http://dx.doi.org/10.1016/j.jemermed.2015.06.028 The Journal of Emergency Medicine, Vol. 49, No. 6, pp. 998 1003, 2015 Copyright Ó 2015 Elsevier Inc. Printed in the USA. All rights reserved 0736-4679/$ - see front matter Brief Reports RESISTANCE PATTERNS OF ESCHERICHIA COLI IN WOMEN WITH UNCOMPLICATED URINARY TRACT INFECTION DO NOT CORRELATE WITH EMERGENCY DEPARTMENT ANTIBIOGRAM Michelle C. Hines, PHARMD,* Tareq Al-Salamah, MBBS, MPH,* Emily L. Heil, PHARMD,* Haney Mallemat, MD, Michael D. Witting, MD, MS, Jennifer K. Johnson, PHD, Michael E. Winters, MD, and Bryan D. Hayes, PHARMD* *Department of Pharmacy Services, University of Maryland Medical Center, Baltimore, Maryland, King Saud University, Riyadh, Kingdom of Saudi Arabia, Department of Emergency Medicine, and Department of Pathology, University of Maryland School of Medicine, Baltimore, Maryland Reprint Address: Bryan D. Hayes, PHARMD, Department of Emergency Medicine, University of Maryland School of Medicine, 22 South Greene St., Baltimore, MD 21201, Abstract Background: Urine cultures are not always performed for female Emergency Department (ED) patients with uncomplicated urinary tract infection (UTI). Accordingly, hospital, and even ED-specific, antibiograms might be skewed toward elderly patients with many comorbidities and relatively high rates of antimicrobial resistance, and thus do not accurately reflect otherwise healthy women. Our ED antibiogram indicates Escherichia coli resistance rates for ciprofloxacin, levofloxacin, and trimethoprim-sulfamethoxazole (TMP-SMX) of 42%, 26%, and 33%, respectively. Objectives: This study aims to compare resistance rates of urinary E. coli from otherwise healthy women with uncomplicated UTI and pyelonephritis in the ED to rates in our ED antibiogram. Methods: Females > 18 years old with acute onset of urinary frequency, urgency, or dysuria with pyuria identified on urinalysis (white blood cell count > 10/high-power field) were prospectively enrolled in the ED of an urban, academic medical center. Exclusion criteria indicating a complicated UTI were consistent with Infectious Diseases Society of America guidelines. Susceptibility patterns of E. coli to ciprofloxacin, levofloxacin, and TMP-SMX in the study group were compared to our ED antibiogram. Results: Forty-five patients grew E. coli. Pyelonephritis The data were presented at the American College of Emergency Physicians Scientific Assembly in October 2014. was suspected in nine (20%) subjects. Compared with the ED antibiogram, significantly lower rates of resistance to ciprofloxacin (2% vs. 42%, p < 0.001), levofloxacin (2% vs. 26%, p < 0.001), and TMP-SMX (16% vs. 33%, p = 0.016) were observed. Six patients grew non-e. coli uropathogens. All were susceptible to both levofloxacin and TMP-SMX. Conclusions: ED antibiograms may overestimate resistance rates for uropathogens causing uncomplicated UTIs. In cases where nitrofurantoin cannot be used, fluoroquinolones and possibly TMP-SMX may remain viable options for treatment of uncomplicated UTI and pyelonephritis in women. Ó 2015 Elsevier Inc., Keywords urinary tract infection; antibiotic resistance; antibiogram; community-acquired infection; pyelonephritis INTRODUCTION Uncomplicated urinary tract infection (UTI) is a common indication for antimicrobial therapy in women. The most recent guidelines issued by the Infectious Diseases Society of America (IDSA) for treatment of uncomplicated cystitis and pyelonephritis recommend that a suspected pathogen be <20% resistant to selected empiric treatment (<10% for fluoroquinolones for pyelonephritis), as use of an agent RECEIVED: 6 February 2015; FINAL SUBMISSION RECEIVED: 4 May 2015; ACCEPTED: 12 June 2015 998

E. Coli Resistance Patterns 999 to which the uropathogen has in vitro resistance is associated with clinical failure (1 3). Recent international surveillance studies have demonstrated increasing resistance to fluoroquinolones and trimethoprimsulfamethoxazole (TMP-SMX) among communityacquired uropathogens (4). Therefore, nitrofurantoin is the recommended first-line therapy for uncomplicated cystitis in women. However, in cases where nitrofurantoin cannot be used for UTI (eg, allergy, creatinine clearance < 40 ml/min), or in cases of pyelonephritis, more information is needed to guide the clinician in prescribing effective empiric therapy (5). Urine cultures are not always performed for women presenting to the Emergency Department (ED) with a suspected uncomplicated UTI (6,7). Accordingly, susceptibility data reported in an institution s antibiogram might be skewed toward patients with many comorbidities and higher rates of antimicrobial resistance, and might not relate to the population of healthy women presenting to the ED for treatment of a community-acquired UTI and pyelonephritis (1,7). The same concept applies even for ED-specific antibiograms. In fact, the IDSA guidelines recommend considering local Escherichia coli resistance patterns to guide empiric antibiotic selection for an uncomplicated UTI (1). Although the bedside clinician may understand that the institutional antibiogram represents a population of more complicated patients, there currently are no data available demonstrating the true difference in susceptibility between an antibiogram and otherwise healthy patients who normally are not cultured. At our institution, an ED-specific antibiogram is published separately from the intensive care units (ICU) and other hospital areas. Given that the ED sees a spectrum of patients ranging from the otherwise healthy to the critically ill, it might be expected that the ED-specific antibiogram is less skewed toward the type of patients who harbor resistant pathogens than the overall institution or ICU antibiograms. According to our ED-specific antibiogram, the rates of E. coli resistance to TMP-SMX and fluoroquinolones are >20% for each (1). Although nitrofurantoin can be used in some cases, the need for alternative therapy is frequent, particularly for pyelonephritis. Our study aims to prospectively compare the resistance rate of E. coli isolated from urine of otherwise healthy women who present to the ED with symptomatic uncomplicated UTI or pyelonephritis to the E. coli resistance rate reported in our ED-specific antibiogram. MATERIALS AND METHODS Participants The study took place at a large, urban, academic medical center in Baltimore, Maryland. Institutional review board approval was obtained. Females > 18 years of age who presented to the ED with acute onset of urinary frequency, urgency, or dysuria with pyuria identified on urinalysis (white blood cell [WBC] count > 10/high power field [HPF]) and clinically suspected uncomplicated cystitis or pyelonephritis were asked if they would participate in the study; written consent was obtained from those who agreed. At our institution, patients # 18 years of age are treated in the pediatric ED. The following patients were excluded: those who had a known urologic abnormality or comorbidity, those with an indwelling Foley catheter or who had a catheter removed within the previous 14 days, anyone with a history of kidney stones, those with diabetes mellitus, those who did not speak English, those who had received TMP-SMX prophylaxis within the previous 6 months, anyone with the human immunodeficiency virus (HIV) with no CD4 count on file or a CD4 count < 350 cells/mm 3, those who had been transferred from another health care facility, and those who were pregnant (8). Pregnant women were excluded, as these patients are considered to have complicated infections according to the IDSA guidelines (1). Design The design is a prospective cohort study. E. coli resistance rates in urine cultures from a convenience sample of prospectively enrolled women with uncomplicated UTI were compared to retrospective data from an ED-specific antibiogram from 2 years prior to the study. Because the study period was restricted to staff hours, an attempt was made to identify patients consecutively. The study period was January to July 2014. Measurements Urine cultures were obtained on all urine specimens that demonstrated a WBC count > 10/HPF on urinalysis. At our institution, bacteria colony counts $ 10 3 colonyforming units (CFU)/mL are reported as a positive culture. Urine cultures containing three or more organisms of approximately the same quantity are reported as mixed microbial flora and were considered contaminated and excluded, in accordance with standard microbiology laboratory protocol. All urine specimens were collected by midstream clean-catch technique. Antimicrobial susceptibility testing for levofloxacin and TMP-SMX was performed using Vitek2 (Biomerieux, Durham, NC), and ciprofloxacin mean inhibitory concentration was determined via EtestÒ test (Biomerieux). Mean inhibitory concentration breakpoints for ciprofloxacin, levofloxacin, and TMP- SMX susceptibility were # 1 mg/ml, # 2 mg/ml, and # 38 mg/ml SMX, respectively, in accordance with guidelines from the Clinical Laboratory Standards

1000 M. C. Hines et al. Institute (9). Separate ciprofloxacin susceptibility testing was not performed for non-e. coli isolates. The 2012 adult ED antibiogram is based on cultures of all specimens collected in the ED from both outpatients and admitted patients during that calendar year. Specimen transport, processing, positive-culture CFU cutoff, and susceptibility testing in the ED antibiogram were the same for specimens in the prospective cohort. The 2012 iteration was the most recent antibiogram available at the time of the study. It contains documentation of 337 E. coli isolates. It indicates resistance rates for ciprofloxacin, levofloxacin, and TMP-SMX of 42% (142/337), 26% (88/337), and 33% (111/337), respectively. Notably, the ED-specific antibiogram and the institutional antibiogram are similar with respect to susceptibility patterns of E. coli isolates for these three antibiotics. Statistical Analysis Based on the IDSA recommendation to select an antimicrobial agent with local resistance # 20%, our sample size was intended to detect, with 90% power and 5% significance, a relative decrease of 50% in the 42% ciprofloxacin resistance rate (1). Based on these goals, we calculated a sample size of 45 E. coli isolates. E. coli resistance rates for the three antimicrobials were compared to those from the ED antibiogram using Fisher s exact test. We calculated confidence intervals for proportions and differences between proportions using exact methods or normal binomial approximation, as appropriate. Statistical calculations, including prevalence ratios, were performed using EpiCalc 2000 (Brixton Health, South London, UK). Study Population RESULTS A total of 149 women with clinically suspected uncomplicated cystitis or pyelonephritis were identified between January 1 and July 31, 2014. Figure 1 shows the study flow. Forty-five patients had a urine culture positive for E. coli, and 6 grew non-e.coli uropathogens. The remaining 30 patients had contaminated specimens, no growth, or grew nonpathogenic bacteria (Figure 1).Themedianagewas35years(range,19 83 years). Of the 45 patients who grew E. coli, 35(78%) were black, 9 (20%) were white, and 1 (2%) was of Asian descent. Pyelonephritis was suspected in 9 (20%) subjects, and 22 (49%) had one or more comorbid medical conditions. Figure 1. Inclusion and exclusion pathways for patients eligible for this study. HIV = human immunodeficiency virus; SMX- TMP = trimethoprim-sulfamethoxazole.

E. Coli Resistance Patterns 1001 Table 1. Comparison of the Prevalence of Resistant E. coli for Uncomplicated UTI vs. ED-specific Antibiogram Antibiotic Resistant Per Antibiogram Data (%) Resistant in Uncomplicated UTI (%) Absolute Difference [95% CI] p-value Ciprofloxacin 142/337 (42) 1/45 (2) 0.4 [0.32 0.48] <0.001 Levofloxacin 88/337 (26) 1/45 (2) 0.24 [0.18 0.3] <0.001 TMP/SMX 111/337 (33) 7/45 (16) 0.17 [0.09 0.3] 0.016 UTI = urinary tract infection; ED = emergency department; CI = confidence interval; TMP/SMX = trimethoprim-sulfamethoxazole. Uropathogen Susceptibility We obtained susceptibility data from 45 E. coli isolates. Resistance to ciprofloxacin and levofloxacin was seen in one case, yielding a prevalence estimate of 0.02 (95% confidence interval [CI] 0.0 0.12); resistance to TMP-SMX was seen in seven cases, for a prevalence estimate of 0.16 (95% CI 0.07 0.30). Compared with the ED-specific antibiogram, significantly lower rates of resistance to ciprofloxacin (2% vs. 42%, p < 0.001), levofloxacin (2% vs. 26%, p < 0.001), and TMP-SMX (16% vs. 33%, p = 0.016) were observed (Table 1). The bacteria other than E. coli that were isolated included Proteus mirabilis (4/6), Citrobacter koseri (1/6), and Klebsiella pneumoniae (1/6); three other cultures showed isolates not generally considered pathogenic. The six uropathogens were 100% susceptible to both levofloxacin and TMP-SMX. Ciprofloxacin susceptibilities were not performed on these organisms. DISCUSSION In this prospective observational study in an urban area with high antimicrobial resistance, urinary E. coli isolated from women presenting to the ED with symptomatic uncomplicated cystitis and pyelonephritis exhibited significantly lower resistance rates to ciprofloxacin, levofloxacin, and TMP-SMX compared with those published in the ED-specific institutional antibiogram. Treatment guidelines recommend consideration of local resistance rates when choosing empiric therapy for uncomplicated UTIs. Nine patients (20%) had clinically suspected or radiographically confirmed pyelonephritis, yet still demonstrated low resistance to TMP-SMX and fluoroquinolones. The results of this study have two important clinical implications. First, even ED-specific antibiograms might greatly overestimate resistance rates for uropathogens causing uncomplicated UTIs, namely, E. coli. Despite the fact that the ED-specific antibiogram includes more otherwise healthy patients than a hospital antibiogram, we still found that it did not accurately reflect the uropathogens cultured from our cohort, comprised entirely of otherwise healthy women. Second, in cases where nitrofurantoin cannot be used, fluoroquinolones and possibly TMP-SMX may remain viable options for treatment of uncomplicated UTI and pyelonephritis in selected women. A study by Moffett et al. compared susceptibility data of E. coli obtained from urinary isolates of ambulatory women presenting to three different EDs with a suspected UTI with institution-specific antibiograms (6). No significant difference in susceptibility rates of E. coli urinary isolates to fluoroquinolones or TMP-SMX and antibiogram data was observed. Our study cohort is narrower the Moffett et al. study included Spanish-speaking patients, those with diabetes, those with HIV, and one pregnant patient. Nitrofurantoin is the first-line preferred treatment for uncomplicated cystitis in women (1).However,nitrofurantoin requires a longer treatment course than fluoroquinolones or TMP-SMX, and cannot be used in patients with significant renal impairment or in the setting of suspected pyelonephritis. Fosfomycin, another treatment option with low E. coli resistance rates, is more expensive than fluoroquinolones and TMP-SMX, and is not universally available in the outpatient setting. Beta-lactam agents are appropriate choices when other recommended agents cannot be used, though they are less well studied and require longer courses of therapy. Particularly for pyelonephritis, the IDSA guidelines state that, compared with TMP-SMX, oral beta-lactams have inferior efficacy and higher relapse rates (1). Therefore, despite high in vitro E. coli susceptibility at many institutions, beta-lactams still should be used with caution for pyelonephritis. Studies on oral cephalosporins similarly demonstrate worse clinical cure rates when compared to fluoroquinolones (10). Because the IDSA guidelines recommend that empiric therapy be based on E. coli resistance patterns, we chose to compare the resistance rates of E. coli isolates only to our institutional antibiogram. However, 12% of the uropathogens isolated in the cultures used for this study were not E. coli. The resistance rates of those pathogens could affect empiric therapy decisions. We observed that all of the non-e. coli uropathogens detected through culture were susceptible to levofloxacin and TMP-SMX. The exclusion criteria were designed to enable us to focus on a generalizable population of otherwise healthy women with uncomplicated UTI and pyelonephritis. To increase the clinical utility of our study, we did not exclude patients for whom antibiotics were recently prescribed (other than those on long-term TMP-SMX for an

1002 M. C. Hines et al. immunosuppressed condition), because this information might not be available to the emergency care provider. Additionally, we did not exclude patients based on recent hospitalization, unless a Foley catheter had been removed within the preceding 14 days. We did, however, exclude patients with diabetes, because some clinicians treat UTI in diabetic patients as a complicated infection (1). Limitations Several limitations of this study should be noted. Although consecutive patients were enrolled when study staff was available, a convenience sample was used. We do not anticipate that patients presenting to the ED at times when the study staff was unavailable would be different than those enrolled. Because this was a singleinstitution study, the marked difference in susceptibility patterns between our antibiogram and the study patients may have limited generalizability to all institutions. Antimicrobial resistance varies by region, and medical centers differ in antibiotic prescribing trends, formularies, and stewardship. Excluding non-english-speaking patients could be a potential problem in some regions, because there is some evidence that Hispanic women have higher resistance rates (6). Pregnant women were excluded, as these patients are considered to have complicated infections according to the IDSA guidelines (1). The EDspecific antibiogram does not differentiate the anatomical origin of cultures; therefore, our prospective cohort consisting solely of urinary isolates was compared to a sample that likely contained a minority of specimens obtained from nonurine sources. Finally, resistance data reflected by the antibiogram lag by 2 years, although this is a clinical reality of antibiogram use, and a large shift in resistance is not expected to occur during this timeframe. CONCLUSIONS Institutional, and even ED-specific, antibiograms appear to be skewed toward patients with many comorbidities and higher rates of antimicrobial resistance, and may greatly overestimate the resistance of urinary pathogens in otherwise healthy women presenting to the ED. Based on our results, we suggest reporting separately the E. coli susceptibilities in patients with uncomplicated UTI to prevent the unnecessary use of broad-spectrum antibiotics. In cases of pyelonephritis, fluoroquinolones can still be considered first-line agents according to the IDSA guidelines. At our institution, fluoroquinolone resistance did not cross the 10% threshold outlined by the IDSA guidelines for which an alternative treatment would be needed. Based on the results of this study, fluoroquinolones, and possibly TMP-SMX, can still be considered for treatment of uncomplicated UTI in women when nitrofurantoin cannot be used at our institution. Further comparative investigations of institutional antibiograms to otherwise healthy patients presenting to EDs could have critical implications for antimicrobial stewardship by reserving broad-spectrum antibiotics for specific patients and thus diminishing the threat of drug-resistant microorganisms. Our study suggests that even ED-specific antibiograms can greatly overestimate resistance rates for uropathogens causing uncomplicated UTIs, namely E. coli. Acknowledgment This study was sponsored by a resident research grant from the Maryland Emergency Medicine Network. We thank the microbiology laboratory staff at the University of Maryland Medical Center for performing specimen workup and microbiological analysis. We thank Linda J. Kesselring, MS, ELS, for copyediting the manuscript. REFERENCES 1. Gupta K, Hooton TM, Naber KG, et al. International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelonephritis in women: a 2010 update by the Infectious Diseases Society of America and the European Society for Microbiology and Infectious Diseases. Clin Infect Dis 2011;52:e103 20. 2. Gupta K, Hooton TM, Roberts PL, Stamm WE. Short-course nitrofurantoin for the treatment of acute uncomplicated cystitis in women. Arch Intern Med 2007;167:2207 12. 3. Raz R, Chazan B, Kennes Y, et al. Empiric use of trimethoprimsulfamethoxazole (TMP-SMX) in the treatment of women with uncomplicated urinary tract infections, in a geographical area with a high prevalence of TMP-SMX-resistant uropathogens. Clin Infect Dis 2002;34:1165 9. 4. Kahlmeter G, Poulsen HO. Antimicrobial susceptibility of Escherichia coli from community-acquired urinary tract infections in Europe: the ECO-SENS study revisited. Int J Antimicrob Agents 2012; 39:45 51. 5. Oplinger M, Andrews CO. Nitrofurantoin contraindication in patients with a creatinine clearance below 60 ml/min: looking for the evidence. Ann Pharmacother 2013;47:106 11. 6. Moffett SE, Frazee BW, Stein JC, et al. Antimicrobial resistance in uncomplicated urinary tract infections in 3 California EDs. Am J Emerg Med 2012;30:942 9. 7. Talan DA, Krishnadasan A, Abrahamian FM, et al. Prevalence and risk factor analysis of trimethoprim-sulfamethoxazole- and fluoroquinolone-resistant Escherichia coli infection among emergency department patients with pyelonephritis. Clin Infect Dis 2008;47:1150 8. 8. Padmavathy K, Padma K, Rajasekaran S. Extended-spectrum b-lactamase/ampc-producing uropathogenic Escherichia coli from HIV patients: do they have a low virulence score? J Med Microbiol 2013; 62:345 51. 9. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; twenty-third informational supplement. CLSI Document M100 S24. Wayne, PA: Clinical and Laboratory Standards Institute; 2014. 10. Cronberg S, Banke S, Bergman B, et al. Fewer bacterial relapses after oral treatment with norfloxacin than with ceftibuten in acute pyelonephritis initially treated with intravenous cefuroxime. Scand J Infect Dis 2001;33:339 43.

E. Coli Resistance Patterns 1003 ARTICLE SUMMARY 1. Why is this topic important? Susceptibility data reported in institutional, and even emergency department (ED)-specific antibiograms, might be skewed toward patients with many comorbidities and higher rates of antimicrobial resistance, and thus may not relate to the population of healthy women presenting to the ED for treatment of community-acquired urinary tract infection (UTI) and pyelonephritis. Whereas nitrofurantoin may be an option for some patients, fluoroquinolones and trimethoprim-sulfamethoxazole (TMP-SMX) demonstrate superior clinical cure rates to other options, particularly in pyelonephritis. 2. What does this study attempt to show? This study aims to prospectively compare the resistance rate of Escherichia coli isolated from urine of otherwise healthy women who present to the ED with symptomatic uncomplicated UTI or pyelonephritis to the E. coli resistance rate reported in our ED-specific antibiogram. Fluoroquinolones and TMP-SMX might still be viable treatment options despite high resistance rates in institutional and ED-specific antibiograms. 3. What are the key findings? Compared with the ED antibiogram, significantly lower rates of resistance to ciprofloxacin (2% vs. 42%, p < 0.001), levofloxacin (2% vs. 26%, p < 0.001), and TMP- SMX (16% vs. 33%, p = 0.016) were observed. Even in the patients who grew non-e. coli uropathogens, all were susceptible to both levofloxacin and TMP-SMX. 4. How is patient care impacted? ED antibiograms may overestimate resistance rates for uropathogens causing uncomplicated UTIs. In cases where nitrofurantoin cannot be used, fluoroquinolones, and possibly TMP-SMX, remain viable options for treatment of uncomplicated UTI and pyelonephritis in women.