Purpose: Vascular access (VA) thrombosis, due to significant stenoses (>50%), is the main cause of VA failure in hemodialysis patients. Flow measurements enable detection of stenoses >70% and not >50%. Flow measurements regularly fail to prevent thrombosis. The purpose of the study was to test a new technique for detection of significant stenoses (>50%). Methods and Materials: A pulsatile in-vitro model of a radio-cephalic arteriovenous fistula with silicone tubes, a reservoir and a pump was created. A 15G needle was introduced at 5 and 10 cm downstream of the anastomosis. Intravascular pulse pressure amplitude (systolic minus diastolic pressure = PP) was measured in the arterial inflow and at the arterial needle. PP ratios were calculated (PP-needle/PP-inflow*100%). A 50% stenosis was introduced in the arterial inflow, between needles and in the venous outflow, successively. Measurements were repeated at different heart rates (60-90 beats/min) and different flows (500-1.300 ml/min). ANOVA analysis and post-hoc tests were used to evaluate the relation between the PP ratio and the presence of a stenosis in different conditions. Results: PP ratios were 20.26 ± 4.55% (no stenosis), 7.69 ± 2.08% (arterial inflow stenosis), 36.20 ± 2.12% (between needles stenosis) and 32.38 ± 2.17% (venous outflow stenosis). Stenoses can be located upstream and downstream of the needle (P < 0.001). Between needles stenoses and venous outflow stenoses could also be distinguished (P < 0.001). Conclusion: Pulse pressure analysis enables detection of 50% stenosis independent of heart rate and flow volume. It also enables stenoses localization, in contrast to flow measurements. This promising new method needs clinical validation.