Unraveling the Length of Hospital Stay for Patients with Urinary Tract Infections: Contributing Factors and Microbial Susceptibility.

Background/Objectives: Length of hospital stay (LOS) is a critical measure of healthcare efficiency. This study investigated factors contributing to prolonged LOS in adult patients with urinary tract infections (UTIs) in Lebanon and assessed microbial susceptibility patterns of causative pathogens. Methods: A retrospective cohort study was conducted across five Lebanese university hospitals (March 2022-December 2023), analyzing 401 patients. Data on microbiological findings and the LOS were extracted from medical records. Statistical analyses, including descriptive statistics, bivariate tests (t-tests, ANOVA, and Pearson's correlation), and multiple linear regression (significance: p ≤ 0.05), were performed using IBM SPSS® 27. Results: The mean LOS was 5.85 ± 2.41 days. Prolonged hospitalization was associated with patient-related factors (age, comorbidities, UTI type, specific symptoms, and multidrug-resistant infections) and treatment-related factors. Empirical use of carbapenems (β = 0.783, p = 0.004) and fluoroquinolones (β = 1.360, p = 0.014), along with inappropriate antibiotic prescriptions (β = 0.609, p = 0.022), significantly extended the LOS. Conversely, antibiotic de-escalation based on culture results reduced the LOS (β = -0.567, p = 0.029). Escherichia coli (61.8%) was the predominant pathogen, followed by Klebsiella pneumoniae (11.9%), Proteus mirabilis (7.8%), and Pseudomonas aeruginosa (7.5%). Notably, susceptibility to antibiotics showed a concerning decline. Conclusions: Inappropriate antibiotic prescriptions were linked to a prolonged LOS, emphasizing the need for judicious antimicrobial use. The positive impact of de-escalation supports culture-guided therapy. Declining antibiotic susceptibility highlights the urgency for robust antimicrobial stewardship programs (ASPs) and a national microbial resistance database to combat antimicrobial resistance (AMR) in Lebanon.