#1922 Patterns and 2-year outcomes of preemptive vascular access creation in France

Abstract

Abstract Background and Aims Most studies on clinical practices regarding the type and the timing of vascular access creation are limited to patients on kidney replacement therapy (KRT) from North American settings. Our aim was to describe patterns of preemptive vascular access creation in patients under nephrology care in France, and to assess the subsequent risks of hemodialysis (HD) initiation with arteriovenous (AV) access use, HD initiation with central venous catheter (CVC) use, initiation of other KRT modality, and competing death. Method From 2013 through 2016, the CKD-REIN cohort included 3033 adult patients with CKD stages 3 to 5 from 40 outpatient nephrology clinics in France, nationally representative geographically and in terms of facility legal status (public, private not for profit, and private for profit). Participants were followed for 5 years or until initiation of kidney replacement therapy (KRT) or death, whichever came first. Patients who underwent a first vascular access creation during follow-up were eligible for this analysis. We described the type, anatomical site and timing of the first vascular access created, as well as patients’ associated characteristics. The number and type of subsequent accesses were also described. Among patients who underwent AV fistula creation first, we used the cumulative incidence functions to estimate the 2-year probabilities of HD initiation with AV access use (independently of its type or order), HD initiation with CVC use, initiation of other KRT modality, and death. Results Over a median follow-up of 5.0 years (interquartile range [IQR], 4.6 to 5.2), 535 patients (18% of the total population) underwent a first vascular access creation. 408 patients (75%) had an AV fistula as their first vascular access, 7 patients (1%), an AV graft, and 127 patients (24%), a CVC. AV fistulas were predominantly located in the lower arm (76%). The median eGFR levels at access creation were 13 (IQR, 10 to 16), 12 (10 to 14) and 9 (6 to 13) ml/min/1.73 m2 for AV fistula, AV graft, and CVC, respectively. Groups defined by their first vascular access did not differ with regards to diabetes status, but patients who received AV fistula as their first vascular access stood out for being younger and having less cardiovascular comorbidities than their counterparts who received AV graft or CVC (Table 1). In contrast to the 7 patients with first AV graft who did not have any subsequent access creation, 22% of those with first AV fistula had at least one subsequent access creation before initiating KRT. The subsequent access was a fistula in 49% of cases (54% of them in the lower arm), a CVC in 48% and a graft in 2%. The 2-year cumulative incidences of HD initiation with AV access use, HD initiation with CVC use, preemptive kidney transplantation and pre-KRT death were 61%, 14%, 2% and 5%, respectively (Fig. 1); 18% of patients undergoing a first AV fistula creation did not experience any event within two years, and 25% did not start HD. No patients received peritoneal dialysis as other KRT modality. Conclusion In this nationwide description of patterns of preemptive vascular access creation, we highlight that lower arm AV fistula is largely prioritized, even after a first failed AV fistula creation. This may contribute to the higher use of AV fistula among HD patients in France compared to North American countries. Most patients who undergo a first AV fistula creation eventually start hemodialysis with an AV access within 2 years of this creation, but 1 out of 4 do not start HD at all (because of death, preemptive transplantation, or stable kidney function). Novel tools assisting clinical decision making in preemptive AV fistula creation are needed.