Percutaneous Management of Dialysis Access Steal Syndrome: Interventions and Outcomes from a Single Institution’s 20-Year Experience

To determine safety and effectiveness of percutaneous interventions performed by interventional radiologists at a single institution over 2 decades in patients with dialysis access steal syndrome (DASS). A retrospective review of fistulograms from 2001 to 2021 (N = 11,658) was performed. In total, 286 fistulograms in 212 patients with surgically created dialysis accesses met inclusion criterion of fistulography for suspected DASS. Chart review collected data regarding patient demographics, comorbidities, access characteristics, fistulography findings, intervention(s) performed, and outcomes. Procedures with and without DASS intervention were compared. Odds ratios (ORs), adjusted for age, sex, comorbidities, access characteristics, and multiple within-patient events, were calculated using logistic regression to determine associations between steal intervention status and outcome variables: (a) major adverse events, (b) access preservation, and (c) follow-up surgery. A percutaneously treatable cause of DASS was present in 128 cases (45%). Treatment of DASS lesions was performed in 118 cases. Fifteen embolizations were also performed in patients without DASS lesions. Technical success of DASS interventions, defined by the Society of Interventional Radiology (SIR) reporting standards, was 94%; 54% of interventions resulted in DASS symptom improvement at a median follow-up of 15 days. Patients with steal intervention had 60% lower odds of follow-up surgery (OR, 0.4; P = .007). There was no difference in major adverse events (P = .98) or access preservation (P = .13) between groups. In this retrospective cohort study, approximately half of DASS fistulograms revealed a percutaneously treatable cause of steal. Over half of DASS interventions resulted in symptomatic relief. Percutaneous intervention was associated with lower odds of follow-up surgery without compromising access preservation. [Display omitted]