Obesity as a predictor of vascular access outcomes: analysis of the USRDS DMMS Wave II study

Arteriovenous fistulae (AVF) are widely regarded as the preferred vascular access in hemodialysis patients due to their primary patency and patient survival benefits. While the obesity paradox has been associated with improved cardiovascular morbidity and all-cause mortality in dialysis patients, its long-term vascular access outcomes are less clear. Recent literature has suggested that obese patients may have increased early and late fistula failure. The purpose of this study was to explore the relationships between obesity and vascular access outcomes. We performed a retrospective cohort analysis using the USRDS DMMS Wave 2 data set. All incident dialysis patients as of January 1, 1996, over the age of 18, receiving only hemodialysis as mode of renal replacement therapy were eligible for inclusion. Among other variables, data collected for the DMMS Wave 2 included: type and location of vascular access, AVF maturity, vascular access revision, and failure. Logistic regression analyses were used to examine the relationships between obesity and vascular access outcomes, adjusting for important covariates. In all, 1486 hemodialysis patients were included. Using body mass index (BMI) <30 kg/m(2) as reference, obesity did not emerge as a factor in predicting vascular access revisions or failures. An increased risk of AVF failure to mature was found only in the highest BMI quartile (>or=35 kg/m(2)) (aOR 3.66 [95% CI 1.27-10.55], p = 0.017). Peripheral vascular disease was independently associated with an increased risk of AVF failure (aOR 2.78 [95% CI 1.01-7.63], p = 0.047) and arteriovenous graft (AVG) failure (aOR 1.65 [95% CI 1.03-2.64], p = 0.036). Obesity was not associated with increased AVF or AVG revision rates or failure and only associated with poorer AVF maturity at highest BMI quartile. We conclude that obesity should not preclude placement of AVF as vascular access of choice, except in the very obese where assessment should be individually based.