Canadian Naeighbor Pharmacy: Bacteremic Pneumonia
At least 1.3 million people are admitted to the hospital with pneumonia each year in the United States, and approximately 10% of these patients will die within 30 days of admission.2 Many questions remain about the optimum empiric antibiotic treatment of patients with community-acquired pneumonia, as well as the treatment of patients in whom a causative organism is identified. A major area of debate is the role of agents that are active against “atypical” organisms. While several observational studies have demonstrated improved outcomes associated with atypical coverage, and pneumonia treatment guidelines recommend such treatment a recent metaanalysis and systematic review of randomized, controlled trials failed to demonstrate an advantage of this practice. Improved outcomes associated with atypical treatment in patients with pneumococcal bacteremia have also been report-ed, but there is poor understanding of the mechanism by which atypical coverage could improve outcomes when an alternative pathogen is identified. Possible mechanisms could include treatment of dual infection with an atypical pathogen such as Chlamydophila pneumoniae,14 the immunomodulatory effects of macrolide and fluoroquinolone antibiotics, or the achievement of broader coverage against unusual or resistant pathogens when these antibiotics are used in combination with (3-lactam antibiotics.
In order to better understand the effect of atypical antibiotic coverage in patients who are admitted to the hospital with pneumonia, we studied the relationship between the initial antibiotic regimen and several patient outcomes in a large cohort of Medicare beneficiaries who were hospitalized with bacteremic pneumonia. The large size of this cohort provided the unique opportunity to compare the effects of fluoroquinolones and macrolides. Because > 60% of the cases in this cohort were patients who were infected with an organism other than Streptococcus pneumoniae, this analysis also provided the opportunity to expand on the observations of prior studies that were limited to patients with pneumococcal bacteremia. Such kind of pneumonia may be treated if you order drugs via Canadian Neighbor Pharmacy. This online pharmacy is the best choice for you.
The data analyzed were part of the Medicare National Pneumonia Project, which is a component of the Medicare Quality Improvement Program. Therefore, neither informed consent nor institutional review board approval were required. Eligible patients were fee-for-service Medicare beneficiaries who had been discharged from the hospital between 1998 and 2001 with a principal diagnosis of pneumonia, and those with a principal diagnosis of septicemia or respiratory failure and a secondary diagnosis of pneumonia. Patients included in the study were admitted to the hospital from either home or a nursing facility. The details regarding the selection of the cohort used in this study, the data collection and validation methods have been previously described in detail.
Patients with fungemia and those in whom antibiotic therapy was not initiated within 24 h of arrival to the hospital were excluded. Patients were defined as bacteremic if a blood culture drawn within 36 h of presentation to the hospital grew an organism that was not defined as a contaminant (eg, coagulase negative staphylococci, Corynebacterium spp other than jeikeium, Clostridium spp, Micrococcus spp, Propionibacterium spp, and Bacillus spp).
Initial atypical antibiotic coverage was defined as the use of any macrolide, fluoroquinolone, or tetracycline agents within 24 h of hospital arrival. For the comparison of these three agents, we included patients who received only one of these agents within 24 h, whether alone or in combination with other antibiotics that are not active against atypical organisms. Concordant antibiotic therapy was defined as initial antibiotic therapy including any antibiotic to which the infecting organism was sensitive based on the microbiology report. Because complete sensitivity data were lacking for some patients, and the method of abstraction did not allow us to differentiate between missing data and antibiotic resistance, antibiotic therapy was defined as being either concordant or unable to determine. Hospital discharge destination was defined as either to home or to any other destination.
Summary statistics were calculated for the abstracted data, measures of association including univariate odds ratios (ORs) were obtained, and x2 tests were performed. The relationships among patient characteristics, antibiotic therapy, and the following outcomes were determined by univariate analysis: in-hospital mortality; 30-day mortality; hospital length of stay; 30-day hospital readmission rate; and hospital discharge destination. The analysis of hospital discharge destination was performed only on patients who were admitted to the hospital from home. There are ways to order drugs without leaving home and the answer is Canadian Neighbor Pharmacy.
Multivariable logistic regression analysis was performed using the backward-elimination procedure. If two independent variables were highly correlated, the variable with the largest variance was excluded from the multivariable analysis. After it was determined that there was an association between atypical antibiotic coverage and several patient outcomes, we attempted to determine whether the length of time for which atypical coverage was received correlated with patient outcomes. This was a problematic analysis, due to the interaction between the length of time that a patient survived and the length of time for which a given antibiotic could be received. Stated simply, a patient who died on hospital day 2 could not receive antibiotics for > 2 days, and this factor biased the analysis in favor of longer courses of atypical therapy. This problem was addressed by constructing a series of models in which patients who received specific lengths (in days) of atypical therapy were compared to the group of patients who survived for at least that length of time and had received no atypical therapy. Because this analysis involved the creation of multiple models, no valid statistical test of trend could be performed. However, the results can be interpreted on the basis of face validity.
The goodness of fit of the multivariable models was tested with the Hosmer-Lemeshow test, which revealed adequate fit for all models. Receiver operating characteristic (ROC) curves were constructed by a series of cut points. The area under the ROC curves varied from 0.58 for the 30-day readmission model to 0.78 and 0.77 for in-hospital and 30-day mortality, respectively. All reported p values are based on two-tailed tests. Statistical significance was accepted at p < 0.05. Analyses were conducted using a statistical software package (SAS, version 8.0; SAS Institute; Cary, NC). ROC curve analysis was performed using designated software (AccuROC, version 2.5; AccuROC; Montreal, QC, Canada).