The obesity paradox in heart failure (HF) is criticized because of the limitations of body mass index (BMI) in correctly characterizing overweight and obese patients, necessitating a better evaluation of nutritional status. The aim of this study was to assess nutritional status, BMI, and significance in terms of HF survival. Anthropometry and biochemical nutritional markers were assessed in 55 HF patients. Undernourishment was defined as the presence of <2 of the following indexes below the normal range: triceps skinfold, subscapular skinfold, arm muscle circumference, albumin, and total lymphocyte count. Patients were also stratified by BMI and followed for a median of 26.7 months. Across BMI strata, no patient was underweight, 31% were normal weight, 42% were overweight, and 27% were obese. Undernourishment was present in 53% of normal-weight patients, 22% of overweight patients, and none of the obese patients (p = 0.001). Undernourished patients had significantly higher mortality (p = 0.009) compared to well-nourished patients. In multivariate analysis, only undernutrition (hazard ratio 3.149, 95% confidence interval 1.367 to 7.253), New York Heart Association functional class (hazard ratio 3.374, 95% confidence interval 1.486 to 7.659), and age (hazard ratio 1.115, 95% confidence interval 1.045 to 1.189) remained in the model. Among nutritional indicators, subscapular skinfold was the best predictor of mortality; patients with subscapular skinfold in the fifth percentile had higher mortality (p = 0.0001). In conclusion, BMI does not indicate true nutritional status in HF. Classifying patients as well nourished or undernourished may improve risk stratification. © 2011 Elsevier Inc.