Damage of large arteries is a major factor in the high cardiovascular morbidity and mortality of patients with end-stage renal disease (ESRD). Increased aortic pulse wave velocity (PWV) and brachial pulse pressure (PP) are the principal arterial markers of cardiovascular mortality described in these patients. Whether central (carotid) PP and brachial-carotid PP amplification may predict all-cause (including cardiovascular) mortality has never been investigated. A cohort of 180 patients with ESRD who were undergoing hemodialysis was studied between January 1990 and March 2000. The mean duration of follow-up was 52+/-36 months (mean+/-SD). Mean age at entry was 51.5+/-16.3 years. Seventy deaths occurred, including both cardiovascular and noncardiovascular fatal events. At entry, patients underwent carotid PP measurements (pulse wave analysis), echocardiography, and aortic PWV (Doppler ultrasonography), together with standard clinical and biochemical analyses. On the basis of Cox analyses, after adjustment of age, time on dialysis before inclusion, and previous cardiovascular events, 3 factors emerged as predictors of all-cause mortality: carotid PP, brachial/carotid PP, and aortic PWV. Adjusted hazard ratios for 1-SD increments were 1.4 (1.1 to 1.8) for carotid PP, 0.5 (0.3 to 0.8) for brachial/carotid PP, and 1.3 (1.0 to 1.7) for PWV. Brachial blood pressure, including PP, had no predictive value for mortality after adjustment. These results provide the first direct evidence that in patients with ESRD, the carotid PP level and, mostly, the disappearance of PP amplification are strong independent predictors of all-cause (including cardiovascular) mortality.