Background: Traditionally, serum albumin concentration has been used for assessing the nutrition status of hemodialysis patients despite evidence that the level is also affected by inflammation and many other underlying disorders frequently present in these individuals. The authors evaluated albumin as a nutrition parameter, comparing it with more specific anthropometric parameters.
Methods: The study included a cohort of 271 patients. The analysis involved data obtained after patients entered the study (1994-2004). Anthropometric measurements included skinfolds, mid-arm circumference, mid-arm muscle circumference, percentage of body fat, body mass index, body height, and dry weight. Kt/V and normalized protein catabolic rate were also determined and laboratory analyses undertaken.
Results: Serum albumin was only weakly correlated with mid-arm circumference (r = 0.12), mid-arm muscle circumference (r = 0.15), and fat-free mass (r = 0.12). Common factor analysis of nutrition parameters uncovered latent variables, but serum albumin was not associated strongly with them. The sensitivity of albumin in detecting malnutrition was 24%, with a specificity of 88% and a predictive value of 74%. Graphic analysis showed disagreement in albumin levels with percentage of body fat and mid-arm muscle circumference.
Conclusion: Serum albumin determination was shown to be a test with low sensitivity and specificity for evaluating malnutrition in hemodialysis patients. The values correlated weakly and showed graphic disagreement with anthropometric parameters. Therefore, methods that measure percentage of body fat and muscle mass should be used together or instead of serum albumin level for assessing the nutrition status of hemodialysis patients.