RESUMO
INTRODUCTION: A successful arteriovenous fistula is essential for effective haemodialysis. We aim to validate the existing failure to maturation equation and to propose a new clinical scoring system by evaluating arteriovenous fistula success predictors. METHODS: Data of end-stage renal disease patients initiated on haemodialysis from January 2010 to December 2012 were retrospectively obtained from medical records with follow-up until 1 January 2014. Application of the failure to maturation equation was evaluated. A nomogram was developed using arteriovenous fistula success predictors and was calibrated with a bootstrapping technique. RESULTS: A total of 694 patients were included with mean duration of follow-up of 2.3 years. Arteriovenous fistula maturation was achieved by 542 patients (78%). Comparing our cohort with the failure to maturation cohort, there were statistically significant differences in mean age, ethnicity and presence of diabetes mellitus. The failure to maturation equation failed to predict arteriovenous fistula outcomes with area under the curve performance of 0.519 on a receiver operating characteristic curve. Multivariate logistic regression showed that Malay patients (odds ratio = 0.628; 95% confidence interval = 0.403-0.978; p < 0.05) and patients requiring preoperative vein mapping (odds ratio = 0.601; 95% confidence interval = 0.410-0.883; p < 0.01) had a lower chance of arteriovenous fistula success, whereas male gender (odds ratio = 1.526; 95% confidence interval = 1.040-2.241; p < 0.05) and presence of postoperative good thrill (odds ratio = 3.137; 95% confidence interval = 2.127-4.625; p < 0.0001) had a higher chance of arteriovenous fistula success. The derived nomogram predicted arteriovenous fistula success (odds ratio = 1.030; 95% confidence interval = 1.022-1.038; p < 0.0001) with the area under the curve of 0.695 on a receiver operating characteristic curve and an adequacy index of 99.86% (p < 0.0001). CONCLUSION: The failure to maturation equation was not validated in our cohort. The clinical utility of our proposed arteriovenous fistula scoring system requires external validation in larger studies.