SAT-053 CENTRE-EFFECTS AND INCIDENT HAEMODIALYSIS VASCULAR ACCESS: A BINATIONAL REGISTRY ANALYSIS.

Abstract

An arteriovenous access, including fistula and graft, is preferred to a central venous catheter (CVC) but the majority of patients start haemodialysis with a CVC. The aims of this study were to examine patient- and centre-level factors associated with incident arteriovenous access versus CVC use and the contribution of these factors to centre variation in incident arteriovenous access use. Using data from the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry, all adult (≥18yo) patients commencing haemodialysis between 1 January 2004 and 31 December 2015 were included. The primary outcome, vascular access type (arteriovenous access versus CVC) at haemodialysis initiation, was analysed using logistic regression models with a random effect for dialysis centre. Patient-level factors examined included sex, age, race, body mass index (BMI), smoking status, primary kidney disease, late referral to nephrology (<3 months), cardiovascular disease (ischaemic heart disease, cerebrovascular disease, peripheral vascular disease), diabetes mellitus, chronic lung disease, and previous renal replacement therapy (RRT). Centre-level factors included centre size, transplantation facility status, percentage of home haemodialysis patients, mean weekly haemodialysis hours, and percentage of patients achieving target blood flow rates, phosphate levels, haemoglobin, and weekly Kt/V. The study included 27,123 patients from 61 centres. The mean proportion of patients starting haemodialysis with an arteriovenous access varied from 15% to 61% (median 39%). Patient-level factors associated with an increased likelihood of starting haemodialysis with an arteriovenous access were male sex, BMI >25kg/m2and polycystic kidney disease as primary kidney disease. Patient-level factors associated with a decreased likelihood of starting haemodialysis with an arteriovenous access included BMI<18.5kg/m2, late nephrologist referral, diabetes mellitus, cardiovascular disease, chronic lung disease, and prior RRT. No centre-level factors associated with increased incident arteriovenous access use were identified. Lower achievement in phosphate targets and higher proportion of home haemodialysis were associated with a decreased likelihood of initiating haemodialysis with an arteriovenous access. Reductions of variation in incident arteriovenous access use across centres after adjustment for patient- and centre-level factors were modest (Figure 1). Figure 1.Variation of arteriovenous access use (expressed as log odds ratios) in incident haemodialysis patients across 61 centres during the period of 2004 through 2015 in unadjusted (green), patient-level adjusted (orange) and multilevel (patient and centre) (blue) models with standard errors. Dialysis centres are ranked by log odds ratio of incident arteriovenous access use. Incident arteriovenous access use varied 4-fold across centres and was not explained by the patient- and centre-level characteristics currently captured in the ANZDATA registry. Identification of processes of care and surgical factors may account and allow for modification of inter-centre variation.