Hemodialysis vascular sound index as a diagnostic tool for vascular access stenosis: a prospective matched observational study.

backgroundVascular access (VA) monitoring is critical for hemodialysis patients. While vascular ultrasound (US) provides high diagnostic accuracy for VA stenosis, its reliance on equipment and trained operators can limit routine implementation. The Hemodialysis Vascular Sound Index (HVSI), derived from vascular murmur analysis, may provide a simple, objective adjunct for triaging patients who require confirmatory US and/or vascular access interventional treatment (VAIVT).

objectivesThis study evaluated the diagnostic performance of HVSI for detecting VA stenosis and reduced brachial artery flow volume (FV), and examined its association with US-based parameters including resistance index (RI).

methodsThis prospective matched observational study included 202 hemodialysis patients: 101 with clinically significant stenosis requiring VAIVT (study group) and 101 ultrasound-confirmed stable controls (control group). Participants were matched by age, sex, dialysis duration, diabetes status, Kt/V, and blood data using propensity score matching. HVSI was measured using an electronic stethoscope placed over the anastomosis before dialysis. FV and RI were assessed using Doppler US. Diagnostic performance was evaluated using receiver operating characteristic (ROC) analyses, including sensitivity, specificity, and area under the curve (AUC). A small verification cohort (n = 20) was also analyzed to explore the reproducibility of predefined HVSI cutoffs.

resultsHVSI showed significant correlation with FV (R2 = 0.58, p < 0.001) and inverse correlation with RI (R2 = 0.32, p < 0.001). For FV thresholds ≤500, ≤400, and ≤350 mL/min, HVSI showed sensitivities of 86.3–94.4%, specificities of 78.7–82.9%, and AUCs of 0.90–0.94. Diagnostic accuracy tended to be higher in non-bifurcated vessels. In the verification cohort, predefined HVSI cutoffs showed high specificity for FV < 400 mL/min and strong concordance for identifying VAIVT necessity.

conclusionHVSI demonstrated clinically meaningful diagnostic accuracy for reduced FV and VA stenosis in this matched observational cohort. However, because this design compared clinically evident stenosis cases with ultrasound-confirmed stable controls, diagnostic performance may be overestimated relative to consecutive real-world screening populations. Therefore, HVSI should be interpreted not as a replacement for vascular ultrasound, but as a simple screening adjunct to triage patients who require confirmatory ultrasound and/or VAIVT. Further studies in consecutively enrolled cohorts are warranted to validate generalizability and optimize implementation.