Trends and predictors of antimicrobial resistance among patients with urinary tract infections at a tertiary hospital facility in Alexandria, Egypt: a retrospective record-based classification and regression tree analysis

Abstract

Abstract Background The incidence of Antimicrobial Resistance (AMR) in uropathogens varies between countries and over time. We aim to study the patterns and potential predictors of AMR among patients with UTIs admitted to the Urology Department at Alexandria University Hospital. Methods An observational retrospective record-based study was conducted on all patients admitted to the Urology department from October 2018 to October 2020. Data collected from patients’ records included: demographic data, diagnosis on admission, history of chronic diseases, duration of hospital stay, insertion of a urinary catheter, duration of the catheter in days, history of the use of antibiotics in the previous three months, and history of urinary tract operations. If UTI was documented, we abstracted data about urine culture, use of antibiotics, results of urine cultures, type of organism isolated, and sensitivity to antibiotics. We conducted a multivariable logistic regression model. We performed Classification and Regression Tree Analysis (CART) for predicting risk factors associated with drug resistance among patients with UTI. Data were analyzed using SPSS statistical package, Version 28.0, and R software (2022). Results This study encompassed 469 patients with UTIs. The most commonly isolated bacterium was Escherichia coli, followed by Klebsiella pneumoniae. Multidrug resistance (MDR) was found in 67.7% (149/220) of patients with hospital-acquired UTIs and in 49.4% (83/168) of patients with community-acquired UTIs. Risk factors independently associated with antimicrobial resistance according to logistic regression analysis were the use of antibiotics within three months (AOR = 5.2, 95% CI 2.19–12.31), hospital-acquired UTI (AOR = 5.7, 95% CI 3.06–10.76), diabetes mellitus (AOR = 3.8, 95% CI 1.24–11.84), age over 60 years (AOR = 2.9, 95% CI 1.27–6.72), and recurrent UTI (AOR = 2.6, 95% CI 1.08–6.20). Classification and regression tree (CART) analysis revealed that antibiotic use in the previous three months was the most significant predictor for developing drug resistance. Conclusion The study concluded a high level of antimicrobial resistance as well as significant MDR predictors among hospitalized patients with UTIs. It is vital to assess resistance patterns in our hospitals frequently to improve rational antibiotic treatment as well as to sustain antimicrobial stewardship programs and a rational strategy in the use of antibiotics. Empirical therapy for UTI treatment should be tailored to the potential pathogens’ susceptibility to ensure optimal treatment. Strategic antibiotic use is essential to prevent further AMR increases. Further research should focus on suggesting new biological systems or designed drugs to combat the resistance of UTI pathogens.

Keywords

• Urinary tract infection
• Drug resistance
• Microbial
• Risk factors
• Multiple drug-resistant
• Bacteria