SGA scores have poor correlation with serum albumin in obese hemodialysis patients: a secondary analysis.

OBJECTIVE: The objective of this study was to determine the relationship between serum albumin and the Subjective Global Assessment (SGA) in a sample of obese hemodialysis (HD) patients. DESIGN AND METHOD: Study subjects (N = 253) included patients who were categorized into well-nourished (68%, SGA score 6-7) and malnourished (score 1-5) groups, and, on the basis of the body mass index (BMI), into obese (BMI > 30 kg/m(2); 35%) and nonobese (BMI < 30 kg/m(2)). The mean baseline data (±standard deviation) were as follows: age, 63.5 ± 14.3 years; BMI, 29 ± 8 kg/m(2); and serum albumin, 3.8 ± 0.4 mg/dL (bromocresol green). The secondary analysis of data from the SGA Validation Project and Nutrition Algorithm Preliminary Report determined the relationship between the 7-point SGA and serum albumin concentrations in a sample of obese HD patients. Data were analyzed at Case Western Reserve University from a total of 253 HD patients. MAIN OUTCOME MEASURE: The SGA scores in the BMI groups were compared with serum albumin as an objective measure of nutrition and inflammation risk. RESULTS: By using analysis of variance, the obese and nonobese populations showed statistically significant differences in SGA scores (obese: P < .3468, nonobese: P < .0080) and serum albumin values (obese, P < .0943; nonobese, P < .0183) between well-nourished and malnourished groups. A multivariable nominal logistic fit was used to predict SGA group with age and gender as covariates. Serum albumin values (P < .0057) and BMI values greater than 30 (P < .0090) predicted SGA group. CONCLUSION: The secondary analysis showed that SGA does not correlate well with serum albumin; thus, it may not be a valid nutrition assessment tool among obese HD patients.