Are nutritional status indicators associated with mortality in the Hemodialysis (HEMO) Study?

Background. The purpose of this study was to determine if indicators of nutritional status were associated with subsequent mortality in hemodialysis patients. Methods. Twelve selected nutrition indicators were measured prior to randomization in the Mortality and Morbidity in Hemodialysis (HEMO) Study. Relative risks (RR) of mortality were assessed at < 6 months and > 6 months of follow-up using Cox regression after controlling for case mix, comorbidity, and treatment assignment (high vs. standard Kt/V and high vs. low membrane flux). Results. Low values of most nutritional status indicators were associated with increased RR of mortality. RRs were greatest over the short term (< 6 months) and diminished with increasing follow-up (> 6 months). Increases in body mass index (BMI) at lower levels (e.g., <= 25 kg/m(2)) and increases in serum albumin at any level were associated with reduced short-term RR, even after adjusting for case mix, treatment assignment, and for the joint effects of equilibrated normalized protein catabolic rate, total cholesterol, and serum creatinine. For > 6 months' follow-up, increases in values among those with lower levels of BMI and serum albumin (<= 3.635 g/dL) and increases in all serum creatinine levels were associated with lower RR. Conclusion. Nutrition indicators are associated with subsequent mortality in a time-dependent manner, with greatest effects at < 6 months of follow-up. The RR for these indicators may also vary within different ranges of values.