MO792: Validation and Applicability of an Innovative Monitoring System for Vascular Access in Haemodialysis—A Multicentre Study

Abstract

Abstract BACKGROUND AND AIMS Routine systems for monitoring vascular access (VA) performance are lacking. We recently showed that with a VA triage system it is possible to improve the average value of a number of HD efficiency parameters and that the triage score was associated with clinical outcomes [1]. The triage categorization is generated monthly by a scoring system which is based on a number of parameters (blood flows, VA pressure values, end HD circuit clots and a dedicated score of the external VA examination) recorded at each session by the staff and by the KT/V value recorded once every month. According to threshold values, each VA is classified as Green (G), Yellow (Y) or Red (R), thus attracting the attention of the staff. In our open-label single centre study, with 3 years of follow-up, the average VA scores improved significantly and G VA associated with lower mortality, compared to the Y and R VA. To validate our triage system with a multicentre, blinded approach in a large number of patients. The primary outcome was to evaluate if the triage identifies VA with an increased rate of complications; the secondary outcome was to verify the association of G VA score with better clinical outcomes. METHOD In this interventional prospective multicentre study, each centre used the electronic spreadsheet without knowledge of the generated score and triage. After 6 months of system implementation, 2 years of follow-up (1 January 2020–1 December 2021) were planned to record VA-related events. Two external reviewers evaluated the records. A minimum of 3 months VA follow-up was necessary for patients’ enrolment. RESULTS From 18 HD centres we enrolled 757 patients, aged 64.5 ± 15.5 years; x/y M/F; 27% diabetics; HD since 24.4 ± 32.4 months; 369 (48.7%) with arteriovenous fistula (AVF) and 388 (51.3%) with permanent central venous catheter (CVC). During 11.4 ± 5.6 months of follow-up (range 3–23), 108 537 HD sessions were recorded on the triage electronic spreadsheet, with 214 total clinical events and an event-free time of 224.5 ± 172 days (range 4–713). The VA-related events were 150 (70.1%) with an event-free time of 230 ± 160 days (range 11–713). As of today, we have the association of VA triage with VA events in a subgroup of 300 patients (66.1 ± 11.7 years; HD vintage 28 ± 18; diabetes 27.6%; 180 with AVF and 120 had CVC) who are not different from the remaining 457. In this subgroup, 54% had G VA triage, 39% Y and 7% R. In the subgroups distinguished by VA type, AVF cases (64.5 ± 12.1 years; HD since 30.3 ± 27.2; 24.4% diabetics) were triaged as G in 54%, Y in 41% and R in 5%, while CVC cases (66.5 ± 11.7 years; HD since 28.8 ± 36.5 months; 34.3% diabetics) were G in 54%, Y in 34% and R in 11%. 83 patients had VA-related events VA during 16.3 ± 2.2 months of follow-up (range 3–23), with an event-free time of 237.8 ± 195.9 days. Figure 1 shows how both AVF and CVC patients, stratified into only two triage groups (G versus Y and R together) according to their average triage, had significantly different time-free from events, which was higher in both populations (AVF P = 0.04; CVC P = 0.001) in the green triage groups. CONCLUSION Our VA triage system identifies 40% of vascular accesses as yellow-red triage, thus deserving clinical surveillance. This score has increased the rate of clinical complications which, according to the time survival curve could be detectable roughly 237.8 days before the event developed.