Determinants of multidrug-resistant urinary tract infections: A retrospective cross-sectional study from a tertiary care hospital in southern Bangladesh

Multidrug resistance (MDR) in urinary tract infections (UTIs) presents a growing global health threat, particularly in resource-limited settings like Bangladesh, where context-specific data remain limited. This retrospective cross-sectional study, conducted from January to December 2023 at a tertiary care hospital in Barishal, Bangladesh, aimed to identify key predictors and risk factors associated with MDR in UTI patients. Of 1,670 urine samples received, 229 with significant bacterial growth were included for antimicrobial susceptibility testing using the disk diffusion method. E. coli (55.9%) was the most common isolate, followed by Pseudomonas spp. (20.5%), Klebsiella spp. (14.8%), and Acinetobacter spp. (8.7%). Of these isolates, 70 (30.56%) were found to be MDR-positive. Multivariate logistic regression revealed that male patients (aOR = 2.2; p < 0.05), samples from specialized units (SUs) (aOR = 6.1; p < 0.001), and private medical settings (PMSs) (aOR = 3.1; p < 0.05) were independently associated with increased odds of MDR. Notably, male patients from PMSs showed significantly elevated MDR risk (aOR = 21.8; p < 0.05), indicating a strong predictor for MDR-UTIs. Compared to Acinetobacter spp., E. coli (aOR = 0.3; p < 0.05) and Pseudomonas spp. (aOR = 0.1; p < 0.001) demonstrated lower MDR odds. Multiple Correspondence Analysis (MCA) showed that MDR-positive status and SUs were the main factors contributing to variations in the data, whereas Acinetobacter spp. and Pseudomonas spp. contributed to secondary patterns of variation. These findings underscore the multifactorial nature of MDR-UTIs and emphasize the critical role of healthcare-associated exposure and bacterial species in MDR development. Targeted antimicrobial stewardship, enhanced surveillance, and evidence-based interventions are urgently needed to curb MDR in UTIs within Bangladesh and similar low-resource contexts.