This retrospective analysis used a prospectively maintained database to evaluate predictors of restenosis after stent graft implantation for dysfunctional hemodialysis vascular access. The authors reviewed 203 unique patients treated between 2018 and 2022. The study aimed to identify procedural and clinical factors associated with loss of stent graft patency and loss of access primary patency, using survival analysis and regression modeling.
The study defined geographic miss as either inadequate lesion coverage or suboptimal stent sizing at the time of implantation. A specific, prespecified measure—longitudinal geographic miss (LGM)—was operationalized as residual stenosis greater than 30% within 5 mm of either stent edge. These definitions were used to categorize cases and test whether incomplete lesion coverage or edge residual disease influenced subsequent patency.
Among the 203 patients included, 81% had arteriovenous grafts (AVGs) and 19% had arteriovenous fistulas (AVFs). The dataset represented consecutive implants recorded in the institutional database over the study interval. Device-related variables captured included stent graft diameter and stent graft length. Clinical variables recorded and analyzed included diabetes mellitus, presentation with thrombosis, and procedural complications such as vessel rupture or dissection.
The primary outcomes were loss of stent graft patency and loss of access primary patency. One-year actuarial rates reported were 63.1% for stent graft patency and 22.7% for access primary patency. Time-to-event analyses used Kaplan-Meier estimates to visualize patency over time, and Cox proportional hazards regression models to evaluate associations between candidate predictors and the two endpoints. Both an exploratory multivariable model and an additional clinically adjusted model were reported.
In the primary exploratory multivariable Cox model, longitudinal geographic miss (LGM) was associated with loss of stent graft patency (hazard ratio [HR] 2.62, 95% confidence interval [CI] 1.34–5.12) and with loss of access primary patency (HR 2.76, 95% CI 1.50–5.05). LGM retained its association with both outcomes in a separately described clinically adjusted model.
Other factors identified as associated with worse patency outcomes included:
The report therefore highlights both procedural (LGM, stent sizing and length) and clinical (diabetes, thrombosis) contributors to restenosis risk in the cohort studied.
The authors interpret their findings to suggest that LGM may be a potentially modifiable procedural factor. Because LGM reflects inadequate lesion coverage or suboptimal sizing at the stent edges, attention to complete lesion coverage and appropriate stent sizing during implantation could plausibly reduce restenosis risk. The study emphasizes the importance of procedural technique and device selection when placing stent grafts for hemodialysis access dysfunction.
This work is retrospective in design, using a prospectively maintained institutional database. The authors note that while LGM was associated with patency loss in multivariable models, prospective studies are needed. Specifically, they call for prospective validation with standardized imaging follow-up to confirm LGM as a causal and modifiable predictor of restenosis. Details such as the exact imaging protocol, frequency of follow-up, or external validation cohorts were not reported in this abstract and would require access to the full article for further information.
In 203 hemodialysis access stent graft implants (2018–2022), one-year stent graft patency was 63.1% and access primary patency 22.7%. Longitudinal geographic miss (LGM)—residual stenosis >30% within 5 mm of a stent edge—was independently associated with about a two- to three-fold higher hazard of both stent graft failure and loss of access primary patency. Other clinical and device factors, including diabetes, thrombosis presentation, vessel injury, smaller stent diameter, and stent graft length, were also associated with poorer outcomes. The authors recommend prospective studies with standardized imaging to validate LGM as a modifiable procedural target.