Epidemiological characteristics of congestive heart failure (CHF) have not been well studied in patients with end-stage renal disease (ESRD). We evaluated the prevalence and clinical correlates of CHF using data from Wave 2 of the US Renal Data System Dialysis Morbidity and Mortality Study, a national random sample of incident hemodialysis and peritoneal dialysis patients in 1996 and 1997 (n = 4,024). CHF was recorded as present in 36% of patients. In multivariate analysis, age, female sex, hypertension, diabetes, measures of atherosclerosis, and structural cardiac abnormalities were significantly associated with the presence of CHF. Elevated serum phosphate level >/= 6.8 mg/dL (versus <6.8 mg/dL) and serum calcium level >/= 8.0 mg/dL (versus <8.0 mg/dL) were associated with significantly more CHF (odds ratios, 1.34 and 1.41, respectively), as were low serum albumin (odds ratio, 1.35 per 1-g/dL lower) and low serum cholesterol levels (odds ratio, 1.03 per 20-mg/dL lower). Of elements of pre-ESRD care, frequent visits to a nephrologist (odds ratio, 0.80) or dietitian (odds ratio, 0.84) were associated with significantly lower odds of CHF at the start of ESRD compared with less frequent visits. This national study shows the association of several measures of atherosclerosis and cardiac abnormalities with the presence of CHF at the start of dialysis therapy. It identifies serum albumin as a strong disease correlate and suggests that elevated serum calcium and phosphate levels may be potential risk factors for CHF. This study also suggests that frequent specialist care during this critical period may impact favorably on the prevalence of CHF at the start of ESRD. Future longitudinal studies are required to evaluate the impact of pre-ESRD care on cardiovascular and other clinical outcomes.