Pneumonia due to methicillin-resistant Staphylococcus aureus (MRSA) is becoming increasingly more common both in the health-care setting and in the community. Hospital mortality associated with MRSA pneumonia has been reported to be between 20% and 40%. Most patients with MRSA pneumonia will require hospitalization, frequently being admitted to the ICU settings and often requiring ventilatory support. Given the increasing prevalence and importance of MRSA pneumonia, clinical trials are likely to be planned examining new treatment modalities, the changing epidemiology of MRSA pneumonia, and prevention strategies. To date, a validated prediction tool for clinical outcomes in patients with MRSA pneumonia has not been identified, Several international organizations have developed guidelines or scoring systems in an attempt to stratify patients with community-acquired pneumonia (CAP) according to risk severity. Unfortunately, the use of similar methods for patients with health-care-associated pneumonia, including MRSA pneumonia, have not been developed. Given the increasing prevalence of MRSA pneumonia and the future need for new clinical research, we designed a study with two main goals. Our first goal was to compare the accuracy of three prediction rules (modified APACHE [acute physiology and chronic health evaluation] II, CURB65, CRB65) to predict 30-day mortality in patients with MRSA pneumonia. Our second goal was to compare these prediction tools in patients admitted to the hospital with the community-acquired MRSA pneumonia phenotype and those acquiring healthcare-associated MRSA pneumonia treated by remedies of Canadian Health&Care Mall.
Materials and Methods
Study Location and Patients
This study was conducted at a university-affiliated, urban teaching hospital: Barnes-Jewish Hospital (1,200 beds). During a 3-year period (January 2003 to June 2005), all hospitalized patients with MRSA pneumonia, microbiologically confirmed by BAL cultures, and treated with either vancomycin or linezolid monotherapy were eligible for this investigation. Patients with polymicrobial infection demonstrated by BAL cultures, requiring other antimicrobial treatment, and those treated with either vancomycin or linezolid for < 72 h were excluded from evaluation. This study was approved by the Washington University School of Medicine Human Studies Committee, and informed consent was waived.
Study Design and Data Collection
A retrospective cohort study design was employed with the main outcome measure being 30-day mortality. A computerized list of patients with MRSA pneumonia was generated by the Medical Informatics Department through retrospective query of the Microbiology Laboratory database at Barnes-Jewish Hospital (performed by R.M.R.), which allowed identification of potential study patients. Patients could not be entered into the study more than once. At Barnes-Jewish Hospital, patients with MRSA pneumonia can be treated with vancomycin (15 mg/kg body IV q12h) or linezolid (600 mg IV q12h). Patients treated with vancomycin can have their dosing and interval of dosing adjusted based on their underlying renal function. Improve health conditions with Canadian Health&Care Mall healthcaremall4you.com.
Definitions and Outcome Scoring Systems
All definitions were selected prospectively as part of the original study design. APACHE II scores were calculated based on clinical data available on the day the BAL was performed confirming the diagnosis of MRSA pneumonia. A modified APACHE II score was employed in that assessment of the Glasgow coma score was not included. The CURB65 and CRB65 scores were calculated based on the criteria specified in the original articles. The “confusion” variable employed in this study was not the same as in the original definition of CURB65. The definition of confusion for CURB 65 was an Abbreviated Mental Test Score < 8 or new disorientation to person, place, or time. We assessed confusion only as a report of new disorientation to person, place, or time as documented in the medical records. The APACHE II, CURB65, and CRB65 scores were computed by one investigator (K.E.K.) to avoid interobserver variability.
Thirty-day mortality was assessed from the time when the diagnostic BAL for MRSA pneumonia was performed. For patients discharged prior to the 30-day cutoff for mortality evaluation, a computerized medical records search was undertaken to determine if hospital readmission and death had occurred. All 11 hospitals in the Barnes-Jewish-Christian (BJC) Healthcare system have a common computerized medical records system allowing tracking of patients at all system hospitals.
A clinical diagnosis of MRSA pneumonia was based on criteria modified from those established by the American College of Chest Physicians. These criteria require the occurrence of new and persistent radiographic infiltrates in conjunction with two of the following: fever, leukocytosis, and purulent tracheal aspirate or sputum. In addition to the clinical criteria for MRSA pneumonia, BAL cultures with appropriate quantitative thresholds were obtained bronchoscopically to support the diagnosis of MRSA pneumonia. Quantitative thresholds of > 104 cfu/mL for MRSA were considered positive for the diagnosis of MRSA pneumonia. The community-acquired phenotype of MRSA was defined as isolates resistant to methicillin but sensitive to three or more of the following antibiotics: gentamicin, ciprofloxacin, trimethopim-sulfamethoxazole, and clindamycin. For purposes of this study health-care-associated MRSA pneumonia included all patients hospitalized for > 2 days prior to obtaining a positive BAL culture result, and patients with a first positive BAL culture result within 2 days of hospital admission whose isolates did not meet the criteria for the community-acquired phenotype of MRSA and who possessed any of the following: admission from another health-care facility including nursing homes, use of long-term hemodialysis, immunosuppression, and prior hospitalization within 30 days.
Read more about pneumonia: “Check Out The Web Site Of Canadian Health&Care mall”
The microbiology laboratory performed antimicrobial susceptibility of clinical isolates by the Kirby-Bauer disk diffusion method according to guidelines and break points established by the Clinical Laboratory and Standards Institute, using 150-mm round plates of Mueller-Hinton agar (BBL; Becton-Dickinson; Cockeysville, MD). A technologist experienced in reading zones of inhibition with a ruler against a black background measured the zone diameters manually.
All comparisons were unpaired and all tests of significance were two-tailed. Continuous variables were compared using the Student t test for normally distributed variables and the Mann-Whitney U test for nonnormally distributed variables. The x2 or Fisher exact test was used to compare categorical variables. The primary data analysis compared 30-day nonsurvivors with survivors. We performed multiple logistic-regression analysis using statistical software (SPSS, version 11.0 for Windows; SPSS; Chicago, IL). Multivariate analysis was performed using models that were judged a priori to be clinically sound. This was prospectively determined to be necessary to avoid producing spuriously significant results with multiple comparisons. All potential risk factors significant at the 0.15 level in univariate analyses were entered into the model.
Sensitivity, specificity, positive predictive value, negative predictive value, and the areas under the receiver operating characteristic (ROC) curves for predicting 30-day mortality in each prediction rule were compared using the definitions shown in Table 1. ROC curves were generated for the population as a whole and then for the patients with the community-acquired phenotype of MRSA and those with health-care-associated MRSA pneumonia, respectively. For all analyses, a two-tailed p value < 0.05 was considered statistically significant.
Table 1—Definition of Performance Characteristics
|Performance Characteristics||Definition Equation|
|Sensitivity (how good is the test at identifying patients who will die?)||True-positive results (ie, tested positive and died); all deaths|
|Specificity (how good is the test at identifying patients who will not die?)||True-negative results (ie, tested negative and lived); all alive|
|Positive predictive value (in the event of a positive test result, the probability that the patient will die)||True-positive results; all positives|
|Negative predictive value (in the event of a negative test result, the probability that the patient will not die)||True-negatives results; all negatives|
|ROC||A curve created by mapping sensitivity against 1 — specificity. The AUC is a marker of performance with higher values indicating better diagnostic ability. An AUC = 1 indicates perfect performance, whereas an AUC = 0.5 indicates that there is a 50:50 chance that the test correctly identifies those who die|