BACKGROUND: In patients chronically treated with hemodialysis, the prevalence of heart failure is high with a consequently poor prognosis. The role played by blood pressure (BP) on cardiovascular (CV) mortality of these patients has not been clearly defined. METHODS: In this follow-up study, we investigated the relationship of pre- and postdialysis measurements of BP with CV and all-cause mortality in a cohort of 557 dialysis patients with a left ventricular (LV) ejection fraction <50%. RESULTS: During the follow-up (mean = 21.6 ± 8.8 months), 179 deaths were recorded. Ninety-eight patients died from CV causes. By the Cox multivariable analysis, we constructed a predictive model of CV mortality including age, duration on dialysis, diabetes, serum albumin, diffusive dialysis technique, predialysis mean arterial pressure (MAP) (hazard ratio (HR) = 0.978; 95% confidence interval (CI) = 0.956-0.999), and postdialysis MAP (HR = 1.035; 95% CI = 1.010-1.061). The relationship with mortality was inverse for predialysis MAP and direct for postdialysis MAP. In a subsequent analysis, we found that pre- and postdialysis systolic BP, but not diastolic BP, were predictive of CV mortality. Predialysis MAP was in a direct relationship with body mass index. Postdialysis MAP had an inverse relationship with weight loss during dialysis session. CONCLUSIONS: CV mortality in dialysis patients with LV dysfunction is associated with both pre- and postdialysis BP interacting in a complex relationship. Nutritional state and fluid balance and removal are possible clues to this relationship.
BACKGROUND: In patients chronically treated with hemodialysis, the prevalence of heart failure is high with a consequently poor prognosis. The role played by blood pressure (BP) on cardiovascular (CV) mortality of these patients has not been clearly defined. METHODS: In this follow-up study, we investigated the relationship of pre- and postdialysis measurements of BP with CV and all-cause mortality in a cohort of 557 dialysis patients with a left ventricular (LV) ejection fraction <50%. RESULTS: During the follow-up (mean = 21.6 ± 8.8 months), 179 deaths were recorded. Ninety-eight patients died from CV causes. By the Cox multivariable analysis, we constructed a predictive model of CV mortality including age, duration on dialysis, diabetes, serum albumin, diffusive dialysis technique, predialysis mean arterial pressure (MAP) (hazard ratio (HR) = 0.978; 95% confidence interval (CI) = 0.956-0.999), and postdialysis MAP (HR = 1.035; 95% CI = 1.010-1.061). The relationship with mortality was inverse for predialysis MAP and direct for postdialysis MAP. In a subsequent analysis, we found that pre- and postdialysis systolic BP, but not diastolic BP, were predictive of CV mortality. Predialysis MAP was in a direct relationship with body mass index. Postdialysis MAP had an inverse relationship with weight loss during dialysis session. CONCLUSIONS: CV mortality in dialysis patients with LV dysfunction is associated with both pre- and postdialysis BP interacting in a complex relationship. Nutritional state and fluid balance and removal are possible clues to this relationship.