MO015HAEMODIALYSIS VASCULAR ACCESS THROMBOSIS: AN ANALYSIS OF EVENTS FROM THE PROACTIVE IV IRON THERAPY IN HAEMODIALYSIS PATIENTS (PIVOTAL) TRIAL

Abstract

Abstract Background and Aims Haemodialysis vascular access thrombosis (VAT) is a common occurrence that undermines the safe and effective delivery of haemodialysis (HD) and has been associated with anaemia therapy. In the Proactive IV Iron Therapy in Haemodialysis Patients trial (PIVOTAL), proactive high-dose intravenous iron dosing was compared with reactive low-dose intravenous iron dosing. VAT was a pre-specified secondary endpoint. Here we present data detailing the VAT events in PIVOTAL, alongside univariate and multivariate analyses of risk association, testing the hypothesis that VAT may be independently associated with a proactive iron dosing regimen. Method In PIVOTAL 2141 adults with end-stage kidney disease within 12 months of initiating regular HD were randomised 1:1 to either high-dose intravenous iron administered proactively, or low-dose intravenous iron administered reactively, on the basis of monthly Ferritin and Transferrin saturation levels. Erythropoiesis Stimulating Agents were dosed sufficiently to maintain a haemoglobin of 100 to 120g/L. Vascular access use was recorded at baseline and on a monthly basis. VAT was regarded as a safety event with investigators mandated to record such events. VAT was defined clinically, with investigators expected to report thrombotic occlusion of the access (either arteriovenous or catheter thrombosis) significant enough to prevent further effective use without intervention. Baseline demographic, blood pressure and laboratory data underwent univariate analysis with regard to first event of VAT. Primary patient demographics (age, gender, primary renal disease, randomised treatment arm) were put forward for inclusion in a multivariable cox proportional hazards model alongside variables that associated on univariate testing with α<0.05. Analyses were performed using SAS software, version 9.4 (SAS Institute). Results 2141 eligible patients were randomised, of whom 877/2141 (41.0%) were using a central venous catheter (CVC), 1209/2141 (56.5%) an arteriovenous fistula (AVF), and 55/2141 (2.6%) an arteriovenous graft (AVG) at baseline. Over a median follow up of 2.1 years, 480/2141 (22.4%) experienced a VAT event; 194/887 (21.9%) of CVC patients; 263/1209 (21.8%) of AVF patients; and 23/55 (41.8%) of AVG patients. On univariate testing VAT events associated with diabetes as a cause of kidney failure (HR 1.44, 95%CI 1.19 to 1.74, p<0.001), AVG use at baseline (HR 2.37, p<0.001), digoxin use (HR 2.35, p=0.002), and diuretic use (HR 1.24, p=0.02). Angiotensin receptor blocker (ARB) use associated with less VAT (HR 0.66, p=0.013), as did male gender (HR 0.74, p=0.002), and never having smoked (HR 0.68, p=0.008). No association was seen with age, AVF use, CVC use, blood pressure, history of cardiovascular (CV) disease, atrial fibrillation, other CV medications, C-reactive protein, haemoglobin, transferrin saturation, ferritin or randomised treatment group (Proactive 262/1093 (24.0%) v Reactive 218/1048 (20.8%), HR 1.15, p=0.12) . Multivariate analysis found diabetes as a cause of kidney failure (HR 1.45, p<0.001), AVG use (HR 2.29, p<0.001), digoxin exposure (HR 2.48, p<0.001) and diuretic exposure (HR 1.25, p=0.018) to independently associate with VAT. Male gender (HR 0.75, p=0.002), never smoked (HR 0.68, p=0.004) and ARB use (HR 0.66, p=0.011) were found to independently associate with lower risk of VAT. Investigation of interaction between significant predictors found diuretic use in AVF patients to reach nominal significance (p=0.033). No other significant interactions being demonstrated. Conclusion In PIVOTAL, VAT affected nearly one quarter of the cohort over a median of 2.1 years. Diabetes as a cause of kidney failure, active smoking, AVG use, female gender, exposure to digoxin, and exposure to diuretic were independently associated with VAT, but this was not the case with higher iron dosing. ARB use was independently associated with lower risk.