Blood Culture Results Do Not Affect Cellulitis Treatment
We conducted a retrospective chart review of adult patients treated in the ED with a diagnosis of cellulitis who were treated with antibiotics and had blood cultures attained. All charts were identified using International Classification of Diseases, ninth revision, codes (681–682.9) for patients older than the age of 18 years with a diagnosis of cellulitis who presented in our ED during the 5-year period from January 1, 2005 to December 31, 2009. The study was approved by the Institutional Review Board. Cases were defined by the presence of a concomitant diagnosis of human immunodeficiency virus/acquired immune deficiency syndrome, diabetes, peripheral vascular disease, organ transplantation, dialysis, or active chemotherapy during the ED visit. The group labeled "immunocompromised" was composed of individuals who did not fit into any of the previously mentioned categories, these included patients with unidentifiable neutropenia, lupus on chronic corticosteroids, common-variable immunodeficiency, rituximab or remicade therapy, active leukemia or lymphoma not currently on chemotherapy, end-stage renal disease not on dialysis, and asplenia. We excluded all patients with a separate identifiable source of infection (i.e., pneumonia or bacterial endocarditis) in which blood cultures would likely be obtained as a part of routine care. Controls were identified as presenting to the ED during the same time period with cellulitis for which a blood culture was obtained without the presence of the previously mentioned medical comorbidities.
A standard data-collection sheet was used to gather information, including age, sex, medical history, location of cellulitis if available, previous antibiotics, initial antibiotics given (including dosages), blood culture results, subsequent inpatient or discharge antibiotic, fever, and death rate. The basic demographic information and medical history were obtainable via the ED record. Blood culture results were available from our electronic records, where we were able to directly review the pathology results and not rely on physician narrative. Finally, the administration of antibiotics was obtained from both the ED note and discharge summary, in addition to the electronic medical record. Two authors (A. R. Poreda and W. F. Paulo) abstracted all data onto a standardized collection form. A clinically significant change in antibiotic management was defined as a change in the class of antibiotic to treat isolated pathogens that were either not covered by the initial treating antibiotic or would not have been recommended as empirical cellulitis treatment based on our local antibiogram. Changes in management were considered significant if the change in antibiotic coverage was either expanded or narrowed. Blood culture contaminants were defined as those likely to be from local skin flora (i.e., Staphylococcus epidermidis) and not a true pathologic organism. Primary analysis was done to assess rates of blood cultures and change in management, and secondary outcomes were rates of death and culture results as compared with fever. Data were entered into a Microsoft Excel (Microsoft, Redmond, WA) spreadsheet and analyzed using IBM-SPSS Statistics 19 (IBM-SPSS, Armonk, NY). We calculated odds ratios (ORs) for primary data analysis, as well as the 95% confidence intervals (CIs) around point estimates.
Methods
We conducted a retrospective chart review of adult patients treated in the ED with a diagnosis of cellulitis who were treated with antibiotics and had blood cultures attained. All charts were identified using International Classification of Diseases, ninth revision, codes (681–682.9) for patients older than the age of 18 years with a diagnosis of cellulitis who presented in our ED during the 5-year period from January 1, 2005 to December 31, 2009. The study was approved by the Institutional Review Board. Cases were defined by the presence of a concomitant diagnosis of human immunodeficiency virus/acquired immune deficiency syndrome, diabetes, peripheral vascular disease, organ transplantation, dialysis, or active chemotherapy during the ED visit. The group labeled "immunocompromised" was composed of individuals who did not fit into any of the previously mentioned categories, these included patients with unidentifiable neutropenia, lupus on chronic corticosteroids, common-variable immunodeficiency, rituximab or remicade therapy, active leukemia or lymphoma not currently on chemotherapy, end-stage renal disease not on dialysis, and asplenia. We excluded all patients with a separate identifiable source of infection (i.e., pneumonia or bacterial endocarditis) in which blood cultures would likely be obtained as a part of routine care. Controls were identified as presenting to the ED during the same time period with cellulitis for which a blood culture was obtained without the presence of the previously mentioned medical comorbidities.
A standard data-collection sheet was used to gather information, including age, sex, medical history, location of cellulitis if available, previous antibiotics, initial antibiotics given (including dosages), blood culture results, subsequent inpatient or discharge antibiotic, fever, and death rate. The basic demographic information and medical history were obtainable via the ED record. Blood culture results were available from our electronic records, where we were able to directly review the pathology results and not rely on physician narrative. Finally, the administration of antibiotics was obtained from both the ED note and discharge summary, in addition to the electronic medical record. Two authors (A. R. Poreda and W. F. Paulo) abstracted all data onto a standardized collection form. A clinically significant change in antibiotic management was defined as a change in the class of antibiotic to treat isolated pathogens that were either not covered by the initial treating antibiotic or would not have been recommended as empirical cellulitis treatment based on our local antibiogram. Changes in management were considered significant if the change in antibiotic coverage was either expanded or narrowed. Blood culture contaminants were defined as those likely to be from local skin flora (i.e., Staphylococcus epidermidis) and not a true pathologic organism. Primary analysis was done to assess rates of blood cultures and change in management, and secondary outcomes were rates of death and culture results as compared with fever. Data were entered into a Microsoft Excel (Microsoft, Redmond, WA) spreadsheet and analyzed using IBM-SPSS Statistics 19 (IBM-SPSS, Armonk, NY). We calculated odds ratios (ORs) for primary data analysis, as well as the 95% confidence intervals (CIs) around point estimates.
SHARE
