ESRD patients without co-morbid risk factors at the start of haemodialysis are ideal as survival comparison population

BACKGROUND: Our aim was to identify co-morbid risk factors in ESRD patients at the start of the treatment, to select patients in the low-risk group (LRG) and to compare overall survival, adjusted overall survival and LRG survival in three centres (A, B and C). METHODS: Population includes 531 patients entering haemodialysis from 1 January 1981 to 31 December 1996 (mean age 59.6 +/- 16.7 years). Demographics and co-morbidity data collected at the start of HD were independent variables for the analysis. Univariate and multivariate analysis of survival were used to identify significant prognostic factors (Kaplan-Meier and Cox hazard regression model respectively). According to this analysis the LRG was defined by the absence of the identified significant mortality risk factors except age. RESULTS: The overall median survival was 7.92 years, with 92.0% survival at 1 year, 65.2% at 5 years and 40.4% at 10 years. The median survival in centre C (10.83 years) was significantly higher than median survival in centre A (7.0 years) and in centre B (7.83 years). Centres A and B were afterwards analysed together (A-B). In the multivariate analysis, five variables (starting age, diabetes, cancer, smoking habit, and arteriosclerotic heart disease) were associated with survival. The variable centre (A-B or C) was not significant. The adjusted survival curve for centres A-B and C were not different. The LRG included patients of any age, without diabetes, cancer, smoking habit, and arteriosclerotic heart disease. The frequency of the patients in the LRG was 66.3% in centre C and 45.7% in centre A-B (P = 0.0004). Taking into account only the LRG, the survival comparison between centres A-B and C, did not show significant differences (P = 0.196). CONCLUSIONS: We conclude that for purposes of comparison of mortality in ESRD, low-risk population is better than overall ESRD population.