Longitudinal comparison of dialysis access methods: risk factors for failure.

PURPOSE To compare dialysis access patency rates and identify risk factors for failure. METHODS All access procedures at our institution from 1987 to 1996 were reviewed. Primary procedures were surgically implanted dual-lumen central venous hemodialysis catheters (SIHCs), peritoneal dialysis catheters (PDCs), arteriovenous fistulas (AVFs), and prosthetic shunts (PTFEs). RESULTS Five hundred eighty-five primary procedures (236 PTFEs, 87 AVFs, 112 SIHCs, and 150 PDCs) and 259 secondary procedures (215 PTFEs, 14 AVFs, 0 SIHCs, and 30 PDCs) were performed on 350 patients. By life table analysis, SIHCs exhibited the lowest primary patency rate (9% at 1 year; p < 0.0001), whereas PDCs had the highest primary patency rate (57% at 1 year; p < 0.02). The primary patency rates of AVFs and PTFEs was similar, with 43% and 41% 1-year patency rates, respectively (p = 0.70). Less-stringent reporting methods would have increased apparent 1-year patency rates by 9% to 41%. With regard to secondary patency, there was no significant difference between PTFEs and PDCs, with 1-year patency rates of 59% and 70%, respectively (p = 0.62), but PTFEs were more frequently revised. In addition, there was no significant difference between AVF and PTFE secondary patency rates, with 1-year patency rates of 46% and 59%, respectively. Early differences in patency rates for AVFs, PTFEs, and PDCs diminished over time, and at 4 years AVFs had the best secondary patency rate (p = 0.6). The most common reasons for access failure were: PTFEs, thrombosis; AVFs, thrombosis and failure to mature; SIHCs, inadequate dialysis; PDCs, infection and inadequate exchange. By regression analysis, a history of a previous unsalvageable PTFE was the only significant risk factor for failure of a subsequent PTFE (p < 0.01), and the risk of graft failure increased exponentially with the number of previous PTFE shunts. Diabetes was the only significant risk factor for failure of PDCs (p < 0.02; odds ratio, 2.0). CONCLUSIONS The patency rate for PTFEs is similar to that for AVFs, but AVFs require fewer revisions. When replacing a failed access graft, the risk of PTFE failure increases with the number of prior unsalvageable PTFE shunts. PDCs have excellent patency rates, but failure rates are doubled in patients with diabetes. Because of poor patency rates and inadequate dialysis flow rates, SIHCs should be avoided when possible. Reporting methods dramatically affect apparent patency rates, and reporting standards are needed to allow meaningful comparisons in the dialysis access literature.