Validation of arteriovenous access stage (AVAS) classification: a prospective, international multicentre study.

Abstract

Background: The arteriovenous access stage (AVAS) classification provides evaluation of upper extremity vessels for vascular access (VA) suitability. It divides patients into classes within three main groups: suitable for native fistula (AVAS1) or prosthetic graft (AVAS2), and patients not suitable for conventional native or prosthetic VA (AVAS3). We validated this system on a prospective dataset.

Methods: A prospective, international observational study (NCT04796558) involved 11 centres from 8 countries. Patient recruitment was from March 2021 to January 2024. Demographic data, risk factors, vessels parameters, VA types, AVAS class and early VA failure were collected. Percentage agreement was used to assess predictive ability of AVAS (comparison of AVAS and created VA) and consistency of AVAS assessment between evaluators. Pearson's Chi-squared test was used for comparison of early failure rate of conventional (predicted by AVAS) and unconventional (not predicted by AVAS) VA.

Results: From 1034 enrolled patients, 935 had arteriovenous fistula or graft, 99 patients did not undergo VA creation due opting for alternative renal replacement therapies, experiencing health complications, death or non-compliance. AVAS1 had 91.2%, AVAS2 7.2% and AVAS3 1.6% of patients. Agreement between evaluators was 89%. The most frequently created VAs were radial-cephalic (46%) and brachial-cephalic (27%) fistulae. The accuracy of AVAS versus created access was 79%. In comparison, VA predicted by clinicians versus created access was 62.1%. Inaccuracy of AVAS prediction was more common with higher AVAS classes, and the most common reason for inaccuracy was creation of distal VA despite less favourable anatomy (17%). Patients with unconventional VA had higher early failure rate than patients with conventional VA (20% vs 9.3%, respectively, P = .002).

Conclusion: AVAS is effective in predicting VA creation, but overall accuracy is reduced at higher AVAS classes when the complexity of decision-making increases and proximal vessels require preservation. When AVAS was followed by clinicians, early failure was significantly decreased.