Bedside Score Predicts Risk of Carbapenem Resistance in Complicated Urinary Tract Infections


In the past 17 years, the number of hospitalizations due to cUTIs has risen by 50%.

A poster presented at the 2018 ASM Microbe meeting in Atlanta, Georgia, has described the preparation and validation of a scoring system that accurately predicts the development of resistance to carbapenem in patients hospitalized with complicated urinary tract infection (cUTI).

Even better, the score can be applied at the bedside to rapidly gauge whether a patient is at risk of infection with a carbapenem-resistant pathogen.

It is a timely study. In the past 17 years, the number of hospitalizations due to cUTIs has risen by 50%. The bacterial culprits are most often Enterobacteriaceae and Pseudomonas aeruginosa. Infections caused by Acinetobacter baumannii are comparatively rare, but often involve strains resistant to carbapenem, so when they occur they are bad.

Carbapenem resistance is increasing. This has been associated with inappropriate therapy that is administered based on experience, which is termed empiric therapy. Getting empiric therapy wrong is a harbinger of a worse outcome.

It is a no-brainer that improved empiric therapy when it comes to treating infections should improve the outcome for cUTI. With this in mind, researchers from OptiStatim, LLC, Melinta Therapeutics, The Medicines Company, and Washington Hospital Center set out to develop a scoring system that could predict carbapenem resistance at the bedside. The results were reported by Marya Zilberberg, MD, MPH, of EviMed Research Group, a research and scientific communication consultancy based in Goshen, Massachusetts,

The researchers exploited the Premier Research database, which contains patient microbiology information from 178 hospitals in the United States. Data from 2009 to 2016 was used. Overall, Escherichia coli was the most frequent cUTI pathogen and accounted for 11.5% of all carbapenem-resistant infections. P aeruginosa was the predominant carbapenem-resistant bacterium (54.5%).

The primary outcome was carbapenem resistance. The cohort of 25,285 patients was split into 2 groups. The first, comprising 60% of the patients, was used to develop the score system. The remaining 40% of the patients were used to validate the system. For ease of use, a maximum of 10 predictors were allowed. The predictors included baseline factors, characteristics of the hospital, and procedures used preceding the infection.

Of the cases, 92% were acquired outside the hospital. Carbapenem resistance was uncommon (5% of cases), with a median rate of resistance for each hospital of 4%. There was no difference in age between patients with carbapenem-resistant cUTIs and those with carbapenem-sensitive infections (mean age 64.1 vs 64.0 years; P = .86).

Differences were apparent between patients with carbapenem-resistant and -sensitive infections in term of significantly higher Charlson score (mean 2.9 vs 2.3; P < .001), male sex (63.0% vs 44.0%; P < .001), prevalence of blacks (18.1% vs 15.0%; P = .011), transfer from an extended care facility (13.2% vs 7.6%; P < .001), greater prevalence of catheter-associated UTIs (76.0% vs 46.8%; P < .001), antibiotic treatment within 90 prior to the detection of the cUTI (50.3% vs 26.1%; P < .001), culture of a carbapenem-resistant isolate in the same 90-day period (11.8% vs 0.9%; P < .001), need for mechanical ventilation during hospitalization (8.0% vs 5.1%; P < .001), and the need for dialysis during hospitalization (2.3% vs 1.5%; P = .025).

Eight factors were used in the scoring system. The points assigned to each factor were based on the weighting resulting from a regression analysis. The factors were admission from an extended care facility (1 point), history of weight loss (1 point), need for mechanical ventilation soon after hospitalization (1 point), age <50 years (2 points), male gender (3 points), catheter-associated UTI (4 points), prior use of antibiotics (4 points), and prior infection with a carbapenem-resistant strain (8 points).

The predictive value of the scoring system was good, particularly in hospitals with low rates of carbapenem-resistant cUTIs. And, because of the design of the study, which included the use of nationwide data, the researchers think that the findings are broadly generalizable.

The study was funded by The Medicine Company and Melinta Therapeutics.


Marya Zilberberg is the CEO of EviMed Research Group.


SUNDAY-705 Predicting Carbapenem Resistance among Gram-Negative Pathogens in Complicated Urinary Tract Infections. Marya Zilberberg, EviMed Research Group LLC, Goshen MA

Brian Hoyle, PhD, is a medical and science writer and editor from Halifax, Nova Scotia, Canada. He has been a full-time freelance writer/editor for over 15 years. Prior to that, he was a research microbiologist and lab manager of a provincial government water testing lab. He can be reached at

Recent Videos
© 2024 MJH Life Sciences

All rights reserved.