Objective: To determine the influence of vascular access site on continuous renal replacement therapy (CRRT) filter survival.
Design, setting and patients: Retrospective study of the records of patients who received CRRT in The Alfred intensive care unit from June 2011 to May 2012.
Main outcome measure: Filter run time.
Methods: We matched filter run time to site and type of vascular access. Mean run times were compared using a linear mixed-effects model between: temporary femoral, internal jugular (IJ) and subclavian catheters, tunnelled semipermanent IJ catheters, and extracorporeal membrane oxygenation (ECMO) circuit access. The Markov chain Monte Carlo method was used to construct 95% confidence intervals, and the Wilcoxon rank sum test was used for post hoc testing of significance.
Results: Filter run-time data were available for 131 patients (191 occasions of vascular access) with a total of 870 individual filters analysed. Mean run times were subclavian, 14.4 h; IJ, 17.1 h; femoral, 20.2 h; tunnelled IJ, 25.2 h; and ECMO, 29.0 h. Differences were significant for all combinations except between subclavian and IJ, and between tunnelled access and ECMO. Sites in order of best performing to worst-performing were ECMO circuit, tunnelled IJ, femoral vein, direct IJ vein, and subclavian vein.
Conclusion: Vascular access for CRRT plays a significant role in determining filter life. Our study suggests that for temporary dialysis catheters the femoral site should be favoured in ICU patients, and if CRRT is likely to continue for an extended period, a tunnelled IJ line should be considered.