Fever Duration in Hospitalized Acute Pyelonephritis Patients Marcel A. Behr, MD, MSc, Robert Drummond, MD, CM, Michael D. Libman, MD, CM, J. Scott Delaney, MD, CM, Joseph S. Dylewski, MD, CM, Montreal, Quebec, Canada PURPOSE: To study persistence of fever in treated pyelonephritis with respect to guidelines recommending investigation and modification of therapy after 2 to 3 days of fever. PATIENTS AND METHODS: A retrospective chart review was made of 70 patients hospitalized for febrile pyelonephritis at a community hospital in Canada. RESULTS: Median duration of fever was 34 hours; persistence of fever at 48 and 72 hours was 26% and 13%, respectively. No patients had complications such as intrarenal or perirenal abscess. Prolonged fever was independently associated with increasing baseline creatinine (P = O.OOOl), younger age (P = 0.027), and increasing total leukocyte count (P = 0.026). Results of ultrasonography and intravenous urograms were not predictors of fever duration. CONCLUSION: Fever in treated pyelonephritis can take 4 days to resolve, and routine urologic investigation after 2 to 3 days of fever may be unwarranted. Am J Med. 1996; 101:277-280. cute pyelonephritis is a common medical illness A that may require hospitalization. Approximately 250,000 cases per year are treated in the United States and fever is present in at least 70%.*The pres- ence of fever is used both for diagnosis and to follow up treatment response. Several infectious disease textbooks3-” recommend that after 2 to 3 days of continued fever, testing should be done to exclude obstruction or abscess formation (Table I). It is also common clinical practice to consider modifying therapy if fever persists after 72 hours of antibiotics.” The objective of this study was to determine the clinical course and the predictors of fever duration in febrile patients admitted for a confirmed diagnosis of acute pyelonephritis. By describing the clinical outcome in these patients, we hoped to be able to From the Departments of Medicine (MAB, MDL. JSD) and Family Medl- clne (RD. JSD). St. Marv’s Hosoital. Montreal, Ouebec. Canada. Requests for reprints>should be addressed to Joseph S. Dylewski, MD, Department of MIcrobiology. St. Mary’s Hospital, 3830 Lacombe, Mon- treal, Quebec, H3T 1M5 tinada. Manuscript submltted December 15, 1995 and accepted in revised form May 28, 1996. I I ‘cl996 by Excerpta Medica, Inc. All rights reserved. describe the frequency of prolonged fever, and whether there was an association with complica- tions (eg, stones, abscess). METHODS AND MATERIALS St. Mary’s Hospital is a 400-bed, university-affili- ated, community hospital serving an urban popula- tion of mixed socioeconomic status. Charts were ob- tained from medical records where the discharge diagnosis listed urinary tract infection as either the principle or a secondary diagnosis. All 234 charts meeting this criterion from August 1991 to July 1994 were retrospectively reviewed. The criteria for py- elonephritis were fever (temperature ~38.0”C oral), a positive urine culture (210” colony-forming units/L (cfu/L) organisms, pure growth), and the absence of another possible source of infection on chart re- view. Clinical data on admission and during hospitaliza- tion were obtained from the chart. Data at presen- tation included past medical and surgical history, such as previous urinary tract infections, pregnancy, bladder catheterization, and nosocomial acquisition. Laboratory data from specimens collected within 12 hours of presentation included blood leukocyte count (WBC) , renal function as assessedby blood creatinine and urea nitrogen (BUN), and organism cultured. Information on duration of illness prior to admission, presence of flank pain, and urinalysis were not routinely available and therefore were not included. Data on antibiotic therapy were collected, although there were often several changes in the first day between choices and doses by emergency and ward physicians. Duration of fever was defined as the interval from initial evaluation until the beginning of the first 12- hour period with a maximum temperature 37.5”C (oral). Reports of radiologic investigations, such as abdominal x-rays, ultrasound (US), and intravenous urograms (IVP) were also recorded. Data Analysis Data analysis was performed using Statistical Ap- plications Software (SAS) . Simple univariate analy- sis was applied to determine descriptive statistics of the variables; for reasons of precision, fever is rounded to the nearest whole number. Bivariate analysis was employed to explore correlations be- OOOZ-9343/96/$15.00 277 PII SOOOZ-9343(96)00173-8