Kidney Week

Abstract: FR-PO776

Which Test Is Best for Detecting Stenosis and Predicting Thrombosis in Arteriovenous Graft? A Diagnostic Accuracy Comparative Study

Session Information

• Hemodialysis: Vascular Access - II
  November 03, 2017 | Location: Hall H, Morial Convention Center
  Abstract Time: 10:00 AM - 10:00 AM

Category: Dialysis

• 603 Hemodialysis: Vascular Access

Authors

• Tessitore, Nicola, AOUI Verona, Nephrology Unit, Verona, Italy

Nicola Tessitore,

• Pessolano, Giuseppina, AOUI Verona, Nephrology Unit, Verona, Italy

Giuseppina Pessolano,

• Lipari, Giovanni, AOUI Verona, Surgery Dpt, Verona, Italy

Giovanni Lipari,

• Mansueto, Giancarlo, AOUI Verona, Diagnostics & Public Health Dpt, Verona, Italy

Giancarlo Mansueto,

• Bedogna, Valeria, AOUI Verona, Nephrology Unit, Verona, Italy

Valeria Bedogna,

• Contro, Alberto, AOUI Verona, Diagnostics & Public Health Dpt, Verona, Italy

Alberto Contro,

• Poli, Albino, AOUI Verona, Diagnostics & Public Health Dpt, Verona, Italy

Albino Poli,

• Lupo, Antonio, AOUI Verona, Nephrology Unit, Verona, Italy

Antonio Lupo,

Background

Guidelines recommend regular screening of grafts for significant >50% stenosis (St) by surveillance (access blood flow[Qa], static venous pressure ratio[sVPR] or Duplex Ultrasound [D]) and state that there is insufficient evidence to prefer one tool over another due to the lack of studies comparing vis-a-vis all of the options

Methods

To identify optimal criteria for St detection and elective repair, we compared in 52 PTFE grafts the Area under Receiver Operator Curve (AUC[95%CI]), sensitivity (SE) & false positive rate (FPR) of clinical monitoring (M), Qa by ultrasound dilution (QaU), sVPR, & D to detect >50% St (StD) and measure Qa (QaD) for >50% St at angiography (StA) and incipient thrombosis (within 4-mo).

Results

Prevalence of StA was 52%. M could not detect StA (AUC 0.60[0.45-0.76]), while all surveillance tools had a similarly significant accuracy for StA: StD (AUC 0.91[0.79-0.97]; SE 85%, FPR 4%), QaU (AUC 0.89[0.77-0.96]; Qa<1300 ml/min: SE 78%, FPR 8%), sVPR (AUC 0.78[0.65-0.89]; sVPR>0.7: SE 89%, FPR 40%), QaD (AUC 0.73[0.59-0.84]; Qa<1300 ml/min: SE 76%, FPR 48%).
31 thromboses occurred during the follow-up. Only QaU (AUC 0.75[0.64-0.84], p<0.001; QaU<1200: SE 68%, FPR 29%), QaD (AUC 0.67[0.53-0.80], p<0.03; QaD<1300: SE 87%, FPR 57%) & StA (AUC 0.61[0.50-0.72], p<0.05; SE 73%, FPR 47%) were equally significant predictors of thrombosis, though their AUC was similar to StD (AUC 0.60[0.46-0.72]) & sVPR (AUC 0.59[0.47-0.70]). At Cox’s multivariate analysis (in a model including StA or StD, QaU or QaD, sVPR, M & acute symptomatic Hypotension during the follow-up) the only significant & independent predictors of incipient thrombosis were Hypotension (with a 4-fold [95%CI:2-18] increased risk, p=0.001), QaU or QaD (with a 20% [95% CI:10-40] greater risk for each 100 ml/min drop in Qa, p<0.01). At univariate analysis, the risk of thrombosis increased significantly at QaU<1200 ml/min (RR 3.7 [95%CI:1.7-8.7], p<0.001) or QaD<1300 ml/min (RR 4.1 [95%CI:0.9-37.1], p<0.05).

Conclusion

Our comparative study suggests that in graft an effective screening program should be based on Qa surveillance (QaU or QaD) & the risk of thrombosis may be contained by avoiding acute hypotension & electively repairing St at a QaU<1200 ml/min or QaD<1300 ml/min.