Escherichia coli is the most common cause of community-acquired pyelonephritis and is becoming increasingly resistant to first-line empiric antibiotics. Traditionally, extended-spectrum beta-lactamase (ESBL)-producing E. coli has been associated with healthcare acquisition and rarely originated in the community. Talan and colleagues aimed to determine the prevalence of antibiotic-resistant E. coli in adults with acute pyelonephritis who were seen in U.S. emergency departments (EDs).

The study was conducted using a network of 10 university-affiliated urban EDs called EMERGEncy ID NET. Enrolled patients included those aged ≥ 18 years who presented with flank pain or costovertebral tenderness; fever ≥ 38°C; a presumptive diagnosis of acute pyelonephritis; and a urine specimen that grew a single uropathogen at ≥ 104 CFU/mL. Urine cultures that grew more than one organism were considered to be contaminated and were excluded from the study. Furthermore, patients were classified as having complicated pyelonephritis if they met one of the following criteria: pregnant, male, preexisting urinary tract abnormality, or current immunocompromising condition.

After exclusions, the study population included 521 patients. The median age was 37 years and the majority of patients (87.3%) were female. Most infections were community-acquired (85.6%) and uncomplicated (54.9%). E. coli was the most common pathogen cultured (86.9%). Among those with uncomplicated pyelonephritis, 17 of 272 E. coli isolates (6.3%) were resistant to fluoroquinolones. The range of prevalence by site was 0.0% to 23.1%. For complicated pyelonephritis, 36 (19.9%) were fluoroquinolone-resistant (range by site was 0.0% to 50.0%). ESBL production was found in seven of 272 E. coli isolates (range 0.0% to 8.3%) from cases of uncomplicated pyelonephritis and in 22 of 181 E. coli isolates (range 0.0% to 17.2%) in patients with complicated pyelonephritis. When analyzed by site, the prevalence of ESBL-producing E. coli corresponded with the prevalence of fluoroquinolone-resistant strains. Only 41% of the ESBL-producing strains were susceptible to trimethoprim-sulfamethoxazole, 18% to ciprofloxacin, 21.7% to levofloxacin, and 41.4% to gentamicin.

Another disconcerting finding from the study was the number of patients with pyelonephritis who received the wrong empiric antibiotics. Of the 53 patients with a fluoroquinolone-resistant uropathogen, 24 were treated initially with an antibiotic that was inactive in vitro. Moreover, 22 out of the 29 patients with an infection from ESBL-producing E. coli were started initially on an inactive antibiotic. Of nine patients with ESBL-producing E. coli pyelonephritis who were discharged from the ED, seven were sent out with an in vitro-inactive antibiotic. The investigators did not include any information regarding outcomes in these patients, such as mortality or whether they re-presented to the ED and were then hospitalized. Among the 20 patients admitted with ESBL-producing E. coli pyelonephritis, 15 initially were treated with inactive antibiotics.


Talan and colleagues have presented data that show an alarming increase in the prevalence of antibiotic-resistant E. coli in patients presenting to EDs with pyelonephritis. This unfortunate situation makes the decision about empiric antibiotics more challenging. The latest Infectious Diseases Society of America (IDSA) guidelines for treating pyelonephritis in women recommend a threshold of 20% for using trimethoprim-sulfamethoxazole, and when this number is exceeded, an alternative antibiotic should be used.1 The present study provides solid evidence that the current IDSA guidelines need to be updated. Moreover, since the spread of antibiotic resistance is ongoing and shows no signs of abating, it seems likely that the resistance patterns presented by Talan will soon be out of date as well.

Until new guidelines are published, clinicians should make empiric antibiotic choices based on local resistance patterns. Also, it is important to keep in mind risk factors for fluoroquinolone resistance and the possibility of an ESBL-producing organism, such as recent antibiotic use, travel outside North America, recent hospitalization, and a previous urinary tract infection from a fluoroquinolone-resistant or ceftriaxone-resistant pathogen. However, in the present study, approximately one-third of patients with an ESBL-producing E. coli had none of these risk factors. One rational empiric antibiotic for pyelonephritis, especially if the patient has a risk factor(s) for antibiotic resistance, is ertapenem. All of the uropathogens in the Talan et al study were susceptible to this drug. However, potential drawbacks of ertapenem include disruption of the anaerobic gut flora, no easy oral conversion, and the very real risk of increasing carbapenem resistance. When treating pyelonephritis, clinicians should be pragmatic about the risks and benefits of empiric antibiotics and mindful about antibiogram data for urinary isolates from their community.


  1. Gupta K, 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.