Valve selection in dialysis-dependent patients can be difficult because long-term survival is diminished and bleeding risks during anticoagulation treatment are greater in patients with renal failure. In this study we analyzed long-term outcomes of dialysis-dependent patients who underwent valve replacement to help guide optimal prosthetic valve type selection. Dialysis-dependent patients who underwent aortic and/or mitral valve replacement at 3 institutions over 20 years were examined. The primary outcome was long-term survival. A Cox regression model was used to estimate survival according to 5 ages, presence of diabetes, and/or heart failure symptoms. Four hundred twenty-three available patients were analyzed; 341 patients had biological and 82 had mechanical valves. Overall complication and 30-day mortality rates were similar between the groups. Thirty-day readmission rates for biological and mechanical groups were 15% (50/341) and 28% (23/82; P = .005). Five-year survival was 23% and 33% for the biological and mechanical groups, respectively. After adjusting for age, New York Heart Association (NYHA) class, and diabetes using a multivariable Cox regression model, survival was similar between groups (hazard ratio, 0.93; 95% confidence interval, 0.66-1.29; P = .8). A Cox regression model on the basis of age, diabetes, and heart failure, estimated that patients only 30 or 40 years old, with NYHA class I-II failure without diabetes had a >50% estimated 5-year survival (P < .001). Dialysis-dependent patients who underwent valve replacement surgery had poor long-term survival. Young patients without diabetes or NYHA III or IV symptoms might survive long enough to justify placement of a mechanical valve; however, a biological valve is suitable for most patients.