Abstract
OBJECTIVE: The Kidney Disease Outcome Quality Initiative and Fistula First Breakthrough Initiative call for the indiscriminate creation of arteriovenous fistulas (AVFs) over arteriovenous grafts (AVGs) without providing patient-specific criteria for vascular access selection. Although the U.S. AVF rate has increased dramatically, several reports have found that this singular focus on increasing AVFs has resulted in increased AVF nonmaturation/early failure and a high prevalence of catheter dependence. The objective of this study was to determine the appropriateness of vascular access procedures in clinical scenarios constructed with combinations of relevant factors potentially influencing outcomes. METHODS: The RAND/UCLA Appropriateness Method was used. Accordingly, a comprehensive literature search was performed and a synthesis of results compiled. The RAND/UCLA Appropriateness Method was applied to 2088 AVF and 1728 AVG clinical scenarios with varying patient characteristics. Eleven international vascular access experts rated the appropriateness of each scenario in two rounds. On the basis of the distribution of the panelists' scores, each scenario was determined to be appropriate, inappropriate, or indeterminate. RESULTS: Panelists achieved agreement in 2964 (77.7%) scenarios; 860 (41%) AVF and 588 (34%) AVG scenarios were scored appropriate, 686 (33%) AVF and 480 (28%) AVG scenarios were scored inappropriate, and 542 (26%) AVF and 660 (38%) AVG scenarios were indeterminate. Younger age, larger outflow vein diameter, normal or obese body mass index (vs morbidly obese), larger inflow artery diameter, and higher patient functional status were associated with appropriateness of AVF creation. Older age, dialysis dependence, and smaller vein size were associated with appropriateness of AVG creation. Gender, diabetes, and coronary artery disease were not associated with AVF or AVG appropriateness. Dialysis status was not associated with AVF appropriateness. Body mass index and functional status were not associated with AVG appropriateness. To simulate the surgeon's decision-making, scenarios were combined to create situations with the same patient characteristics and both AVF and AVG options for access. Of these 864 clinical situations, 311 (36%) were rated appropriate for AVG but inappropriate or indeterminate for AVF. CONCLUSIONS: The results of this study indicate that patient-specific situations exist wherein AVG is as appropriate as or more appropriate than AVF. These results provide patient-specific recommendations for clinicians to optimize vascular access selection criteria, to standardize care, and to inform payers and policy. Indeterminate scenarios will guide future research.