed their SSTI symptoms as a S. aureus infection. This is one of the first reports in the United States of an association between nasal carriage of livestock-associated S. aureus and SSTI among individuals with frequent and intensive exposure to industrial hog production, as well as of SSTI prevalence among their minor child household members. This finding warrants further investigation to assess the directionality of the relation between S. aureus nasal carriage and SSTI in this population.
Few studies have surveyed SSTI among individuals employed in livestock production. In this cross-sectional analysis, we observed a SSTI prevalence of 6.6% (12/183). A retrospective analysis of reported injuries among pork meatpacking and poultry processing workers from 2004-2009 determined average annual SSTI prevalence to be 33% and 12%, respectively [41]. A cross-sectional online and paper-based survey of US pork producers identified a selfreported MRSA infection prevalence of 5/135 (3.7%) [42]. In one other published study of SSTI among hog farm owners and operators that we are aware of, Wardyn et al. observed an incidence of 6.6 cases/1,000 person-months [7]. The present study differed from Wardyn et al. in several ways. First, Wardyn et al. only examined farm owners and operators, for whom exposure to animals and the confinement barn environment may be less intensive than for hired laborers whose job tasks primarily involve direct contact with livestock. Second, participants in Wardyn et al. 's study were primarily Caucasian (98.9%), while participants in this study were primarily Hispanic (89%). Hispanic livestock production workers are believed to significantly underreport occupational injuries compared to Caucasian workers [43]. Third, Wardyn et al. ascertained S. aureus infections using prospectively-collected skin swabs, while we only assessed SSTI through participants' recall of recent symptoms. Participants' recall of SSTI symptoms may have included infections prior to the three months before enrollment, skin afflictions that were not related to a bacterial infection, and/or SSTIs with a bacterial etiology other than S. aureus. In order to minimize the possibility of reporting errors, we showed participants pictures of S. aureus SSTI (provided by the CDC and Dr. Tara Smith, Kent State University) prior to asking them to report SSTI symptoms. In our study, any potential misclassification due to recall was non-differential with respect to S. aureus nasal carriage status since participants were blinded to their lab result at enrollment. However, whether recent SSTI preceded or was a consequence of participants' S. aureus nasal carriage could not be investigated with our study design. Fourth, Wardyn et al. did not describe S. aureus nasal carriage status among individuals who reported SSTI. Thus, any association between antecedent S. aureus nasal carriage and SSTI in their study population is not known. The results presented here provide some of the first insights into the association between nasal carriage of S. aureus with multiple indicators of livestock association and prevalence of SSTI among IHO workers and their household contacts in the United States.
Because we asked participants about SSTI during the three months prior to and including the enrollment session, we were unable to assess whether the S. aureus present in participants' nares were the same strains responsible for recent SSTI. However, we observed associations between nasal carriage of S. aureus, MDRSA, and S. aureus with indicators of livestock association (absence of scn) and self-reported SSTI. Additionally, risk factors for community-associated and healthcare-associated antibiotic-resistant S. aureus infection (e.g. recent use of antibiotics, playing contact sports, recent use of a gym) were uncommon among IHO workers and their household members. Some risk factors were more common among minors (specifically, playing contact sports and recent use of a gym).
S5 Table depicts the frequency and distribution of S. aureus spa types observed among workers and household members. Fifty-nine of the 81 households enrolled in this study contained more than one study participant. Ten of these households contained at least two individuals who shared the same S. aureus spa type at baseline (S6 Table). Seven of these ten were worker-household member pairs, two of ten were worker-worker pairs, and one of ten was a household member-household member pair. Shared spa types included: t094, t233, t337, t4976, t5739, t645, t659, t701 and t7226. S. aureus strains that were concordant between workers and their household members generally did not have characteristics of livestock association (CC398, CC9, tetracycline resistance or absence of scn). One worker-household member pair shared a tetracycline-resistant spa type (t701) and one worker-worker pair shared a S. aureus spa type associated with CC9 (t337), which was also scn-negative. One worker-household member pair shared the same spa type (t4976), but only the worker carried S. aureus with an indicator of livestock association (tetracycline resistance).
Although ten of 59 households contained participants who shared the same S. aureus spa type, we observed little evidence of household transmission of livestock-associated S. aureus. Other studies have detected S. aureus CC398 transmission between occupationally-exposed individuals (i.e. livestock veterinarians) and their household members [44]. We may not have observed transmission of livestock-associated S. aureus in this study population because it does not occur, or because of limitations in our study design. First, we examined transmission events in the context of a cross-sectional analysis. Since transmission events may be infrequent [16], a longitudinal study design might be better suited to capture transmission events. Further, we only examined one S. aureus isolate per nasal swab for spa type and indicators of livestock association. Since a portion (~7%) of S. aureus-colonized individuals carries multiple strain types simultaneously [45], it is possible we missed some instances of concordance. Finally, we assessed transmission events based on shared spa types between household members. However, recent whole genome sequencing studies have demonstrated that shared spa types may not indicate true transmission events; both false positives and false negatives are possible [46,47].
There were several limitations to our assessment of S. aureus-related carriage outcomes. First, because we directly plated nasal swabs, our sensitivity to detect MRSA may have been lower than if we used an antibiotic-supplemented enrichment broth [48]. However, MRSA prevalence among IHO workers in this study (1%) was comparable to a previous study of IHO workers we conducted in this same region (3%) [38], in which an antibiotic-supplemented enrichment broth was used. Second, since~7% of S. aureus-colonized individuals carry multiple S. aureus strains simultaneously [45], as mentioned previously, we may not have captured the true distribution of S. aureus-related outcomes in this population. Third, we only assessed S. aureus nasal carriage to maximize participant acceptability. However, S. aureus colonization has been documented at additional body sites [49]. It is possible that we underestimated S. aureus carriage by not swabbing other body sites. However, it is likely that such underestimation would have been non-differential with respect to the outcome (i.e., similar among individuals with and without recent SSTI), which could have had an attenuating influence on our PR (95% CI) estimates. Other common S. aureus body colonization sites of IHO workers (e.g. axillary, inguinal, and rectal areas) are more likely to be protected in the IHO environment than the anterior nares. Thus, the possibility that we greatly underestimated S. aureus body carriage is likely to be low. Fourth, IHO workers and their household members volunteered to participate in this study; participants were not randomly selected from an enumerated population (e.g. employee roster/records). Thus, it is unclear how generalizable our findings are to all livestock workers in North Carolina or the United States. Nevertheless, the potential for exposure to and infection with antibiotic-resistant S. aureus among the estimated 292,000 livestock workers employed in the United States in 2012 [50][51][52] merits further investigation. Based on the demographics of our study population (>90% Hispanic; <50% health insurance) IHO workers in NC may be particularly vulnerable, as they may not have access to healthcare systems and thus may be difficult to capture via medical records-based or passive surveillance studies.
Additional research is needed to establish the direction and temporality of the association between nasal carriage of livestock-associated, antibiotic-resistant S. aureus and SSTI. In this study population, we observed that nasal carriage of S. aureus, including MDRSA, and scn-negative S. aureus, was more common among individuals who reported recent symptoms of SSTI. This association is in accordance with previously observed associations between S. aureus nasal carriage and infection in clinical settings [12], and provides some of the first evidence in the United States of a potential link between nasal carriage of livestock-associated S. aureus (a modifiable potential risk factor for infection) and SSTI. Because few cases of SSTI were reported, future studies should examine this association in a larger cohort and over time to assess its repeatability and directionality. Overall, this study adds to the growing body of evidence suggesting that individuals exposed to livestock-associated, antibiotic-resistant S. aureus, MRSA, and MDRSA via hog production work in the United States could be at risk for SSTI.

Acknowledgments
This study would not have been possible without a strong partnership between researchers and community-based organizations that have the trust of members of communities in areas where the density of industrial hog production is high. The authors would like to thank the workers and their household members who participated in this study. The authors would also like to acknowledge Norma Mejia, Paul Baker, and Sherri Basnight for assistance with data collection and Nicole Kwiatkowsi and Tracy Howard from the Johns Hopkins Hospital Medical Microbiology Laboratory for assistance with microbiology procedures and sample analysis.