Catheter-related bloodstream infection by catheter type and insertion site

Central venous catheters (CVCs) are indispensable in critically ill patients but carry a risk of catheter-related bloodstream infection (CRBSI). We evaluated CRBSI according to catheter type and insertion site and described its microbiological profile. This prospective observational study was conducted in five intensive care units at a single tertiary-care hospital (Sanliurfa, Türkiye) between January 1, 2025, and January 1, 2026. The unit of analysis was the catheter episode. CRBSI incidence was expressed per 1,000 catheter-days, and adjusted associations with catheter type and insertion site were evaluated using multivariable Poisson regression, cause-specific Cox regression, and post hoc Fine–Gray regression with patient-clustered sandwich standard errors. Among 601 catheter episodes from 455 patients, 385 involved standard triple-lumen CVC and 216 dialysis catheters; 383 were femoral, 188 jugular, and 30 subclavian. Overall, 121 CRBSI events occurred during 8,149 catheter-days, yielding an incidence rate of 14.85 per 1,000 catheter-days (95%-CI, 12.32–17.74). Adjusted models included 473 complete catheter episodes from 367 patients. Standard triple-lumen CVCs had a higher adjusted CRBSI incidence rate than dialysis catheters (adjusted incidence rate ratio [IRR], 2.20; 95%-CI, 1.24–3.91; P = 0.007). Compared with jugular access, neither femoral (adjusted IRR, 1.33; 95%-CI, 0.90–1.96; P = 0.155) nor subclavian access (adjusted IRR, 1.78; 95%-CI, 0.93–3.39; P = 0.081) was clearly associated with CRBSI incidence. Cause-specific Cox analyses yielded similar findings for insertion site. Post hoc Fine–Gray analysis suggested a femoral association and a time-varying catheter-type association. Gram-negative bacteria accounted for 57.0% of CRBSI episodes, with Klebsiella pneumoniae and Acinetobacter baumannii the most frequent isolates. In the adjusted Poisson analysis, standard triple-lumen CVC use was associated with higher CRBSI incidence density than dialysis catheter use. Insertion-site findings differed across models. Gram-negative organisms predominated.