Graft flow as a predictor of thrombosis in hemodialysis grafts

Background. The effort to reduce the incidence of graft thrombosis is mainly based on predicting venous stenosis by measuring venous drip chamber pressures. In this study we evaluated whether graft flow measurements, using an ultrasound dilution technique, would be of additional value to identify patients at risk for thrombosis. Methods. In fifty consecutive patients with a bridge graft we measured graft flow and venous drip chamber pressure at a dialyzer blood flow of 200 ml/min. The results of these flow measurements were not used for selection of patients, nor for a diagnostic or therapeutic procedure. All thrombotic events and (radiological or surgical) interventions were registered. Results. A total of 17 patient-years were analyzed. In 17 patients an intervention was done, and in 18 patients thrombosis occurred. The incidence rate of thrombosis was higher in patients with a flow < 600 ml/min (N = 13) compared to patients with a flow > 600 ml/min (N = 37; rate ratio 7.2; 95% CI, range 2.84 to 15.24, P < 0.001). In 4 patients with a flow < 600 ml/min an intervention was done within the first two months after the how measurement. In the remaining 9 patients, 6 grafts thrombosed within this period. Five interventions were done in patients with a flow > 600 ml/min. In the remaining 32 patients only two developed spontaneous thrombosis. Remarkably, venous drip chamber pressure measurements did not discriminate between patients with graft flow > or < 600 ml/min, and showed a wide range in patients who developed spontaneous thrombosis within two months. Conclusion. We suggest that graft flow measurements are helpful in selecting patients at risk for graft thrombosis.