JSTOR

Methods

From November 2005 through December 2006, respiratory tract specimens were collected from consecutive patients at Thomas Jefferson University Hospital, a tertiary care center in Philadelphia. Patients who had specimens positive for MRSA and a diagnosis of VAP were further analyzed. Pneumonia was defined as the presence of radiographic evidence of lung infiltrates in conjunction with 2 or more of the following signs and symptoms: purulent endotracheal secretions, temperature of 38.0°C or greater, white blood cell count greater than 11,000 cells/mm³ or less than 3,500 cells/mm³, and increasing oxygen requirements. VAP was defined as pneumonia that occurred more than 48 hours after endotracheal intubation.⁶ For all patients with MRSA VAP, the following information was collected: demographic characteristics, presenting illness and clinical course, comorbidities, antimicrobial therapy, and clinical outcome. Additional data collected included the durations of hospitalization and mechanical ventilation before the diagnosis of VAP, overall and intensive care unit–specific lengths of stay, highest or lowest white blood cell count, ratio of the partial pressure of arterial oxygen (Pao₂) to the fraction of inspired oxygen (Fio₂), results of cultures of respiratory tract and blood specimens obtained within the 4 weeks prior to diagnosis, and findings of chest radiography and/or computed tomography (at diagnosis and at discharge or death). Crude mortality rates were calculated on the basis of the time of discharge.

Isolates were identified as MRSA by the MicroScan system (Dade Behring). A multiplex polymerase chain reaction (PCR) with 4 different SCCmec primer sets and a standard PCR for Panton‐Valentine leukocidin (PVL) were performed on each isolate, as described elsewhere.^7,8 Control strains from the Centers for Disease Control and Prevention were tested for SCCmec type I‐IV and PVL by means of the PCR assays described above.

Basic descriptive statistical analysis was performed using SAS software (SAS Institute). A P value of less than .05 was considered to be statistically significant. The study was approved by the institutional review board.

Results

MRSA isolates from 91 patients were collected during the study period. Of 50 isolates excluded from analysis, 21 were not associated with pneumonia, 11 were from patients with hospital‐associated pneumonia, 5 were from patients with healthcare‐associated pneumonia, 4 were from patients with community‐associated pneumonia, 2 isolates were lost, and 7 isolates tested negative for mecA by PCR. Repeat MicroScan tests and a Kirby‐Bauer disk diffusion test for cefoxitin revealed that the latter 7 isolates were methicillin susceptible.

Eight (19.5%) of the 41 MRSA isolates associated with VAP were SCCmec type IV strains; 6 (75.0%) of these 8 were PVL positive. All remaining isolates (33 [80.5%]) were positive for SCCmec type II and negative for PVL. Analysis of baseline characteristics of the patients, as well as clinical, laboratory, microbiologic, and radiographic data on enrollment, is presented in Table 1. A detailed description of the VAP cases due to SCCmec type IV MRSA is presented in Table 2. Six (75.0%) of 8 patients in the SCCmec type IV MRSA group were previously healthy males who were admitted with acute surgical conditions. In contrast, patients with SCCmec type II MRSA VAP were older, were predominately women (52.5%), and had a greater prevalence of underlying diseases. The majority of respiratory tract specimens collected were tracheal aspirates. Bronchoalveolar lavage was performed in 1 patient with SCCmec type IV, compared with 5 patients with SCCmec type II MRSA VAP. Minimum inhibitory concentrations of vancomycin for 10 random MRSA isolates (5 SCCmec type IV organisms and 5 SCCmec type II organisms) were measured by an Etest (AB Biodisk), and all were 1 μg/mL or less.

Table 1.  Comparison of Characteristics for Patients With Ventilator‐Associated Pneumonia (VAP) Due to Staphylococcal Cassette Chromosome mec (SCCmec) Type IV or Type II Methicillin‐Resistant Staphylococcus aureus (MRSA)

Open New Window

Table 2.  Detailed Characteristics of Patients With Ventilator‐Associated Pneumonia Due to Staphylococcal Cassette Chromosome mec Type IV Methicillin‐Resistant Staphylococcus aureus (MRSA)

Open New Window

In patients with SCCmec type IV strains, the white blood cell count was significantly greater than that of patients with SCCmec type II strains, and the ratio of Pao₂ to Fio₂ was significantly less than that for patients with SCCmec type II strains. Vancomycin was administered to all patients with SCCmec type IV and type II MRSA VAP; 5 of 33 patients with SCCmec type II strains also received linezolid concomitantly and/or sequentially with vancomycin. There was a trend toward a longer hospitalization duration after the diagnosis of SCCmec type IV MRSA VAP. However, comparison of the following characteristics revealed no statistically significant differences: overall or intensive care unit–specific length of stay, prevalence of concurrent MRSA bacteremia, and radiographic and computed tomographic findings on discharge (data not shown). Two patients (25.0%) with SCCmec type IV MRSA VAP died, compared with 7 patients (21.2%) in the SCCmec type II MRSA group ( ).

Discussion

To our knowledge, this is the first report to identify SCCmec type IV MRSA strains in patients with VAP (incidence, 19.5%). Although none of the patients with SCCmec type IV MRSA VAP had prior MRSA‐positive cultures of respiratory tract specimens, it is likely that most were colonized with SCCmec type IV MRSA on admission to the hospital. In particular, before admission, patient 3 had been incarcerated, which is a known risk factor for SCCmec type IV MRSA colonization,⁹ and patients 5 and 8 developed MRSA VAP soon (ie, 4‐5 days) after admission, suggesting preadmission colonization. In contrast, for the neonate hospitalized since birth, SCCmec type IV MRSA was presumably transmitted from a family member or, less likely, from a healthcare worker.

No significant difference was observed in VAP severity or crude mortality between patients infected with SCCmec type IV MRSA and those infected with SCCmec type II MRSA. The observed mortality rate for SCCmec type IV MRSA VAP (25%) was also significantly lower than the mean mortality rate reported for SCCmec type IV CAP (50%).¹⁰ Similar to other series, 75% of the SCCmec type IV MRSA strains we analyzed were PVL positive. None of the cases were associated with necrotizing, cavitary pneumonia. These data suggest that cofactors may be important for the development of severe, necrotizing SCCmec type IV MRSA pneumonia.

Patients who develop SCCmec type IV MRSA CAP frequently have a prodrome influenza‐like syndrome.¹⁰ Respiratory viral infections can predispose individuals to superinfection with S. aureus, because of desquamation of the respiratory epithelium and enhanced bacterial adhesion.² Therefore, the absence of a previous viral infection, as well as prompt initiation of antibiotic treatment, in contrast to the delays in diagnosis and initiation of therapy that are frequently encountered in cases of SCCmec type IV MRSA CAP, may partly explain the low mortality rates observed. Additional studies are needed to evaluate the pathogenicity of SCCmec type IV MRSA strains in the respiratory tract.

Our study has several limitations. The total number of MRSA VAP cases analyzed was small, so our comparative analyses have limited statistical value. Cases were reviewed retrospectively, and few cultures of bronchoaveolar lavage specimens were available. Therefore, it was difficult to distinguish between colonization and pneumonia in some cases. Except for 1 case, surveillance cultures for MRSA were not available. Therefore, the role of prior colonization with SCCmec type IV MRSA cannot be determined. Pulsed‐field gel electrophoresis to identify specific MRSA USA clones was not performed.

In summary, to our knowledge this study is the first report of cases of VAP associated with SCCmec type IV MRSA strains. Nearly 20% of cases of MRSA VAP at our institution were associated with SCCmec type IV strains. Although no clusters were observed in our study, patients infected with SCCmec type IV MRSA pose a risk for nosocomial transmission. Therefore, these data support the recommendation of the Society for Healthcare Epidemiology of America that active surveillance for MRSA be performed in intensive care units.¹¹