JSTOR

Methods

All inpatients with CA‐MRSA infections who presented to St. John Hospital and Medical Center, a 600‐bed urban academic medical center in Detroit, from July 1, 2003, until April 30, 2004, and who were seen by the Infectious Diseases Service were included as cases. Both adult and pediatric cases were identified. An MRSA infection was defined as clinical evidence of infection and identification of MRSA from a sterile body site or from drainage obtained at a nonsterile site with evidence of infection and no other pathogens identified. Patients were excluded from the study if they had received systemic antistaphylococcal antibiotic therapy within the year before presentation, because this may be a factor in colonization with resistant organisms. Other exclusion criteria included dialysis, an indwelling catheter, a prior positive culture for MRSA, or residence in a long‐term care facility. The St. John Hospital and Medical Center Investigational Review Board approved the study, and procedures were followed in accordance with their ethical standards. Data were collected by patient interview and medical record review, including demographics, underlying conditions, site of infection, treatment, and outcome.

The clinical microbiology laboratory used the Vitek 2 method to identify organisms. Susceptibility testing was performed by the microdilution broth method, and inducible clindamycin resistance was detected by D‐test following Clinical and Laboratory Standards Institute guidelines.¹⁸ Thirteen CA‐MRSA isolates obtained during a community outbreak from 1980 were included in the genetic analysis.³ Pulsed‐field gel electrophoresis (PFGE) was performed using SmaI restriction endonuclease¹⁹ and interpreted according to standard guidelines.²⁰ Dendrogram analysis was performed using Dendron 3.0 software (Solltech) and analyzed using the Dice coefficient. A sample of USA 300 and 400 was obtained and compared with our isolates, using PFGE.²¹ The SCC mec types were determined by a polymerase chain reaction–based multiplex assay described elsewhere.²² Polymerase chain reaction amplification of the cassette chromosome recombinase (ccr) gene was performed using the primers described by Okuma et al.²³ Positive controls for each assay included S. aureus HPV 107/ATCC BAA‐44 (type I), NYBK 2464/ATCC BAA‐41 (type II), HUSA 304/ATCC BAA‐39 (type III), and HDE 288/ATCC BAA‐42 (type IV). The mec A gene served as an internal control for the multiplex assay. Polymerase chain amplification of the genes encoding PVL luks‐PV and lukf‐PV was performed using primers described elsewhere.¹⁰ The positive control for the PVL assay was ATCC 49775, and sterile water served as the negative control. Samples from the PFGE groups A1, C1, D, and F were typed by multilocus sequence typing (MLST) as previously described by Enright et al.²⁴ The only change to the procedure was the use of the Visible Genetics Open Gene System for sequencing rather than the ABI Prism 377.

Results

Patient Characteristics

A total of 23 patients with CA‐MRSA infection meeting our case definition were prospectively identified from July 1, 2003, until April 30, 2004. The case rate was 0.9 per 1,000 nonelective admissions during this period. Before this period, the occurrence of CA‐MRSA was rare. All patients lived in the Detroit area; however, this study did not show any epidemiologic association among the patients. Patients’ home addresses were from a total of 11 different ZIP codes, reflecting the population served by our hospital. Twelve patients were equally divided between 2 of these ZIP codes. The mean age of patients was 40.3 years (median, 41 years; range, 7 months to 66 years); 2 patients were younger than 18 years and 2 were older than 65 years. Most patients (70%) were African American males. Demographic characteristics and risk factors for MRSA infection are given in Table 1. Among those with underlying medical conditions, 7 (30%) had diabetes mellitus, 3 (13%) had chronic obstructive pulmonary disease, and 2 (9%) had human immunodeficiency virus infection.

Table 1.  Demographic Characteristics and Risk Factors of Patients with Community‐Associated Methicillin‐Resistant Staphylococcus aureus Infection

Open New Window

Susceptibility Data

Susceptibilities of the 23 CA‐MRSA isolates are provided in Table 2. The pattern of resistance is similar to that shown in other reports, with a high rate of resistance to erythromycin and a higher proportion of isolates that are susceptible to other antibiotics, compared with hospital‐associated strains of MRSA.^13‐17 Among the fluoroquinolones, moxifloxacin showed superior activity compared with other agents. Although only 2 isolates exhibited clindamycin resistance, an additional 2 isolates (9%) had positive D‐test results, indicating inducible macrolide‐lincosamide‐streptogramin B resistance.

Table 2.  Antibiotic Susceptibility of 2003‐2004 Community‐Associated Methicillin‐Resistant Staphylococcus aureus Isolates

Open New Window

Molecular Typing

Among the 2003‐2004 isolates, 4 strains were identified (Figures 1 and 2). Strain A included 16 identical isolates (A1) and 4 isolates that were identical and differed from A1 by one band (A2). Strain A was identical by PFGE to USA 300, which is a frequently identified strain associated with CA‐MRSA infection.²¹ Strain C included 2 isolates (C1 and C2) that differed by 2 bands and were unrelated (more than 7 bands different) from strain A. One isolate (strain D) was unrelated to either strain A or C. The 13 isolates from 1980 represented a single strain (strain F) that was unrelated to present strains. Isolates from strain A all possessed the SCC mec type IV determinant and PVL genes. Strains C1 and D were SCC mec type II; C2 was untypable by the multiplex procedure but was type IV according to the method of Okuma et al.²³ All isolates from strain F were SCC mec type IVA by the multiplex method, but we were unable to assign a ccr type when using the primers of Okuma et al. No evidence existed of the PVL genes in strains C, D, and F.

(32 KB)

Figure 1.  Dendrogram analysis of community‐associated methicillin‐resistant Staphylococcus aureus isolates. Strains A1, A2, C1, C2, and D represent isolates from the years 2003‐2004; strain F represents isolates from 1980. The scale at the left indicates the degree of relatedness. NT = not typable; PVL = Panton‐Valentine leukocidin gene; SCC = staphylococcal cassette chromosome.

Open New Window

(153 KB)

Figure 2.  Pulsed‐field gel electrophoresis patterns of representative community‐associated methicillin‐resistant Staphylococcus aureus strains. Lane 1, USA 300; lane 2, strain A1; lane 3, strain A2; lane 4, USA 400; lane 5, strain C1; lane 6, strain C2; lane 7, strain D; lane 8, strain F (1980 isolate); lane 9, NCTC 8325 (global standard); and lane 10, λ ladder.

Open New Window

Four distinct allelic profiles were identified using MLST (Table 3). The sample representing strain A1 was sequence type 5 and matched the MLST and pulsed‐field profile for USA 300. This was the predominant strain in our study, representing 16 of 23 of our isolates. Our C1 strain did not match all alleles for any of the sequence types available in the database. Strain D was sequence type 83 and MLST strain CHL5. The CHL5 strain was isolated in Chile in 1997 and was listed as a hospital‐acquired MRSA. The 1980 sample (strain F) was assigned sequence type 74, MLST strain MR108, which corresponds to a community‐acquired MRSA Japanese isolate from 1981.

Table 3.  Results of Multilocus Sequence Typing of Community‐Associated Methicillin‐Resistant Staphylococcus aureus Isolates

Open New Window

Among the 2003‐2004 isolates, patients infected with strains C1, C2, and D all had diabetes, whereas 4 (21%) of 19 patients infected with strain A and A1 had diabetes mellitus. In addition, the patient infected with strain C2 had a healthcare worker in the household, and the spouse of the patient infected with strain D is a resident of a long‐term care facility.

Discussion

To our knowledge, this is the first report of CA‐MRSA infection occurring in our community since the first several case series were reported in the early 1980s.^3,4,25 The new cases of CA‐MRSA infection that we have seen are similar to many other recent studies in which patients with limited healthcare exposure present with skin and soft‐tissue infections.^12‐14^,16,17,26‐29 In addition, the CA‐MRSA isolates were consistent with previous studies that showed susceptibility to numerous other antimicrobials except erythromycin. As in other studies, most of our isolates (87%) were identical by PFGE^12‐14^,16,17,27‐29 and carried the SCC mec type IV determinant^15,26‐28 and the PVL genes.^16,17,27,29

The traditional risk factors for CA‐MRSA infection (injection drug use, prior hospitalizations and antibiotic use, chronic medical conditions) are infrequent in our present study. A number of epidemiologic exposures have been linked to recent outbreaks of CA‐MRSA infection, including jails,^28,30 saunas,²⁹ and daycare settings.³¹ Given the high rate of unemployment and the wide range of ZIP codes that patients gave as home addresses, the likelihood of a single common exposure is small. In addition, because only 1 of 20 patients queried had children at home (in addition to the 2 pediatric cases), transmission through a daycare setting is also unlikely. The preponderance of cases from individuals with low socioeconomic status (44% uninsured and unemployed) is consistent with other series,^5,12,14 although this association remains unexplained.

The spread of methicillin resistance among community isolates has only recently been linked to the SCC mec type IV element.^15,32 The presence of this genetic element among our previously described CA‐MRSA isolates indicates that a similar pattern of methicillin resistance may explain the early outbreaks as well. The original source of this element remains unknown, although it has been suggested to originate from coagulase‐negative staphylococci.³³ The characteristic PVL genes that are seen among present isolates were not detected in the earlier outbreak. Thus, although earlier outbreaks may have also been related to the SCC mec type IV element, PVL production appears to be unique to more recent outbreaks. The PVL toxin has been specifically linked to necrotizing soft‐tissue infections and pneumonia, typically among community‐acquired S. aureus infections.³⁴ The reason for the absence of PVL among hospital‐acquired staphylococcal infections is not well understood.

As with most series that involve CA‐MRSA soft‐tissue infections, patients generally responded well to combination incision and drainage and adjuvant antibiotics. The duration of intravenous antibiotic therapy with activity against MRSA was generally short followed by a more prolonged course of oral therapy. The most frequently used oral agent was trimethoprim‐sulfamethoxazole, which was effective in all but one patient and has been shown to be effective against even deep‐seated MRSA infections.³⁵ The optimal therapeutic regimen remains unknown, and several reports have described patients whose CA‐MRSA skin and soft‐tissue infections were cured with incision and drainage alone.^17,36,³⁷ A number of antibiotics are approved for the treatment of MRSA infections, including vancomycin, linezolid, and daptomycin.

Isolates from the United States are generally susceptible to clindamycin, although there have been reports of resistant isolates from Taiwan.^17,36 Isolates should be screened for macrolide‐lincosamide‐streptogramin B resistance before clindamycin use in MRSA infections.³⁸ Our in vitro susceptibility data suggest that newer fluoroquinolones (especially moxifloxacin) may be therapeutic options, but they have not been proven in clinical studies. In addition, specific concerns regarding the use of these agents include cost and the rapid development of ciprofloxacin resistance among nosocomial MRSA isolates after the introduction of ciprofloxacin.³⁹

One limitation of our study is that cases were identified by referral through the Infectious Diseases Service. Thus, the number of cases and the rate of CA‐MRSA infection might have been higher if all patients with MRSA infections who presented to the institution were reviewed. The use of vancomycin is restricted to the Infectious Diseases Service; thus, it is likely that most cases were identified. Another limitation is the inclusion of 4 individuals with a prior hospitalization within the previous year. All 4 individuals were infected with strain A, suggesting that their exposure to MRSA occurred in the community rather than the nosocomial setting. In addition, they were admitted more than 1 month earlier at 4 different healthcare institutions for brief periods (less than 1 week) for noninfectious conditions.

In conclusion, the recent occurrence of CA‐MRSA infections in our community is similar to recent reports in other communities. The infections are due to predominantly one strain that is highly susceptible to non—β‐lactam agents, carries the PVL genes, and is genetically identical to USA 300. The presence of the SCC mec type IV determinant among early and recent CA‐MRSA infections suggests that the development of MRSA in the community may have occurred previously in a similar fashion. The presence of a strain from Japan in 1981 with an MLST pattern that is identical to our 1980 isolate suggests that the early outbreaks of CA‐MRSA infection may have also represented spread of a single or small number of clones to a wide geographic distribution.