Prevalence of USA 300 colonization or infection and associated variables during an outbreak of community-associated methicillin-resistant Staphylococcus aureus in a marginalized urban population

357 1Canadian Field Epidemiology Program, Public Health Agency of Canada, Ottawa, Ontario; 2Department of Health Care and Epidemiology, University of British Columbia, Vancouver, British Columbia; 3Calgary Health Region; Departments of 4Community Health Sciences; 5Pathology and Laboratory Medicine, 6Medicine; 7Microbiology & Infectious Diseases, University of Calgary; 8Alberta Provincial Laboratory for Public Health; 9Calgary Laboratory Services; 10Centre for Antimicrobial Resistance, University of Calgary, Calgary, Alberta Correspondence: Dr Mark Gilbert, Division of STI/HIV Prevention and Control, BC Centre for Disease Control, 655 West 12th Avenue, Vancouver, British Columbia V5Z 4R4. Telephone 604-660-6149, fax 604-775-0808, e-mail mark.gilbert@bccdc.ca Received for publication May 22, 2007. Accepted August 7, 2007

T here is growing concern about the current epidemic of community-associated methicillin-resistant Staphylococcus aureus (MRSA) infections globally.The USA300 strain of MRSA (the CMRSA10 strain by Canadian nomenclature), in particular, has emerged as a dominant clone and public health threat (1,2) in the United States (3)(4)(5) and Canada (6).This strain has caused numerous outbreaks among professional athletes (7), military recruits (8), men who have sex with men (9), prison inmates (10), homeless youth (11) and tattoo recipients (12).The USA300 strain can cause severe disease including sepsis, necrotizing pneumonia and necrotizing fasciitis (13)(14)(15).This strain has unique genetic characteristics, including carriage of the type IVa staphylococcal cassette chromosome mec (SCCmec), and toxin and virulence factors including the Panton-Valentine leukocidin (PVL) genes, which may explain the severity of associated disease and the strain's ability to affect diverse populations (16,17).
In 2004, an outbreak of USA300 MRSA infections in the Calgary Health Region (CHR), Calgary, Alberta, was investigated, which centred on an urban population with histories of illicit drug use, homelessness or incarceration (6).Based on this investigation, other relevant studies and biological plausibility, four main hypotheses were identified for transmission of the USA300 strain in this population -residence in group living facilities (ie, homeless shelters and prisons), as suggested by outbreaks among soldiers and prison inmates (8,18); membership in a cocaine-or crack cocaine-using social network through drug use in unhygienic settings (ie, crack houses) or sharing crack pipes, as suggested by evidence for similar transmission of S aureus and our isolation of S aureus from crack pipes (6,19,20); injection drug use, through sharing needles or other equipment (11,21); and manipulation of USA300 skin infections (eg, squeezing, popping, or cutting of one's own or someone else's skin infection) (18,22).
During 2005, a cross-sectional prevalence study was conducted in this marginalized population.The objectives of the present study were to measure the prevalence of colonization or infection with the USA300 strain of MRSA, to test the hypotheses for transmission and to identify additional factors associated with USA300 colonization or infection.

Sample size
A sample size of 137 individuals allows for estimating colonization or infection with USA300 with 80% power and 95% confidence based on an expected 5% to 10% prevalence of USA300 in the study population (EpiInfo6 version 6.04, Centers for Disease Control and Prevention, USA).The present study attempted to recruit 300 individuals to improve the ability to detect associations between colonization or infection with USA300 and variables studied.

Recruitment
Between February and May 2005, participants were recruited at five types of study sites in Calgary -an outreach needleexchange van; homeless shelters; detoxification centres and residential substance treatment programs; an inner-city medical clinic; and new admissions to a local corrections facility.All individuals at the study sites were eligible, except at the medical clinic in which potential participants were assessed by a clinic nurse, and only included if they had a history of at least one of illicit drug use, homelessness or incarceration in the previous six months.There were no exclusion criteria.All participants provided verbal consent and were offered a $5 grocery store voucher to facilitate participation.Ethics approval for the present study was obtained from the Conjoint Health Research Ethics Board of the University of Calgary, Calgary, Alberta.

Data collection
A questionnaire was designed to test the hypotheses and to collect additional data, which was piloted with members of the study population.The questionnaire was transcribed into a personal digital assistant database application using Pendragon Forms 4.0 and Distribution Toolbox 4.0 (Pendragon Software Corporation, USA).Trained health care providers at each site administered the questionnaire.After instruction, participants self-collected one swab of both anterior nares and one swab of both axillae, and interviewers collected one to two swabs of any skin infections meeting clinical criteria for MRSA infectionwounds with purulent drainage; pustules, vesicles or boils with or without purulent drainage; or two of localized pain or tenderness, swelling, redness or heat at an infected site.In the midpoint of the study period, a nonrandom sample of used crack pipes was collected from a small number of participants.

Laboratory methods
The interior and exterior of the crack pipes were swabbed and plated on to a blood agar plate and a mannitol salt plate without oxacillin.Pipe and other swabs were inoculated into an overnight tryptic soy broth with 6.5% sodium chloride, and were then subcultured to mannitol salt agar without oxacillin.S aureus colonies were identified using standard microbiology techniques.Oxacillin-resistant S aureus isolates were identified using cefoxiten and oxacillin disk diffusion tests (23), and were confirmed as MRSA by testing for penicillin-binding protein 2a production (MRSA-Screen Corporation, Limited; Denka Seiken, Tokyo).The polymerase chain reaction assays were used to further confirm methicillin resistance (mecA gene), detect the presence of PVL (lukS-PV and lukF-PV genes) and classify according to SCCmec type (24,25).Typing was performed using pulsedfield gel electrophoresis (26), staphylococcal protein A (spa) typing (27), and multilocus sequence typing (MLST) (28).The identification of MRSA isolates matching the USA300 strain was based on the similarity of pulsed-field gel electrophoresis patterns and the presence of PVL, SCCmec type IVa, spa type t008 and MLST type ST8.

Data analysis
Participants were described as colonized if any one of a nasal, axillary or pipe swab was positive for USA300, and were described as infected if an infection swab was positive for USA300.Cases were defined as individuals colonized or infected with USA300, and all other participants were controls (while risk factors for colonization and infection may differ, this approach was chosen to identify participants with any exposure to the USA300 strain).Data were downloaded from personal digital assistants and exported to SPSS version 12.0.1 (Apache Software Foundation, USA) for analysis.Participants with missing questionnaire or laboratory data were excluded from the analysis.Univariate analysis was performed using Pearson's χ 2 or Fisher's exact test for categorical data, and independent samples t test or Wilcoxon's rank sum test for continuous data.ORs and exact binomial 95% CIs were calculated.The level of statistical significance was set at P<0.05.
To determine if results were unique to the USA300 strain, analysis was repeated defining cases by colonization or infection with any strain of MRSA and any strain of methicillinsusceptible S aureus (MSSA).For each analysis, the remainder of participants served as controls.

Recruitment
Overall, 274 individuals were recruited; three were excluded due to missing data (net participation 271).Recruitment by site was 105 (38.7%) from homeless shelters, 84 (31.0%) from the outreach needle-exchange van, 40 (14.8%) from the local corrections facility, 32 (11.8%) from detoxification centres and residential substance treatment programs, and 10 (3.7%) from the inner-city medical clinic.There were no significant differences in the distribution of USA300 cases by recruitment site (data not shown).

Laboratory results (Table 1)
Colonization or infection with the USA300 strain of MRSA was detected in 15 participants, for an overall prevalence of 5.5% (95% CI 3.1% to 9.0%); the prevalence of colonization or infection with the USA300 strain was 4.8% and 1.8%, respectively.Four crack pipes were collected for analysis.Of these, two of four (50.0%) were positive for the USA300 strain of MRSA (one of the positive pipes came from a participant negative for USA300 at other sites).All USA300 isolates were positive for SCCmec type IVa and PVL, and were spa type t008 and MLST type ST8.
The prevalence estimates for MSSA and any strain of MRSA were 36.5% (95% CI 30.8% to 42.6%) and 7.4% (95% CI 4.6% to 11.2%), respectively.The majority of individuals were exclusively colonized or infected with only one of the USA300 strain, MSSA or any MRSA strain, with the exception of three individuals who were colonized or infected with USA300 and also nasally colonized with MSSA.

Study participants (Table 2)
Overall, 149 of 271 (55%) study participants had a history of homelessness and 258 of 271 (95.2%) had a history of illicit drug use.The majority had been residents in the CHR for longer than one year (20 of 268 participants [7.5%] had been residents for less than three months).Cases were more likely to self-report a diagnosis of hepatitis C (OR 5.90; 95% CI 1.29 to 26.94).No other significant differences were identified.

Hypothesis testing (Table 3)
Residence in crowded or group living facilities, cocaine or crack cocaine use, borrowing crack pipes, using drugs at crack houses or injection-related behaviours were not found to be   associated with colonization or infection with the USA300 strain.Cases were not more likely to report self-manipulation of any skin infections; however, cases were more likely to report that others had manipulated their skin infections (OR 9.55; 95% CI 2.74 to 33.26).

Other variables (Table 4)
Several variables describing the drug-use environment and drug-use frequency were found to be associated with colonization or infection with the USA300 strain.Cases were more likely to report any drug use with a sex trade worker (STW) or as a STW with a client (OR There was no association found between colonization or infection with the USA300 strain and use of other illicit drugs or risk factors for nosocomial MRSA acquisition (data not shown).

Analysis using MRSA or MSSA
The analysis was repeated on variables listed in Tables 2, 3 and 4 using two other case definitions -colonization or infection with any strain of MRSA or with any strain of MSSA.Results using colonization or infection with MRSA were similar to

DISCUSSION
We found the prevalence of colonization or infection with the USA300 strain of MRSA in this marginalized urban population of the CHR to be 5.5%, approximately one year after its first appearance in this population (6).Most studies of MRSA population prevalence have not adopted a strain-specific approach; however, our prevalence of nasal colonization with any strain of MRSA (4.8%) is similar to estimates in other urban poor or injection drug using populations in North America (11,(29)(30)(31).By contrast, the prevalence of nasal colonization with MRSA in the general population of the United States is estimated at 0.84% (32).
Of our original hypotheses, we identified manipulation of skin infections as a potential explanation for transmission of the USA300 strain.This may reflect increased severity of disease because patients were also more likely to report seeking medical attention and self-treatment with antibiotics from old prescriptions for skin infections (which is a concern because this may promote further antibiotic resistance in this population).We did detect the USA300 strain on crack pipes belonging to study participants; however, borrowing crack pipes, use of cocaine or crack cocaine, or smoking drugs was not significantly associated with colonization or infection.While the laboratory results suggest the plausibility of our hypothesis, sharing crack pipes did not appear to be a major route of transmission in the study population.
We identified new hypotheses for transmission.While we did not directly measure sexual activity in our study, we identified that drug use in a sexual context (eg, with STW or casual sex partners) was associated with colonization or infection with the USA300 strain.Participants colonized or infected with MSSA were less likely to report drug use with or as a STW -the study variable most closely connected to sexual activity -suggesting that this association may be unique to the USA300 strain.While we have no local data or published reports of genital colonization or infection of STW with community-associated MRSA, sexual transmission has recently been proposed as a route of transmission for the USA300 strain in a case series among heterosexual couples (33) and in an outbreak among men who have sex with men (9), either through skin-to-skin contact or direct genital transmission.Evidence in support of direct genital transmission includes an association between condom use and a decreased risk of infection (9), and that S aureus -including the USA300 strain -can be found among oral, vaginal and anal flora in women (33)(34)(35).Based on these recently published reports, it is plausible that sexual activity may be contributing to transmission of the USA300 strain in the CHR and may explain these findings.Future studies to test this hypothesis are warranted.
As with S aureus, the social environment of drug use likely plays an important role in the transmission of the USA300 strain in the study population (19).Cases were more likely to report drug use with strangers and in hotels or motels, and more frequent and longer binge use of drugs.The significance of these findings is unknown.These behaviours may be associated with high-risk sexual behaviours (eg, multiple sexual partners) or with poor personal hygiene practices, an identified risk factor for transmission of USA300 (36).
There are limitations to the present study.Control measures instituted before the study (including recommendations for improved infection control in group living facilities) may have limited our ability to test our hypotheses.We only collected four crack pipes, and discordance between nasal colonization and drug paraphernalia positivity for S aureus has been demonstrated (20).Misclassification of outcome is possible because pipes were not systematically solicited; however, our use of nasal, axillary and pipe swabs may have increased our overall ability to detect colonization with the USA300 strain and other S aureus strains.In addition, self-collection of nasal and axillary swabs may have affected test results despite observation by study interviewers.Finally, there are statistical limitations -the small number of cases precluded multivariate analysis, the sample size may have limited our ability to detect small associations and we conducted multiple comparisons, increasing the possibility of type I error.
The findings of the present study may be useful to public health officials and investigators involved with the care of similar populations elsewhere in North America.While our a priori hypotheses were for the most part not confirmed in the present study, our data supports continuation of efforts in the present population at the CHR to promote the appropriate care of skin infections and to caution against the use of expired antibiotic prescriptions.We have also identified an intriguing new hypothesis which suggests that sexual activity may be contributing to transmission of this specific strain of MRSA in this population.Evidence is emerging to support this mode of transmission for the USA300 strain.We are currently planning further studies in this population to explore this possibility.
Outbreak-related prevalence of USA300 Can J Infect Dis Med Microbiol Vol 18 No 6 November/December 2007 359 Dis Med Microbiol Vol 18 No 6 November/December 2007 360 IVa; three participants), not classifiable (two participants); † Number of participants contributing swabs for analysis (swab type) or eligible for classification; ‡ Not mutually exclusive.MRSA Methicillin-resistant Staphylococcus aureus; MSSA Methicillin-susceptible S aureus

TABLE 3
Analysis of variables associated with a priori hypotheses