CAEP/ACMU 2000 Scientific Abstracts: 54-57

CAEP Abstracts

CJEM 2001;2(3):169-187

054 Care of Community-Acquired Pneumonia: The Impact of Adding Levofloxacin to a Hospital Formulary.
Marras TK, Chan CK. University Health Network, University of Toronto, Toronto, Ont.

INTRODUCTION: Community-acquired pneumonia (CAP) is a common disease with significant morbidity, mortality and financial burden. The administration of antibiotics within 8 hours of presentation to the emergency department is associated with lower mortality. Recent CAP guidelines recommend newer-generation fluoroquinolones, such as levofloxacin, in several settings. OBJECTIVES: To assess clinical outcomes and treatment costs before and after introducing levofloxacin on to our formulary, and adherence to and impact of applying recent CAP treatment guidelines.
METHODS: Retrospective cohort study, reviewing charts of all CAP patients admitted between October 1997 and June 2000 to the Toronto General and Toronto Western Hospitals. Patients will be risk-stratified according to the Pneumonia Severity Index (PSI), and treatment costs and clinical outcomes will be analyzed. RESULTS: We expect to identify 600­1,000 cases. To date, we have identified 24 cases from the pre-quinolone era and 52 from the post-quinolone era. Interim analysis of the post-quinolone era, comparing patients who received levofloxacin at some time during their hospital course (ever-levofloxacin, 24 cases), with patients who never received levofloxacin (never-levofloxacin, 28 cases), shows that both groups were composed of elderly patients (median age 75 and 78 respectively, p = 0.16); 17% and 14% respectively came from a long-term care facility (p = 0.83); and the median PSI class was 4.5 and 4 respectively (p = 0.48). These very preliminary data show trends to lower mortality (4% vs. 11%, p = 0.40) and lower median drug cost per patient ($54.00 vs. $62.21, p = 0.59), but a slightly longer median hospital stay (7.5 vs. 7 days, p = 0.92) in the ever-levofloxacin group. No trend reached statistical significance. CONCLUSIONS: Interim results regarding the use of levofloxacin in the inpatient treatment of CAP are encouraging and support current guidelines. A much larger data sample is required before making firm conclusions. Analysis of 2­3 times the current number of cases will be presented.
Key words: pneumonia, therapy, guideline

055 Determinates of HIV and Hepatitis C Infection in Two Northern, Urban Emergency Departments: Results of Anonymous Unlinked Surveillance.
Rowe BH, Mashinter L, Joffe M, Mackey D, Preiksaitis Y, Houston S. University of Alberta, Edmonton, Alta.

OBJECTIVES: Transmission of blood borne pathogens like HIV and Hepatitis C (HC) are potential risks to all emergency staff. Data regarding the epidemiology of these infections in non-US ED settings are limited.
METHODS: Consecutive patients aged 15 to 54 were included if they presented to either of 2 urban Canadian EDs and had a CBC drawn. Left over blood was collected and serotested for HIV and HC after removal of all personal identifiers. The list of study patients was cross-referenced against local databases of known HIV and HC seropositive patients prior to unlinking and serotesting. Univariate and multivariate analyses were performed. RESULTS: Samples from 3057 individuals were analyzed. All age groups were equally represented, 7% (213) were aboriginal, and 1654 (51%) were male. Medical diagnoses were common (2193 [71%]) and injury accounted for 643 (21%) cases. Overall, 39 (1.3%; 95% CI, 1.0%­2.0%) patients had proven HIV infection which was most closely associated with known and proven HC infection (p < 0.001). More patients (302 [9.8%; 95% CI, 9.0%­11.0%]) were HC seropositive which was closely associated with aboriginal status (p < 0.001), increasing age (p < 0.001), male gender (p < 0.001), hospital (p < 0.001), and proven HIV infection (p < 0.001). CONCLUSIONS: The prevalence of HC was high in unselected ED attendees at these 2 sites while the prevalence of HIV was low. Most HIV infection was associated with HC seropositivity, suggesting intravenous drug use transmission. These results emphasize the importance of observing universal precautions in ED practice and educating staff about exposure risks.
Key words: hepatitis C, HIV, seropositivity

056 Clinical Utility of Blood Cultures in the Emergency Department.
Innes G, Roland K, Grafstein E, Christenson JM. St. Paul's Hospital, University of British Columbia, Vancouver, BC.

OBJECTIVES: ED blood cultures (BC) may seldom affect patient outcomes. Our objectives were to estimate the rate of true positive BC, the proportion of BC that are potentially useful, and to identify clinical characteristics of patients having useful BC.
METHODS: During a 6 month period, adults who had BC drawn in our inner city ED were identified using the lab database. The study included all patients with positive BC and a concurrent random sample of negative cultures. Demographics, comorbidity, clinical predictors and diagnostic data were gathered from ED charts. BC were defined as potentially helpful if 3 parameters were met: 1) The BC grew a non-contaminant organism, 2) Antibiotic Rx was changed after the BC result, 3) The organism was not grown in primary site cultures (e.g., urine, CSF). RESULTS: 767 ED patients had BC drawn. Of these, 23 (3%) grew contaminants and 83 (11%) grew pathogens. Antibiotics were subsequently changed in 47 cases (6%); however, in 21 of these, the pathogen also grew in primary site cultures and in 16, appropriate antibiotics had already been initiated in the ED. BC may have been useful in 16 patients (2.1%) by causing appropriate treatment changes (n = 10), by leading to a diagnosis of endocarditis (n = 4) or by identifying an unexpected organism in patients already appropriately treated (n = 2).

CONCLUSIONS: Relatively few ED BC are helpful. Accepted predictors (e.g., WBC, fever) may not identify patients likely to benefit. A clinical prediction rule for ED BC should be developed.
Key words: blood cultures, utilization, utility

057 Emergency Department Point of Care Pregnancy Testing.
Filiatrault L, Ng DC, Whitlow K, Pudek M. Vancouver General Hospital, Vancouver, BC.

OBJECTIVES: Point of care ED testing may facilitate patient management. We sought to determine the user satisfaction, time impact and diagnostic performance of a urine pregnancy test utilized by nurses in a tertiary ED.
METHODS: After delivery of a standardized training program, we prospectively evaluated the Abbott TestPack Plus hCG Combo with On Board Controls urine pregnancy test (ATP). All patients deemed by an emergency physician to require pregnancy testing were eligible. Subjects had simultaneous ED point of care ATP testing done by an emergency nurse and central laboratory urine or serum pregnancy testing using a Tandem ICON 2 ImmunoConcentration Assay. The laboratory result was considered the gold standard. ED staff were blinded to the laboratory result when interpreting the ATP and were surveyed to determine user satisfaction and time impact of the test.
RESULTS: A convenience sample of 128 patients were enrolled. There were 16 positive laboratory results for a pregnancy prevalence of 12.5%. The ATP had one false positive and one false negative result giving a sensitivity of 93.8% (95% CI 69.8%­99.8%) and a specificity of 99.1% (95% CI 95.1%­99.9%). Overall ATP accuracy was 98.4% (95% CI 94.5%­99.8%). The survey indicated that the laboratory result was available in a median time of 62 minutes versus 5 minutes for the ATP (p < 0.001) and that users were highly satisfied with the test and believed it accelerated patient disposition.
CONCLUSIONS: Point of care urine pregnancy testing by ED nurses is rapid, highly accurate, and accelerates patient disposition. At our institution, the ATP has now replaced central laboratory pregnancy testing.
Key words: pregnancy, diagnosis, point-of-care