Bacteremic Urinary Tract Infection in Hospitalized Older Patients—are Any Currently Available Diagnostic Criteria Sensitive Enough?

To the Editor: Atypical disease presentations and the high prevalence of asymptomatic bacteriuria make diagnosing urinary tract infection (UTI) in older adults difficult.1,2 Misdiagnosis rates of hospitalized elderly patients may be as high as 40%.3 No reliable diagnostic test exists. Criteria based on collection of clinical features have been developed, although these are primarily designed to guide antibiotic treatment in nursing home settings.4–6 The need to reduce unnecessary antibiotic use has led to the suggestion that such criteria may aid diagnosis in acute hospital settings. A sample of patients with bacteremia caused by a known uropathogen with concomitant isolation of the same organism from the urinary tract without a more likely alternative source was identified. Bacteremic UTI (bUTI) has a worse prognosis than non-bacteremic UTI in hospitalized older adults (mortality 15–33% vs 6%).3,7,8 Any useful diagnostic criteria would need to be sensitive enough to detect such cases most of the time. Patients were identified with concurrent positive blood and urine cultures for the same organism from the pathology laboratory computer system. Patients' records were then retrospectively obtained to extract relevant clinical information. Patients with a likely alternative source of the bacteremia with secondary urinary tract seeding were excluded. Patients were divided into two groups—aged 18 to 74 and 75 and older. Three sets of proposed diagnostic criteria were applied retrospectively for the older patients to try to evaluate their sensitivity to detect significant (bacteremic) UTI in hospitalized patients.4–6 Data were available for 61 patients (mean age 76.7; 55.7% male). Thirty-seven were age 75 and older (mean age 84.7, 56.8% male, 21.6% residing in a care home), and 24 were aged 18 to 74 (mean age 66.5, 58.3% male, none residing in a care home). Symptoms and clinical signs are summarized in Table 1. There was a low rate of reported urinary tract–specific symptoms in both age groups, particularly for dysuria in older patients, which was recorded in just one of the 37 cases. Of the patients aged 75 and older, 18 (48.6%) did not report any urinary tract symptoms. The most frequently detected clinical sign was the presence of at least mild pyrexia (>37.0°C), present in all of the younger group and 91.9% of the older group. Just 73.0% of the older group met the stricter criteria of temperature greater than 37.9°C. Based on the assumption that some older, frailer adults may start at a lower baseline temperature or be incapable of mounting an optimal febrile response to infection, some authors have suggested a rise of 1.5°C or more above baseline on two or more occasions could represent significant infection. Baseline temperatures were estimated from an average of three convalescent values. The use of this criterion as an alternative to temperature greater than 37.9°C led to just two more cases out of 37 being identified. Baseline temperatures averaged 36.3°C in the older group (range 35.1–36.8°C) and in the younger group (range 35.5–36.7°C). The overall mortality rate for patients with bUTI was 18.0%. The rate was higher in the older group (27.0%) than the younger (4.1%). The diagnostic criteria of McGeer4 would have detected 16 of the 37 cases of bacteremia in patients aged 75 and older (a false negative rate of 56.8%). The criteria of Loeb5,6 had a false negative rate of 43.2% (21 of 37 cases correctly identified). The finding that only approximately half of elderly patients with UTI have any urinary tract symptoms is consistent with previous studies.3,9 Nonspecific declines in function or symptoms suggesting a respiratory origin have previously been associated with bUTI in older adults.9 Despite proven bacteremia, 8.1% of patients aged 75 and older did not have even mild pyrexia (>37.0°C) at any stage of their illness. Twenty-seven percent of patients aged 75 and older did not reach the more stringent criteria of a temperature greater than 37.9°C. Previous studies of bUTI have also found that 37% to 49% of older patients do not have significant pyrexia at the time of presentation.7,9 The observed mortality rate of 27% in the group aged 75 and older is similar to that reported in other studies (15–33%).3,7,8 None of the schemes evaluated against the data in this series showed reliable sensitivity to detect bUTI in a hospitalized patient group. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Author Contributions: HJW designed the study, reviewed the patients' records to extract data, and wrote the article. CG, MM, and JM obtained bacteriological data. Sponsor's Role: NA.