Presentation Authors: Jonathan Dokter, Rochester, MI, Laura Nguyen, Lauren Tennyson*, Esther Han, Larry T Sirls, Royal Oak, MI
Introduction: When treating urinary tract infection (UTI) urine culture delay forces clinicians to decide whether to start empiric antibiotics, possibly incorrectly. We evaluated empiric antibiotic use and subsequent change in a FPMRS practice.
Methods: Patients with suspected UTI and urine culture from Jan 2016 to June 2016 were retrospectively reviewed. Exclusion criteria included indwelling catheter, institutionalized, recent urologic surgery, pregnancy, and urinary diversion. Patients with a positive culture (defined as > 103 CFU) were categorized by whether or not they were treated empirically. Empiric treatment was evaluated for associations with clinical-demographic data, symptoms and urinalysis (UA). Antibiotic change was evaluated with clinical-demographic data, urine culture results and resistance patterns. Statistical significance of differences between groups was assessed with chi-square or Fisher's exact test for categorical variables and t-test for continuous measures.
Results: 916 urine cultures (636 patients) were included. 391 (43%) cultures were positive, and 164 (42%) were treated empirically. Demographics including age, gender, BMI, history of diabetes, or immunosuppression did not differ between groups. Those treated empirically had more UTI symptoms (93% vs 58% P < 0.001), and UA abnormalities including positive nitrites (51% vs 29% P < 0.001), 3+ leukocyte esterase (27% vs 19% P = 0.002) and 3+ blood (13% vs 4% P = 0.005). The most common bacteria in both groups were E. coli, followed by Klebsiella and Enterococcus. The most common empiric treatments were Macrobid (24%), TMP-SMX (22%), ciprofloxacin (15%) and doxycycline (13%). Of those treated empirically, 42/164 (26%) required an antibiotic change. Antibiotic change was associated with immunosuppression (12% vs 2% P = 0.027) and > 3 antibiotic resistance (33% vs. 20%, P = 0.039). Patients requiring empiric antibiotic change also had higher rates of fluoroquinolone (50% vs 30% P = 0.016), monobactam (19% vs 7% P = 0.042) and TMP-SMX (52% vs 19% P < 0.001) resistance.
Conclusions: Almost half of patients with clinical UTI were treated empirically. Factors associated with empiric treatment included symptoms and abnormal UA. Antibiotic change was driven largely by bacterial resistance. New technologies that allow rapid urine bacterial identification and sensitivities may greatly improve patient care.