Journal Article

Epidemiology of Vascular Access-Associated Infections in Hemodialysis: A Single-Center Retrospective Study in Tbilisi, Georgia (January 2022–January 2025)

Kasradze, Tamar, Tchokhonelidze, Irma, Didbaridze, Tamar

2025

Overview

Background Infections related to vascular access (VA) remain one of the most serious complications in hemodialysis (HD), contributing substantially to patient morbidity and mortality. Central venous catheters (CVCs) confer a substantially higher infection risk compared to arteriovenous fistulas (AVFs), but local epidemiological data are limited. Objective We aimed to investigate the incidence of VA-related infections by access type (CVC vs. AVF, using access-days) and to identify risk factors, with secondary outcomes including infection-related hospitalization and mortality, in HD patients at the largest single center in Tbilisi, Georgia (2022-2025). Methods A retrospective cohort study was conducted on 386 adult HD patients from January 2022 to January 2025. Demographic, clinical, and VA data were analyzed. Cox proportional hazards and logistic regression models assessed associations between VA type, diabetes, and infection or hospitalization, accounting for person-time at risk. Kaplan-Meier curves compared time-to-infection, and incidence rates were calculated per 1,000 access-days with corresponding incidence rate ratios (IRRs) and attributable fractions. Results Out of 386 patients, 208 (53.9%) used CVCs and 178 (46.1%) used AVFs. A total of 116 (30.1%) patients developed bacterial infections, of which 69 (59.5%) were access-related. The infection rate was markedly higher among CVC users compared with AVF users (0.731 vs. 0.042 infections per 1,000 access-days; incidence rate ratio (IRR) = 17.25; 95% confidence interval (CI): 8.21-41.79; p < 0.001). The incidence rate difference was 0.689 infections per 1,000 access-days (95% CI: 0.50-0.88; p < 0.001), with an attributable fraction among the exposed of 94.2% (95% CI: 87.8-97.6). In the univariate Cox regression analysis, CVC use was strongly associated with an increased hazard of bloodstream infections (BSIs) (hazard ratio (HR) = 19.18; 95% CI: 8.72-42.23; p < 0.001) and infection-related hospitalization (HR = 4.10; 95% CI: 2.50-6.71; p < 0.001). Diabetes mellitus was not significantly associated with BSIs (HR = 0.91; 95% CI: 0.55-1.51; p = 0.715) but showed a weak association with hospitalization (HR = 1.04; 95% CI: 1.00-1.07; p = 0.041). Age was not significantly related to either outcome. Infection-related mortality occurred in 23 (19.8%) patients. Conclusions Using CVCs is a significant and modifiable risk factor for VA-related infections in HD patients. Strategies to reduce CVC use focus on timely AVF placement, and strengthening infection control protocols is essential to decrease infection-related morbidity and mortality.