Urea index and other predictors of hemodialysis patient survival.

The mortality of dialysis patients in the United States has been a concern since the US Renal Data System 1989 report, which showed a lower survival rate in the United States compared with Europe. The differences were thought to be multifactorial, including case mix, malnutrition, and adequacy of dialysis. We reviewed the Regional Kidney Disease Program's 1976 to 1989 database for the pattern of co-morbidity, dialysis therapy, and low serum albumin in 1,082 nondiabetic and 691 diabetic patients followed to September 15, 1991. The number of patients over 60 years of age has increased by 15% to 20% per year since 1982. The primary renal diagnoses for nondiabetic patients have shifted from 26% glomerulonephritis, 18% vascular and hypertension, and 15% interstitial kidney disease cases during the period from 1976 to 1982 to 19% glomerulonephritis, 35% vascular and hypertension, and 10% interstitial disease cases during the period from 1986 to 1989. Co-morbidity as single and multiple conditions has increased from 66% to 85% in diabetic patients and from 57% to 66% in nondiabetic patients. The number of patients with two or more co-morbid conditions has increased 1.5-fold in nondiabetic patients and twofold in diabetic patients from the 1976 to 1982 period to the 1986 to 1989 period. Urea index and albumin were determined for these patients and were averaged as a risk factor. A Cox regression analysis was used to determine the relative risk of death of such characteristics as age, renal diagnosis, co-morbidity, urea index (KT/V), and albumin in nondiabetic and diabetic patients. Urea index was divided into ranges of less than 1.0 (mean, 0.9), 1.0 to less than 1.2, 1.2 to less than 1.4, and > or = 1.4 (mean, 1.6). The relative risk of death in nondiabetic patients was 0.65 (P = 0.0012) for KT/V 1.2 to less than 1.4 and 0.67 (P = 0.0029) for KT/V > or = 1.4 (mean, 1.6) compared with 1.0 to less than 1.2 as the baseline. In the diabetic patients, the relative risk of death was 0.70 (P = 0.009) for KT/V 1.2 to less than 1.4 and 0.59 (P = 0.0001) for KT/V > or = 1.4 (mean, 1.6) compared with 1.0 to less than 1.2. In the diabetic patients, KT/V > or = 1.4 also was significantly different from 1.2 to less than 1.4 as it impacted on death rates.(ABSTRACT TRUNCATED AT 400 WORDS)