Race and diabetes as death risk predictors in hemodialysis patients

Case mix and laboratory predictors of death risk were evaluated in 17,185 hemodialysis patients. The laboratory variables most closely associated with the increased death risk borne by diabetic patients (relative to non-diabetics) and White patients (relative to non-Whites) were identified. The analyses of laboratory death risk predictors were similar to those previously reported. Serum albumin concentration is the most powerful death risk predictor among all of the variables, both case mix and laboratory. Statistical models including only case mix variables reveal both race (RRWhites = 1.42) and diabetes (RRdiabetes = 1.43) as significant predictors. Adding creatinine, albumin, and BUN concentrations to the model eliminated diabetes as a significant predictor. Creatinine and albumin accounted for most of the change. Adding only creatinine eliminated race. The data suggest that reduced visceral and somatic protein mass and/or metabolism may be important determinants of mortality in dialysis patients. Because differences in the concentrations of creatinine and albumin explain much of the risk associated with being White or diabetic, differences in nutritional status may explain the reduced survival observed in those groups. Therefore, clinicians should not simply accept without question the notion that diabetics and Whites are doomed to inferior survival.