INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Staphylococcus aureus is a major cause of nosocomial infections that uses numerous virulence factors such as extracellular toxins and enzymes (13, 41). Studies from the Centers for Disease Control and Prevention (18) showed that during the period from 1985 to 1988, S. aureus was an important cause of postoperative wound infections and nosocomial bacteremia and pneumonia. Current therapy for infections due to S. aureus consists of removal of infected devices, surgical drainage of abscesses, and the use of antibiotics effective against the pathogen.

In the preantibiotic era S. aureus bacteremia resulted in 80% mortality (37); however, with the advent of antibiotics, the organism was reported to be susceptible to the earliest antimicrobial agents, the sulfonamides and penicillin. The widespread use of these antibiotics in the 1950s induced the predominance of -lactamase-producing resistant strains. To solve the problem, the -lactamase-resistant penicillins were developed, but reports of resistance to this new group started to appear in the 1960s in Europe (20) and in the 1970s in the United States (30).

The emergence of antibiotic-resistant strains of S. aureus is now considered to be a major problem in most hospitals. Virtually, all nosocomial strains produce a -lactamase(s) and thus are resistant to penicillins. Moreover, data from the Centers for Disease Control and Prevention indicate that throughout the United States there has been an increase in the frequency of methicillin-resistant S. aureus (MRSA) strains resistant to multiple antibiotics in both large and small hospitals (19). Thus far, all strains of MRSA have been susceptible to vancomycin, although certain strains have exhibited tolerance. It is possible, however, that vancomycin-resistant gram-positive cocci such as Enterococcus spp. (2, 40) and Staphylococcus haemolyticus (36) may transmit the gene(s) responsible for this phenotype to S. aureus, leaving few if any options for antimicrobial chemotherapy of infections caused by the organism.

Previous studies with different populations indicated that S. aureus ranks second only to Staphylococcus epidermidis as a cause of nosocomial bacteremia, accounting for 13% of all cases in 1988 and increasing to 19% of the total in 1990 (23). In this study we investigated the prevalence of MRSA among isolates causing infection in patients admitted to the intensive care unit (ICU) of the University of Maryland Shock Trauma Center (ST-ICU). Our goal was to determine the relatedness of the patient nasal isolates and infection-related isolates by studying their characteristics by phenotypic methods (capsular and antibiogram typing) and genotypic methods (pulsed-field gel electrophoresis [PFGE] of genomic DNA) and to appraise the ability of phenotypic markers to discriminate among strains whose identities have been confirmed by modern strain delineation techniques based on DNA content.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Study design and data collection. A total of 785 patients admitted to the Shock Trauma Center between March 1992 and March 1994 were screened for nasal carriage of S. aureus at admission. Carriage of S. aureus was determined by obtaining swab specimens from both anterior nares of patients at various times, as described below. The swabs were inoculated onto 5% sheep blood agar plates (BBL, Cockeysville, Md.), and the plates were incubated for 24 h at 35°C and observed for the growth of suspected S. aureus colonies. After the identity of the isolates was confirmed (6), they were stored at 70°C in freezer vials pending further analysis.

Capsular polysaccharide typing. Capsular serotyping of the coded isolates (no patient identifiers were used) was performed by direct cell agglutination and immunoprecipitation of cell extracts as described previously (21). Two different types of rabbit serum samples were used to identify the capsular type of each isolate: sera from animals vaccinated with a whole-cell vaccine (21) and monospecific sera against either type 5 or type 8 capsular polysaccharide conjugated vaccine prepared in rabbits (11, 12). Monoclonal antibodies to capsular types 5 and 8 produced and standardized by NABI (Rockville, Md.) were used to confirm the capsular serotypes of the isolates.

Antimicrobial susceptibility testing. The susceptibilities of all isolates to different antimicrobial agents were tested by the disk-agar method as standardized by the National Committee for Clinical Laboratory Standards (29). The following antimicrobial disks and concentrations were used: ampicillin, 10 µg; cephalothin, 30 µg; clindamycin, 2 µg; erythromycin, 15 µg; oxacillin, 1 µg; penicillin, 10 U; tetracycline, 30 µg; and vancomycin, 30 µg (Becton Dickinson Microbiology Systems, BBL, Cockeysville, Md.). The results were recorded after 24 h of incubation at 35°C. A standard strain of S. aureus (ATCC 25923) was used as a control. MRSA isolates were detected by two standard methods: by the disk-agar diffusion susceptibility method with 1-µg oxacillin disks, incubation at 35°C for 24 h, and inspection of the plates for colonies after 48 h (29) and by the ability of the isolates to grow on Mueller-Hinton agar supplemented with 4% sodium chloride and 6 µg of oxacillin per ml (MRSA Screen Agar; BBL) after incubation at 35°C for 24 h (25, 43).

-Lactamase production and inhibition tests. The ability of the isolates to produce -lactamase was determined as recommended by the National Committee for Clinical Laboratory Standards (29). Confirmatory tests were performed by the disk-agar diffusion method with amoxicillin-clavulanic acid (AMC 30) and ampicillin-sulbactam (SAM 20) disks (Difco Laboratories, Detroit, Mich.), as recommended previously (29).

Genomic DNA Analysis by PFGE. The genomic DNA was extracted from logarithmic-phase S. aureus cultures grown in brain heart infusion broth (BBL) as described previously by Murray et al. (27) and modified by Wanger et al. (42). The DNA size standards used were a bacteriophage lambda ladder consisting of concatemers starting at 48.5 kbp and increasing to approximately 1,000 kbp (Bio-Rad Laboratories, Hercules, Calif.). The gels were stained with ethidium bromide, rinsed, and photographed under UV light.

Statistical analysis. Tests of statistical significance were performed by the use of the statistical packages SAS for Windows (SAS Institute, Cary, N.C.) and Sigma-Stat for Windows (Jandel Scientific, San Rafael, Calif.). Means are reported as X ± 2 standard errors. Differences in proportions were tested for significance by the chi-square test with Yate's correction or by Fisher's exact test (for frequencies of less than 5). Differences in means were tested for significance by Student's t test. All P values are two-sided.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Demographics. A total of 785 patients were enrolled in the study, with the majority being males (73.3%). Most of the patients were Caucasians (60%), followed by blacks (36%) and people of other races (4.0%). The median age of all patients enrolled in the study was 32 years (age range, 12 to 92 years). The mean age of the patients who developed a nosocomial infection was 46.2 ± 19.6 years, while that of the patients who did not develop any infection was 37.2 ± 18.4 years. The difference between the mean ages was found to be significant by the t test (P = 0.001). Of the 776 patients with adequate nasal specimen cultures, 150 (19.4%) were positive for S. aureus. Fifty-six patients (7.2%) were reported to have acquired 65 nosocomial S. aureus infections, namely, bacteremia (36; 55.4%), pneumonia (21; 32.3%), and wound and other infections (8; 12.3%). Seven of the patients developed multiple infections.

Capsular and antibiogram typing. Among nasal carriers of S. aureus at admission, the prevalent capsule type was type 8 (65.9%). Among the isolates causing infections, however, there was an almost equal prevalence of both capsular types 5 and 8 (43.1 and 44.6%, respectively). When the infections were classified by the capsular type of the organisms causing the infection, it was found that among the isolates causing pneumonia, S. aureus strains of type 8 were prevalent (61.9%), while among the isolates causing bacteremia and wound infections, type 5 isolates were more common. Some isolates noted not to express either the type 5 or the 8 capsule (12.3%) were isolated from patients with bacteremia and wound infections but not from patients with pneumonia (28).

Production of -lactamase and methicillin resistance. S. aureus strains capable of producing -lactamase were, as expected, prevalent (92.1%). Only 3 of the 20 isolates found not to be -lactamase producers were infection-related isolates, constituting only 4.6% of the infection-related isolates.

DNA fingerprinting. A total of 74 S. aureus isolates obtained from the 60 patients who suffered from infections were fingerprinted by PFGE. In addition to the infection-related isolates, the total represents nasal S. aureus isolates from positive cultures of nasal specimens taken from infected patients upon admission, infection, or discharge, when available. Of these, 32 (43.2%) expressed a type 5 capsule, 33 (44.6%) expressed a type 8 capsule, and 9 (12.2%) were not typed by the monoclonal capsular type 5 and capsular type 8 antibodies. It was notable that the percentages of retrieval of the different capsular serotypes were not changed after adding the data for the types of the few nasal isolates from infected patients.

View larger version (101K): [in this window] [in a new window]   FIG. 1.   Contour-clamped homogeneous electric field characterization of isolates by PFGE. Chromosomal DNA was digested with the endonuclease SmaI and separated by PFGE and was then photographed under UV light. The different patterns shown were obtained for the isolates that expressed the type 8 capsular polysaccharide. Lanes , size markers.

View larger version (64K): [in this window] [in a new window]   FIG. 2.   Contour-clamped homogenous electric field characterization of isolates by PFGE. Chromosomal DNA was digested with endonuclease SmaI and separated by PFGE and was then photographed under UV light. The different patterns were obtained from the isolates that expressed the type 5 capsular polysaccharide (patterns S to h) and the patterns obtained from the isolates denoted as being non-type 5, non-type 8 strains (patterns i to k). Patterns S to V represent the patterns obtained for the MRSA strains. Lane , size markers.

View larger version (84K): [in this window] [in a new window]   FIG. 3.   Contour-clamped homogeneous electric field characterization of isolates from bacteremic patients by PFGE. Chromosomal DNA was digested with endonuclease SmaI and separated by PFGE and was then photographed under UV light. The different patterns are shown in groups for different patients to determine strain delineation. Patterns designated a and b are those for isolates from blood (pattern a) and nasal swab at infection (pattern b) from patient 1; patterns designated 1, 2, 3, and 4 refer to S. aureus isolates from patients with nasal carriage at admission (pattern 1), infection (bacteremia) (pattern 2), infection (gall bladder) (pattern 3), and nasal carriage at discharge (pattern 4), respectively. Lanes A, B, C, and three isolates from blood, nasal swab at infection, and nasal swab at discharge, respectively, from patient 3; Lane , size markers.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study was designed to determine the prevalence of MRSA isolates among infection-related isolates from patients admitted to the ST-ICU, characterize the infection-related isolates phenotypically and genotypically to determine the genetic relatedness of the organisms involved, and evaluate whether the different typing methods are efficient tools in routine epidemiological investigations.

Most of the S. aureus isolates retrieved in this study (91.0%) expressed either capsular polysaccharide type 5 or capsular polysaccharide type 8 (Table 3), in compliance with previous reports (38, 42). The relationship between nasal carriage of S. aureus and subsequent infection has been well established by many investigators (7, 28, 33, 41, 44). Notably, the type 8 capsules were expressed more in the isolates from the nares of patients at admission to the hospital, while no significant prevalence of any of the types was noted among the infection-related isolates. When the infections were classified by the capsular type of the organism causing the infection, the type 5 capsular type was noted to predominate in patients with bacteremia, while capsular type 8 was the prevalent type in patients with pneumonia (Table 3).

The antibiogram typing of the isolates revealed that antibiogram type I (resistant only to penicillin G and ampicillin) was predominant for 73.2% of all the nasal and infection-related isolates (Table 2). This result reflects the predominance of relatively susceptible strains in our setting. The retrieval of more resistant strains among the infection-related isolates was expected (Table 3). An interesting finding was not only that some isolates with specific antibiogram types expressed a specific type of capsule (all type III isolates had a type 5 capsule) but also that there existed a significant relationship (P < 0.0001) between the recovery of certain antibiogram types from isolates with a specific capsular type (Table 4).

Of the isolates tested, 6.3% were found to be resistant to methicillin. The results presented in Table 5 indicate that most of the MRSA isolates are infection-related isolates, implying that MRSA is present in a reservoir in the hospital environment. The finding that only 16.7% of the bacteremias in this study were caused by MRSA reflected a proper hospital management plan for MRSA, because MRSA comprised 20% of all isolates causing bacteremia at the University of Maryland Hospital in 1988, a proportion that increased dramatically to 50% in 1990 (23).

Our study disclosed a significant relationship between the property of methicillin resistance and the possession of a specific type of capsule. All our MRSA isolates (100%) possessed a type 5 capsule, in agreement with previous reports noting the prevalence of type 5 capsules in MRSA isolates (5, 14, 42).

The finding that most infection-related S. aureus isolates belonged to capsular polysaccharide type 5 or 8 did not signify whether these strains were derived from resident (patient), commensal isolates or were the result of the nosocomial spread of a single endemic strain. To perform proper epidemiological typing and delineation of genetic relatedness, a number of methods were introduced and used effectively by other investigators. These methods included bacteriophage typing (22, 44), plasmid pattern analysis (4, 8), whole-cell protein profile analysis (9), immunoblotting profile analysis (15), ribotyping (10, 39), restriction endonuclease analysis of plasmid DNA (17, 34), and PFGE of restriction enzyme-digested genomic DNA (1, 24, 32). We chose to use PFGE for our study because the exceptional properties of PFGE have, with time, turned it into a favored epidemiological tool (3, 27, 31). PFGE detects the distribution of restriction sites throughout the chromosome, giving fingerprints that reflect the structural organization of the bacterial chromosome (31). This distribution of restriction sites, which varies because of ancestral strain-to-strain mutational differences and because of the variations in the gene content of the staphylococcal chromosome, give the method its acknowledged discriminatory ability.

The DNA components of all of the isolates from infected patients were fingerprinted by PFGE. Each of the genotypes obtained by PFGE of DNA for the S. aureus isolates tested was given a different designation, without consideration of band differences and relatedness, in order to present large numbers of patterns, as was previously done by others (31, 35). Except for the few exceptions mentioned previously, the 36 patterns obtained were clearly separated on the basis of capsular type. The isolates that expressed the type 8 capsular polysaccharide were found to have to a specific set of patterns (Fig. 1), while those that expressed the type 5 capsular polysaccharide had a different set of patterns (Fig. 2). The isolates that did not produce either the type 5 or the type 8 capsule were found to have one of three patterns (patterns i, j, and k in Fig. 2). An interesting result was the finding that the MRSA isolates that were found to produce a type 5 capsule had one of only four possible patterns that were not shared by any of the other isolates that produced type 5 patterns (patterns S to V in Fig. 2), with pattern S being the most prevalent (56.3%). This finding confirmed the clonal relatedness of the MRSA strains isolated in the study.

Figure 3 presents the PFGE patterns for MRSA isolates from three different bacteremic patients. The data demonstrate the clonal relationships of these isolates. Isolates from the same patient had identical DNA patterns. Furthermore, isolates from the different patients had similar patterns (a difference of only one or two bands). These data establish the presence of a resident MRSA population endemic in our setting, probably as nosocomial colonization of hospital personnel. This finding is a reasonable incentive for initiating an infection control study that aims at detecting and treating MRSA carriers in shock trauma units.

PFGE demonstrated typeability, reproducibility, and an ability to discriminate between strains. These properties demarcate an excellent tool for epidemiological studies; however, the only problem with the routine use of this method is the difficulty in performing the technique and the need for expensive instruments and chemicals.

The trend toward developing precise molecular biology-based methods has diverted attention from the development of simpler and more common typing methods (like capsular and antibiogram typing techniques). Attempts were made to use a combination of antibiogram and capsular typing methods, both of which are simple techniques that are routinely performed in many hospitals, to assess the possibility of using them as discriminatory tools for epidemiological studies. For practical purposes, the epidemiological conclusions based on the results of the tests based on the use of the previous combination of methods correlated well with those based on the use of the DNA fingerprinting by PFGE, especially with regard to the MRSA isolates.

The previous results indicated the possibility of using the two typing methods in combination in order to provide a more effective tool for epidemiological studies. Because antibiogram typing has been used as an effective tool in associating isolates to outbreaks in hospitals (42) and because both capsular typing and antibiogram typing are easy and reproducible techniques, we suggest that a combination of these simple typing methods be used for the routine processing of clinical specimens in hospitals, because the results obtained may provide valuable epidemiological information. However, we suggest increasing the number of antibiotics used in antibiogram typing, because this will increase the technique's potential for discrimination.

We conclude that (i) the nosocomial infection isolates from the shock trauma patients were not found to belong to a single clone, although the predominance of one methicillin-resistant genotype was noted. (ii) Most S. aureus infections were caused by organisms relatively susceptible to antibiotics, although a well-defined multidrug resistant hospital strain of MRSA was present in our setting. (iii) For practical purposes, we suggest the use of a combination of the results of capsular and antibiogram typing as an epidemiological marker, based on the fact that in our study the combined results correlated well with those based on the use of the DNA fingerprinting by PFGE.