Epidemiology, antimicrobial resistance, and mortality of bloodstream infections in hemodialysis patients: an 11-year retrospective cohort study from southern Saudi Arabia

BACKGROUND: Bloodstream infections (BSI) cause major morbidity and mortality in hemodialysis patients, yet contemporary Saudi data on pathogens, antimicrobial resistance, and outcomes remain limited. We described epidemiology, antimicrobial susceptibility, multidrug-resistant (MDR) phenotype prevalence, and unadjusted associations between MDR phenotypes and 30-day in-hospital mortality among maintenance hemodialysis BSI episodes. METHODS: This single-center retrospective cohort included maintenance hemodialysis patients with positive blood cultures at Aseer Central Hospital, southern Saudi Arabia, from January 2013 through July 2024. Analyses were structured by unit of analysis: unique patients, BSI episodes (after 14-day deduplication), isolates, and hospitalized episodes for mortality. Susceptibility used CLSI M100 categorical results. Unadjusted risk ratios with 95% confidence intervals and Fisher exact tests were descriptive; no multivariable model was fitted (sparse events). RESULTS: 201 patients contributed 297 BSI episodes and 389 isolates; 234 hospitalized. Gram-positive organisms predominated (69.0%), led by staphylococci other than Staphylococcus aureus (SOSA, 42.8%), Enterobacterales (21.2%), and S. aureus (17.2%). The 30-day in-hospital mortality was 11.1% (26/234). Linezolid (99.2%) and vancomycin (96.0%) were most active; fluoroquinolone susceptibility was low (45-49%). MDR phenotypes occurred in 24.2% of isolates (MRSA 56.4%, ESC-R 65.0%, CRE 27.3%, VRE 29.2%, CRAB 80.0%). Organism-restricted analysis found CRE-positive episodes with higher mortality than carbapenem-susceptible Enterobacterales (3/8 vs. 1/21), without conventional significance (risk ratio 7.88, 95% CI 0.95-65.06; P = 0.053). CONCLUSIONS: BSI were predominantly gram-positive with a substantial gram-negative and MDR burden. The CRE-mortality association was unadjusted, imprecise, and hypothesis-generating, requiring confirmation in larger adjusted cohorts. Findings support local antimicrobial stewardship and active MDR surveillance.