By Richard R. Watkins, MD, MS, FACP, FIDSA

Associate Professor of Internal Medicine, Northeast Ohio Medical University; Division of Infectious Diseases, Cleveland Clinic Akron General, Akron, OH

Dr. Watkins reports no financial relationships relevant to this field of study.

SYNOPSIS: The authors of a multicenter, retrospective, cohort study from southern and eastern Europe identified predictive factors for multidrug-resistant complicated urinary tract infections.

SOURCE: Gomila A, Shaw E, Carratalà J, et al. Predictive factors for multidrug-resistant gram-negative bacteria among hospitalised patients with complicated urinary tract infections. Antimicrob Resist Infect Control 2018;7:111.

Hospitalized patients with suspected complicated urinary tract infection (cUTI) often are treated initially with broad-spectrum antibiotics. Using a model to predict which patients are at high risk for multidrug-resistant (MDR) pathogens would be useful to help make better antibiotic choices, thereby leading to improved antibiotic stewardship.

Gomila et al conducted a retrospective cohort study from hospitals in southern and eastern Europe, Turkey, and Israel. They defined MDR as nonsusceptibility to at least one drug in three or more antimicrobial categories. The investigators obtained data from patients who were admitted for a cUTI and from those who were admitted for other reasons and developed a cUTI during hospitalization. Each hospital contributed 50-60 consecutive patients to reduce selection bias. Inclusion criteria were age > 18 years; patients with a UTI with an indwelling Foley catheter, a neurogenic bladder, urinary retention, renal impairment with a glomerular filtration rate < 60 mL/min, a renal transplant, pyelonephritis, or an ileal loop/pouch; signs or symptoms of a UTI; and a urine culture with > 105 colony-forming units of a uropathogen or at least one positive blood culture growing a possible uropathogen and no other evident site of infection (e.g., an intra-abdominal source or pneumonia). Patients were excluded if a polymicrobial urine culture returned positive, a culture revealed Candida spp. as the sole isolate, received a diagnosis of prostatitis, or experienced an uncomplicated UTI. The primary outcome was the presence of an MDR cUTI. The secondary outcomes included estimates of the prevalence of MDR in each country, definition of the cUTI microbiology, and assessment of the resistance rates of the uropathogens to different classes of antibiotics.

A total of 948 patients met inclusion criteria, from which investigators obtained 1,074 bacterial isolates. Most patients (56%) were female, and the mean age was 65.8 years. The most frequent isolates were Escherichia coli (52%), with 14.5% MDR; Klebsiella pneumoniae (15.6%), with 54.2% MDR; Pseudomonas aeruginosa (9%), with 38.1% MDR; Proteus mirabilis (7%), with 24.1% MDR; and Enterococcus spp. (3%), with no MDR rate given. Carbapenem resistance occurred in 2.3% of E. coli, 19.6% of K. pneumoniae, and 32.6% of P. aeruginosa. The MDR rate was < 20% in Hungary and Spain and approximately 60% in Bulgaria and Greece. In a final predictive model, factors that predicted MDR were male sex (odds ratio [OR], 1.66; 95% confidence interval [CI], 1.20-2.29), acquiring a cUTI in a healthcare facility (OR, 2.59; 95% CI, 1.80-3.71), presence of a Foley catheter (OR, 1.44; 95% CI, 0.99-2.10), contracting a UTI during the previous year (OR, 1.89; 95% CI, 1.28-2.79), and receiving an antibiotic in the preceding 30 days (OR, 1.68; 95% CI, 1.13-2.50).


Antibiotic misuse is the main culprit driving the global spread of antimicrobial resistance. UTIs are diagnosed in hospitalized patients in Europe and North America commonly. Many patients receive inappropriately broad empiric antibiotic therapy. Thus, the Gomila et al study is significant and useful because it helps define which patients are at risk for a cUTI due to an MDR pathogen and inform antibiotic choices. Hospitalized patients at high risk for an MDR pathogen might be started empirically on a broader agent (e.g., a carbapenem, ceftazidime/avibactam, or meropenem/vaborbactam), while those at intermediate or low risk might receive a narrower spectrum drug (e.g., third-generation cephalosporin or a quinolone). De-escalation is appropriate when culture data become available except in a few circumstances, such as septic shock or neutropenic fever. A recent study showed many of the same risk factors also to be associated with MDR in healthy young adults with community-acquired UTIs.1 Of note, nitrofurantoin was an effective choice and carried a low risk for inducing MDR.

There were some limitations to this study. First, it included hospitals in several countries that reported higher rates of MDR pathogens compared to the United States. Therefore, further validation of the results for different regions is needed. Second, not all important risk factors, such as history of an MDR pathogen, were included in the analysis. Finally, because of the retrospective observational design, the study may have been influenced by unmeasured confounding variables. The predictive factors Gomila et al identified could serve as the basis for developing a score to identify patients at high risk for MDR cUTIs. Deciding about empiric antibiotic therapy requires balancing the need for an active drug with the risks associated with coverage that is too broad, such as promoting antimicrobial resistance and Clostridioides difficile infection. This study is a significant contribution to the evidence for treating cUTIs and should help clinicians choose appropriate empiric antibiotic therapy.


  1. Brosh-Nissimov T, Navon-Venezia S, Keller N, Amit S. Risk analysis of antimicrobial resistance in outpatient urinary tract infections of young healthy adults. J Antimicrob Chemother 2018; doi: 10.1093/jac/dky424. [Epub ahead of print].