Urine culture susceptibility reports may be misleading: Cultured patients may be different than the majority treated empirically

Abstract

Study objectives: Hospital antimicrobial susceptibility reports (antibiograms) are recommended for guiding empiric therapy, but guidelines often recommend culturing only patients likely to have resistant or complicated infections. Therefore, antibiograms may not reflect the organisms present in the majority of empirically treated low-risk patients, resulting in accelerated development of antimicrobial resistance in the community. We asked how many emergency department (ED) patients treated for urinary tract infection (UTI) had urine cultures (UCS) performed and what the results were relative to empiric treatment decisions. Our specific hypothesis was that cultured patients would differ clinically from noncultured patients in risk of resistant or complicated infection. Methods: This was a retrospective cross-sectional study. Electronic and clinical records for consecutive adults treated for UTI in an urban public teaching hospital were abstracted for demographic, comorbidity, severity of illness, laboratory test result, UCS, diagnosis, treatment, and disposition data. Cases were identified as any patient older than 17 years and with any discharge or admission diagnosis of UTI or one of its synonyms. Factors known to be associated with resistant infections were measured. In the specific comparisons of clinical factors that might predict either receiving a UCS or having a positive UCS result, statistical significance was tested using Student's t test for continuous and χ 2 analysis for dichotomous variables. The α chosen was 0.001 because 11 factors were examined. Results: Of 685 adult UTI patients identified, 30 (4%) were excluded because of unavailable medical records. There were 655 eligible patients with a mean age of 40 years (SD 17 years), 90 (14%) were men, 133 (20%) were diagnosed with pyelonephritis or sepsis, 124 (19%) had underlying genitourinary comorbidity or devices, 79 (12%) were diabetic, 18 (3%) had cancer, and 52 (8%) had renal stones. There were 216 (33%) patients who received UCS (95% confidence interval [CI] 29% to 37%). Those receiving UCS were significantly older, had higher blood urea nitrogen and creatinine levels, and were more likely to be male and diabetic and to have underlying genitourinary problems ( P P 1%) patients received nitrofurantoin. Choice of antibiotic was not significantly associated with performing UCS or positive UCS. Only 33 (5%) patients were hospitalized or treated in our observation unit. An additional 57 (9%) patients received parenteral antibiotics and fluids in the ED before discharge. Conclusion: Although 55% of patients were discharged on fluoroquinolones, this clinical decision may be based on a hospital antibiogram reflecting a smaller group of patients with higher risk of resistance or adverse outcomes than those empirically treated. Only 33% of our sample received UCS. Those patients had higher rates of comorbidity and severity of illness indicators predictive of resistant infection than those who did not receive UCS. Even so, only 2% of the total sample had a culture-documented requirement for fluoroquinolone treatment. Also, the older, less costly nitrofurantoin had better activity against the recovered organisms than fluoroquinolones did but was used to treat only 4% of patients. Over a 3-year period, our hospital antibiogram has reflected a 10% reduction in the effectiveness of fluoroquinolones in eradicating typical urinary pathogens, which may be due to their excessive use in ambulatory low-risk patients who would benefit from less broad-spectrum agents. Our findings suggest that intermittent surveillance cultures in low-risk UTI patients may promote improved antimicrobial choices and cost savings.