JSTOR

Methods

Design. We performed a cross‐sectional study carried out from October 2001 to May 2003.

Setting. We performed the study at Germans Trias i Pujol Hospital, a 600‐bed referral center serving a population of 780,000 inhabitants and surrounded by several acute care hospitals and chronic‐care centers. MRSA has been endemic in this center since a nosocomial outbreak originating in the intensive care unit (ICU) in 1990 that spread widely throughout the hospital.^4,5

Target population. Study patients included patients who were admitted from the emergency department or had a scheduled admission to any hospital ward.

Study population. Individuals were randomly selected from the target population. The following groups of patients were excluded: (1) patients who, at the time of microbiological sample collection, had been in the hospital for more than 24 hours; (2) patients who had previously been diagnosed with MRSA colonization or infection in the hospital or in another healthcare center; and (3) patients admitted to the obstetrics or pediatrics departments (because of the typical lack of cases in these departments).

A sample size of 1,100 patients was calculated with an estimated prevalence of 3%, a significance level of 5%, and a precision level of 1%. The hospital has approximately 15,000 admissions annually, excluding admissions to the obstetrics or pediatrics departments. It was estimated that 40‐50 admissions would occur each day, 68% of which would be emergency cases; the remainder would be scheduled admissions. The patient selection period lasted for 18 months, and 15 patients were included per week. Samples were collected once per week, for logistical reasons.

A total of 15 patients were randomly selected, a group that included patients admitted to the emergency department ( ) and patients with scheduled admissions ( ). If the patient selected did not satisfy the inclusion criteria, the next patient listed was chosen.

Data collection. Clinical and epidemiological data were collected from each subject (after obtaining oral consent) using a questionnaire during an interview with the patient or a relative, or by reviewing the clinical history, if necessary. MRSA carriage was reported for patients who had been hospitalized less than 24 hours and had an MRSA‐positive nasal swab sample and/or an ulcer or wound culture positive for MRSA. The variables studied were as follows: age; sex; location prior to admission (home, long‐term care center, or another hospital); admission to another hospital, ICU, or long‐term care facility in the past year; community contact (ie, occupational or household exposure); history of outpatient visits in the past year; chronic disease; severity of underlying disease (ie, McCabe score); intravenous drug use; presence of a urinary or vascular catheter on admission; presence of pressure sores or open wounds; outpatient receipt of dialysis; receipt of antibiotic treatment in the past 6 months; and functional index of dependence (Barthel index).

Microbiologic studies. Within 24 hours after hospital admission, a nasal swab sample was obtained. Additional samples were obtained from patients who had pressure sores, open wounds, or permanent urinary catheters. Samples were processed in enriched blood agar media, media selective for staphylococci (Chapman agar), and media specific for MRSA (MRSA agar; Oxoid). Identification of S. aureus isolates was also performed by conventional methods.

Susceptibilities to oxacillin, trimethoprin‐sulphametoxazole, ciprofloxacin, erythromycin, clindamycin, gentamicin, rifampicin, and vancomycin were determined by use of the Kirby‐Bauer disk‐diffusion method, in accordance with Clinical and Laboratory Standards Institute guidelines.⁶ Genomic DNA from MRSA isolates was prepared in agarose plugs and digested with the SmaI restriction enzyme, as described elsewhere.⁴ Digested DNA fragments were genotyped by means of PFGE.

Statistical analysis was performed using the SPSS statistical package, version 11.5 (SPSS). Univariate analysis was performed using the χ² and Fisher exact tests for the qualitative variables and the Student t test for quantitative variables. Significant variables were included in the logistic regression analysis.

Results

From October 16, 2001, to May 26, 2003, a total of 1,128 patients were included in the study. Nasal swab samples were obtained from all patients; samples from surgical wounds or pressure sores were collected from 12 patients, urine samples from 8 patients, and a sputum sample from 1 patient.

MRSA was isolated from 17 (1.5%) of 1,128 patients. A total of 15 (88.2%) of these patients had nasal carriage, 3 (17.6%) had colonization elsewhere (1 in the respiratory tract, 1 in pressure sores, and 1 in the urinary tract), and 1 patient (5.8%) had septic arthritis. The latter patient did not have nasal carriage at admission and was excluded from the analysis of risk factors for MRSA carriage.

The prevalence of previously unidentified MRSA carriage at hospital admission was 1.4% (95% confidence interval [CI], 0.8%‐2.4%). The prevalence was significantly greater in patients admitted directly from another healthcare center (5.2%), particularly those admitted from long‐term care facilities (11.6%), compared with those admitted from their own home (1%) ( ).

Risk factors related to the presence of MRSA carriage on admission were advanced age, transfer from another healthcare center (ie, a long‐term care facility or hospital), emergency department admission, previous hospitalization or a stay in a long‐term care facility within the past year, and a lower Barthel index (Table). The variables that remained statistically significant on multivariate analysis were mean age (odds ratio [OR], 1.04 [95% CI, 1.00‐1.08]; ), previous hospitalization in the past year (OR, 4.06 [95% CI, 1.13‐14.61]; ) and admission to a long‐term care facility in the past year (OR, 6.47 [95% CI, 1.97‐21.18]; ).

Table.  Risk Factors Associated With Methicillin‐Resistant Staphylococcus aureus Carriage on Admission to a Hospital in Barcelona, Spain

Open New Window

Of the 16 patients with MRSA carriage, 11 were admitted from home, and 5 were admitted from another healthcare center (4 from long‐term care facilities and 1 from a hospital). The mean Barthel index was significantly lower in patients with MRSA carriage who were admitted from other healthcare centers, compared with those admitted from home (mean ± SD, 41 ± 36.98 vs 81.5 ± 27.99; ).

A total of 14 MRSA isolates were tested for susceptibility to antimicrobial agents other than oxacillin. All isolates (100%) were susceptible to vancomycin and trimethoprim‐sulfamethoxazole, 10 (71.4%) were susceptible to clindamycin, 9 (64.3%) were susceptible to gentamicin, 6 (42.9%) were susceptible to erythromycin, and 1 (7.1%) was susceptible to ciprofloxacin.

Genotypes were determined for 12 of the 17 MRSA isolates, which showed 4 distinct PFGE types with identical or closely related patterns (Figure).⁷ There were 3 PFGE type A subtypes (A, A1, and A2) identified in 6 isolates, 3 type B subtypes (B, B1, and B2) identified in 4 isolates, and 1 isolate each of types C and D. No correlation was observed between the antibiotic resistance profiles observed in our study and PFGE patterns.

(89 KB)

Figure.  Pulsed‐field gel electrophoresis patterns of 12 methicillin‐resistant Staphylococcus aureus isolates. Lanes 2 and 10‐12, subtype A; lane 8, subtype A1; lane 3, subtype A2; lanes 1 and 6, subtype B; lane 7, subtype B1; lane 4, subtype B2; lane 5, type C; lane 9, type D.

Open New Window

Discussion

The results of this cross‐sectional study demonstrate that the prevalence of newly identified MRSA carriage at hospital admission is low in our area, and the associated risk factors are previous admission to a long‐term care facility and/or hospital within the past year and advanced age.

A total of 1.4% of the study population had previously unidentified MRSA carriage at admission, a result similar to that observed in studies like ours that excluded patients previously diagnosed with MRSA infection or colonization.^2,8‐11 These latter patients represent up to one‐third of the imported case patients in some studies,³ and may be detected through the hospital database.¹² The prevalence of MRSA carriage at admission was 4‐fold higher among patients admitted directly from another healthcare center than among patients admitted from home, and 7‐fold higher among those admitted directly from a long‐term care facility.¹³

Risk factors for healthcare‐associated MRSA acquisition have been described elsewhere.^8‐11^,13‐20 In our study, the risk factors associated with MRSA carriage on admission were advanced age (OR, 1.05)^{9,12,13,15,16,18,19} and admission to an acute healthcare center (OR, 4.06)^14,20 or a long‐term care facility (OR, 6.47) in the past year.^13,15 These factors may be associated with other risk factors for nosocomial acquisition of MRSA infection or colonization and a high degree of functional dependence.

More than half of the MRSA isolates identified in the present study were susceptible to antibiotics other than ciprofloxacin (to which only 7.1% of the isolates were susceptible) and erythromycin (to which 42.9% of the isolates were susceptible), a result similar to those obtained in other studies describing a reduction in the number of nosocomial multidrug‐resistant clones and increase in the prevalence of other, more sensitive clones.²¹ Contrary to previous studies performed in acute care settings, several PFGE types were detected that had no correlation with several antibiotic resistance profiles.⁴ This may indicate several clones circulating in the hospital and the community.

This study has some potential limitations. The low prevalence of MRSA carriage may not have allowed the detection of differences between the 2 groups as a result of low statistical power. Recall bias may occur when reporting possible past exposures, particularly with respect to previous antibiotic treatment or admissions to hospitals or healthcare centers. Nonetheless, the information was systematically collected by a trained interviewer after obtaining the nasal swab sample.

In conclusion, the prevalence of previously undetected MRSA carriage at admission to our acute care center in Barcelona was low, particularly among patients admitted directly from home. Most patients with MRSA carriage who had been admitted from the community had other nosocomial risk factors for acquisition, and the prevalence of MRSA carriage among individuals with no risk factors was lower. However, most of the patients with MRSA carriage would not have been detected if they had not been screened. Consequently, patients from nursing homes and older patients who have been hospitalized within the past year should be screened for MRSA carriage at hospital admission because of this increased risk.