PURPOSE: The aim of this study is to determine clinical and ultrasound intra-operative factors related to 1-month autogenous arteriovenous fistula (AVF) thrombosis in end-stage renal failure patients. METHODS: A prospective study was designed, including AVF performed between October 2009 and May 2010. Patient characteristics and intra-operative measurements (clinical and ultrasound findings in both artery and vein: diameters, peak-systolic, end-diastolic and mean velocities, flow and resistance index) were recorded. At 1-month follow-up, AVF primary patency was analyzed. Stepwise logistic regression and ROC curves of the resulting test were used. RESULTS: 111 autogenous end-to-side AVF (44 radiocephalic, 45 brachiocephalic, 22 brachiobasilic) in 101 patients were performed. One-month primary patency rate was 84.7%. Intra-operative absence of bruit following skin closure could predict 1-month AVF thrombosis (70.6% sensitivity and 80.9% specificity, better than absence of thrill: 35.3% and 87.2%). However, logistic regression identified intra-operative end-diastolic velocity in the proximal feeding artery after AVF creation (EDV) as the best independent predictor of 1-month AVF thrombosis (OR=1.072, 95%CI 1.036-1.109; 76.5% sensitivity and 84.0% specificity for EDV<24.5 cm/s). This is a slight improvement on isolated clinical findings, but nevertheless a low positive predictive value (46.4%) is attained. CONCLUSIONS: Prediction of AVF thrombosis with intra-operative ultrasound measurements (proximal artery EDV under 24.5 cm/s) can slightly improve isolated clinical findings, helping to establish an intra-operative criterion to review AVF and increase surgical efficiency, assuming a relatively low positive predictive value.
PURPOSE: The aim of this study is to determine clinical and ultrasound intra-operative factors related to 1-month autogenous arteriovenous fistula (AVF) thrombosis in end-stage renal failurepatients. METHODS: A prospective study was designed, including AVF performed between October 2009 and May 2010. Patient characteristics and intra-operative measurements (clinical and ultrasound findings in both artery and vein: diameters, peak-systolic, end-diastolic and mean velocities, flow and resistance index) were recorded. At 1-month follow-up, AVF primary patency was analyzed. Stepwise logistic regression and ROC curves of the resulting test were used. RESULTS: 111 autogenous end-to-side AVF (44 radiocephalic, 45 brachiocephalic, 22 brachiobasilic) in 101 patients were performed. One-month primary patency rate was 84.7%. Intra-operative absence of bruit following skin closure could predict 1-month AVF thrombosis (70.6% sensitivity and 80.9% specificity, better than absence of thrill: 35.3% and 87.2%). However, logistic regression identified intra-operative end-diastolic velocity in the proximal feeding artery after AVF creation (EDV) as the best independent predictor of 1-month AVF thrombosis (OR=1.072, 95%CI 1.036-1.109; 76.5% sensitivity and 84.0% specificity for EDV<24.5 cm/s). This is a slight improvement on isolated clinical findings, but nevertheless a low positive predictive value (46.4%) is attained. CONCLUSIONS: Prediction of AVF thrombosis with intra-operative ultrasound measurements (proximal artery EDV under 24.5 cm/s) can slightly improve isolated clinical findings, helping to establish an intra-operative criterion to review AVF and increase surgical efficiency, assuming a relatively low positive predictive value.