Patients on dialysis have an age-adjusted death rate 3.5 times that of
the general population. The most common cause of death in patients on
dialysis is cardiovascular disease. We prospectively followed a cohor
t of 433 patients in three centers for a mean of 41 months. Mean hemog
lobin level at the beginning of dialysis was 8.39 (+/-1.7) g/dL, and t
he mean hemoglobin level during follow-up was 8.84 (+/-1.5) g/dL. Usin
g Cox's regression model, we found that anemia predicted mortality ind
ependently of age, diabetes mellitus, cardiac failure, hypoalbuminemia
, serum creatinine, mean arterial pressure, or echocardiographic heart
disease. The independent relative risk (RR) of mortality was 1.18 per
1.0 g/dL decrease in hemoglobin level. Anemia also independently pred
icted the de novo occurrence of congestive heart failure when the same
covariates were controlled for (RR, 1.49 per 1.0 g/dL decrease). Anem
ia was also independently predictive of heart failure at the start of
dialysis (RR, 1.14 per 1.0 g/dL decrease) and heart failure recurrence
(RR, 1.25 per 1.0 g/dL decrease). Left ventricular hypertrophy is pre
sent in 75% of patients on dialysis at the start of therapy for end-st
age renal disease. It independently predicts mortality. Our prospectiv
e cohort study identified increasing age, hypertension, and anemia as
risk factors for its development One controlled study and several unco
ntrolled studies demonstrated improvement (but not complete regression
) of elevated left ventricular mass in patients on dialysis treated wi
th recombinant human erythropoietin (epoetin). (C) 1995 by the Nationa
l Kidney Foundation, Inc.