Temporal and anatomic determinants of central-line-associated bloodstream infection risk in a tertiary PICU: a 3-year time-to-event and competing-risk analysis.

Background: Central line-associated bloodstream infections remain a major cause of preventable morbidity in pediatric intensive care units, yet most prevention strategies treat infection risk as static over catheter dwell time. Objective: We aimed to characterize the temporal and anatomic determinants of central line-associated bloodstream infection in a tertiary pediatric intensive care unit using patient-day survival data. Methods: In this retrospective cohort study, 300 central venous catheter insertions in children admitted to a tertiary pediatric intensive care unit between January 2020 and December 2022 were screened for eligibility. The daily risk of first central line-associated bloodstream infection was modelled using discrete-time pooled-logistic regression with ridge regularization, incorporating baseline covariates, catheter dwell time (0-14, 15-22, and ≥23 days), time-varying clinical markers lagged by 24 hours, and catheter insertion site (femoral, internal jugular, subclavian). Fine-Gray subdistribution hazard models accounted for competing events (catheter removal, discharge, or death). Model performance was evaluated using bootstrap-corrected Harrell's C-index and calibration plots. Results: After exclusions, 270 non-tunneled central venous catheters contributed 4,304 catheter-days. Over 4,304 catheter-days, 13 central line-associated bloodstream infections occurred (incidence 3.0 per 1,000 catheter-days). Compared with dwell time less than 14 days, adjusted odds of infection were 4.33 (95% CI 1.89-9.92) for days 15-22 and 11.71 (95% CI 3.22-42.62) for dwell time of 23 days or more. Right internal jugular insertion was associated with lower adjusted odds relative to femoral access (adjusted odds ratio 0.34; 95% CI 0.14-0.81). Fever in the preceding 24 hours was associated with a threefold higher odds of infection (adjusted odds ratio 3.00; 95% CI 0.94-9.52). The model demonstrated good discrimination (bias-corrected C-index 0.78; 95% CI 0.70-0.86) and acceptable calibration. Conclusion: In this tertiary pediatric intensive care unit, central line-associated bloodstream infection risk was strongly phase-dependent, rising sharply after the second week of catheter dwell and particularly beyond 23 days, and was lower with right internal jugular than femoral insertion. These temporal and anatomic findings support time-triggered line-necessity reviews and careful consideration of insertion site within the broader clinical context when seeking to reduce infection risk in critically ill children.