Hemodialysis access cost comparisons among incident tunneled catheter patients

Arteriovenous fistula is the ideal hemodialysis access, but most patients start with tunneled dialysis catheter. Arteriovenous fistula and arteriovenous graft surgery may reduce tunneled dialysis catheter use and also increase procedural expenses. We compared Medicare costs associated with arteriovenous fistula, arteriovenous graft, and tunneled dialysis catheter.

Using the US Renal Data System, we identified incident hemodialysis patients in 2008 who started with tunneled dialysis catheter, survived at least 90 days, and had adequate Medicare records for analysis. We followed them until death or end of 2011; access modality was based on billing evidence of arteriovenous fistula or arteriovenous graft creation. We assumed patients without such records remained with tunneled dialysis catheter. We generated multivariate linear regression models predicting Medicare expenditures, censoring costs when patients died; we included all payments to physicians and institutions. We also created algorithms to identify access-related costs.

There were 113,505 patients in the US Renal Data System who started hemodialysis in 2008, of whom 51,002 Medicare patients met inclusion criteria. Of that group, 41,532 (81%) began with tunneled dialysis catheter; 27,064 patients were in the final analysis file. In the first 90 days after hemodialysis initiation, 6100 (22.5%) received arteriovenous fistula, 1813 (6.7%) arteriovenous graft, and 19,151 (70.8%) stayed with tunneled dialysis catheter. Annualized access costs by modality were tunneled dialysis catheter US$13,625 (95% confidence interval: US$13,426-US$13,285); arteriovenous fistula US$16,864 (95% confidence interval: US$16,533-US$17,194); and arteriovenous graft US$20,961 (95% confidence interval: US$20,967-US$21,654; p < .001). Multivariate linear regression demonstrated that staying with tunneled dialysis catheter had lowest access-related costs, arteriovenous fistula was intermediate, and those who underwent arteriovenous graft surgery were highest (p < .021). Access type was not significantly associated with total costs. Additional arteriovenous fistula and arteriovenous graft creation (US$3525 and US$3804 per access per year, respectively) and open and endovascular access-related interventions (US$3102 and US$3569 per procedure per year, respectively; all p < .001) were important predictors of increased cost.

Among patients starting hemodialysis with tunneled dialysis catheter, continued tunneled dialysis catheter use is associated with lowest access-related cost. Both endovascular and open interventions are associated with significant additional costs. Further investigation is warranted to develop efficient patient-centered strategies for hemodialysis access.