Epidemiologic Investigation

We conducted a prospectively enrolled case–control investigation at 3 emergency departments and 3 urgent care clinics in Georgia from September 6 through October 31, 2005. Two low-cost urgent care clinics that serve primarily low-income populations and all emergency departments in a 3-county area were included in an attempt to capture sites where methamphetamine users might seek medical care for SSTI. The third urgent care clinic was affiliated with one of the participating hospitals but was located in a neighboring county. According to the 2000 US Census, 43.9% of the population of these 3 counties lives in rural areas (13).

We defined a case-patient as a person >12 years of age with a laboratory culture–confirmed SSTI who came to a participating emergency department or clinic for treatment during the investigation period. Clinicians at participating institutions identified patients with culturable SSTIs and were asked to incise, drain, and culture all infected skin and soft tissue. Patients with SSTIs that were not culturable, such as simple cellulitis, were not included. Patients whose primary language was not English were enrolled if they could speak English fluently enough to answer survey questions. Patients with new or recurrent SSTI could also be enrolled; however, we excluded patients who had previously enrolled in the investigation.

Controls were patients >12 years of age with no current skin infection who were frequency matched by investigation site at a rate of 3 controls to 1 case-patient with MRSA infection. Controls were excluded if they reported a current skin infection or if infection was identified on physical examination. Persons could be enrolled as control patients if illness was minor and comparable in severity to an SSTI. For example, patients with major trauma and critically ill patients were excluded from control selection.

Upon seeking treatment, patients voluntarily consented to be interviewed by trained staff of the participating healthcare facilities, local public health departments, or CDC to identify SSTI case-patients. To ensure as much privacy as possible, the interviews were usually conducted in the patient’s room with no family or friends present. The interview survey contained questions about demographics, clinical history, and potential risk factors for SSTI. Each patient was asked a specific question about methamphetamine use: “In the past 3 months, have you used methamphetamine (crystal meth or meth)?” If the patient answered yes, 2 follow-up questions were asked: 1) “How did you take methamphetamine?” with the choices “smoked or inhaled,” “injected,” or “swallowed or took pills,” and 2) “Have you shared drug equipment or rinse water with anyone else, including a significant other?” To identify healthcare exposure, patients were asked whether they had had surgery or dialysis or if they had stayed overnight in a hospital within the previous 3 months. All patients, and their parents if the patients were <18 years of age, were given a letter explaining the investigation and asked to give verbal informed consent to enroll in the investigation. This investigation was deemed exempt from review by the CDC Institutional Review Board because it was part of a public health response by CDC and the Georgia Division of Public Health. We examined trends in S. aureus skin infections and cultures at one of the main emergency departments in our investigation by reviewing billing codes and laboratory microbiology reports from January 2004 through September 2005, the start of the case–control survey investigation.

Laboratory Investigation

Specimens were obtained from at least 1 infection site in all case-patients. Staff at all 3 hospital emergency departments and the urgent care clinic affiliated with the respective hospital collected cultures and performed antimicrobial drug susceptibility testing at their facility. Two low-cost, urgent care clinics sent all cultures to CDC for culture and antimicrobial drug susceptibility testing. All 6 investigation sites sent both MRSA and methicillin-susceptible S. aureus (MSSA) isolates to CDC for further characterization.

All available isolates from methamphetamine users and a random sample of isolates not related to methamphetamine use from each of the 6 investigation sites were tested at CDC for antimicrobial susceptibility by the Clinical and Laboratory Standards Institute broth microdilution method (14). We tested for susceptibility to chloramphenicol, clindamycin, daptomycin, doxycycline, erythromycin, gentamicin, levofloxacin, linezolid, oxacillin, penicillin, rifampin, tetracycline, trimethoprim-sulfamethoxazole, and vancomycin. In addition, we performed the cefoxitin disk diffusion test to predict mecA-mediated resistance to oxacillin (14) and the D-zone test for inducible clindamycin resistance (15). Isolates were also tested by using PCR for genes encoding the staphylococcal cassette chromosome mec (SCCmec) resistance complex, Panton-Valentine leukocidin (PVL) cytotoxin, and toxic shock syndrome toxin (16). Chromosomal DNA was analyzed by pulsed-field gel electrophoresis (PFGE) after digestion with SmaI restriction endonuclease (17). The relatedness of PFGE patterns in different isolates was defined by using Dice coefficients and 80% relatedness by the unweighted pair-group method with arithmetic averages (Applied Maths, BioNumerics, Austin, TX, USA) (18).

Statistical Methods

We conducted univariate analysis of the data to describe patient demographics and compared binary and categorical variables with the χ² test; continuous variables were compared by using the t test with unequal variances. We evaluated risk factors for MRSA SSTIs by using conditional logistic regression with stratification by investigation site. Risk estimates were adjusted for age (categorized as <18 years, 19–34 years, 35–64 years, and >65 years), sex, and race (categorized as white and nonwhite) because they were potential confounding variables.

Epidemiologic Investigation

We identified 119 case-patients with skin infections in the investigation. MRSA was isolated from 81 (68.1%) of the skin and soft tissue cultures, MSSA from 20 (16.8%), and bacteria other than S. aureus from 18 (15.1%) (Table 1). Compared with controls with no skin infection, a higher percentage of patients with MRSA SSTIs were male (p<0.001). The proportion of patients that were male did not differ significantly between controls and patients with either MSSA or non–S. aureus SSTIs (p = 0.67 for MSSA, p = 0.12 for non–S. aureus) or between patients with MRSA and MSSA SSTIs (p = 0.16).

Fifteen patients who reported recently using methamphetamine were identified: 8 with MRSA SSTIs, 2 with MSSA SSTIs, and 5 controls. Half (8 [53.3%]) of the methamphetamine users were male. Ten percent of patients with MRSA skin infections (8/81) reported using methamphetamine in the past 3 months, significantly more than the 2% of controls (5/283) who reported this behavior (p<0.001). After adjusting for age, sex, and race, we determined that patients with MRSA SSTI were significantly more likely to have recently used methamphetamine than were controls (adjusted odds ratio [AOR] 5.10, 95% confidence interval [CI] 1.55–16.79) (Table 2). Of the 8 methamphetamine users with MRSA SSTIs, most (5 [62.5%]) smoked or inhaled the drug. Only 1 (12.5%) injected the drug, and 1 (12.5%) took the drug orally. For 1 methamphetamine user with MRSA SSTI, we could not determine the route of drug administration. Of the 8 methamphetamine users with MRSA SSTIs in our investigation, 2 (25.0%) reported sharing drug equipment or rinse water with other persons; we did not have information on drug-sharing behavior for 1 methamphetamine user with a MRSA SSTI.

In our study population, having had a skin infection within the previous 3 months was the factor most strongly associated with current MRSA skin infection (AOR 7.92, 95% CI 4.10–15.28) (Table 2). Recent sexual contact with someone with a skin infection was also a significant risk factor for MRSA skin disease (AOR 5.42, 95% CI 1.68–17.50), when compared with recent sexual contact with a person without a skin infection. Frequent skin-picking behavior was independently associated with MRSA SSTI (AOR 2.53, 95% CI 1.22–5.23). Crowded living conditions, defined as >1 person per bedroom, had a small but significant association with MRSA SSTI (AOR 1.78, 95% CI 1.004–3.15).

Only 10% of MRSA case-patients had healthcare-associated risk factors traditionally associated with MRSA infection, namely, recent hospitalization, surgery, or dialysis. Additional factors not significantly associated with MRSA SSTI in our study population included use of antimicrobial agents in the previous 6 months, recent stays in a jail or prison, bathing less than daily, history of diabetes or liver disease, recent tattoo or body piercing, and participation in contact sports in the previous 3 months. In addition, very few or no patients were HIV positive (2 [0.5%]), homeless (0), or recently had sex with someone of the same sex (7 [1.6%]), suggesting that none of these were significant risk factors for MRSA SSTI in this population.

Figure 1

Figure 1. Number of Staphylococcus aureus skin infections at a southeastern United States emergency department, January 2004–September 2005.

The number of visits for S. aureus skin infections at one of the main emergency departments in our investigation increased from ≈1 per 1,000 emergency department visits to 12 per 1,000 visits over the 20 months leading up to the investigation (Figure 1). This emergency department accounted for 46.2% of all study participants in our investigation. Over the same period, MRSA infections increased from 2 to 38 per month in the same emergency department. Most emergency department S. aureus cultures for both SSTIs and non-SSTIs were resistant to methicillin, with the prevalence of methicillin-resistance remaining stable over the same 20-month period (median 82%, range 50–100%).

Laboratory Investigation

MRSA (n = 32) and MSSA (n = 13) isolates tested were commonly susceptible to clindamycin, daptomycin, doxycycline, gentamicin, levofloxacin, linezolid, rifampin, tetracycline, trimethoprim-sulfamethoxazole, and vancomycin (Table 3). None of the MRSA isolates and only 1 (7.7%) of the MSSA isolates had inducible clindamycin resistance. MRSA susceptibility patterns of isolates from methamphetamine users and nonusers were similar, except that both MRSA isolates susceptible to erythromycin were found in those who did not use methamphetamine. The MSSA isolate from a methamphetamine user was susceptible to all but penicillin.

Figure 2

Figure 2. Dendrogram of pulsed-field types for methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA) isolated from methamphetamine users.

We detected genes for PVL in all MRSA isolates and 5 (41.7%) MSSA isolates; however, the MSSA isolate from a methamphetamine user did not carry the PVL locus. All available MRSA isolates from 6 methamphetamine users and 21 nonusers of methamphetamine had type IV SCCmec resistance complex and were PFGE type USA300. Most of the MRSA isolates were a single strain, PFGE type USA300-0114 (4 [66.7%] were methamphetamine users, 15 [71.4%] were non-methamphetamine users) (Figure 2). One third (33.3%) of MRSA isolates from methamphetamine users and one fifth (19.0%) of MRSA isolates from non-methamphetamine users were variants of USA300-0114, such as USA300-0047.