IJCS | Volume 32, Nº2, March/April 2019

145 Rocha et al. Nutritional assessment in heart failure Int J Cardiovasc Sci. 2019;32(2)143-151 Original Article right foot at the midpoint between the distal prominences of the radius and ulna of the right wrist and between the medial and lateral malleolus of the right ankle. 13 For the diagnosis of protein malnutrition, the fat free mass index (FFMI) was used, corresponding to the ratio between the amount in kilograms of fat free mass (FFM) obtainedby electrical bioimpedance and the squaredheight (FFM/height²), considering the following cutoff points: 15kg/m² forwomenand17kg/m² formen. 18 Excess fatwas classifiedaccording to sex, age and fat percentage evaluated by the EBI, using the following reference values: forwomen aged 20-39 years (≥ 33%); 40-59 years (≥ 34%); 60-79 years (≥ 36%), formen aged 20-39 years (≥ 20%), aged 40-59 years (≥ 22%); 60-79 years (≥ 25%). 19,20 Statistical analyses In the descriptive analysis of the continuous variables, mean and standard deviation were used. Absolute and percentage frequencies were calculated for the binary and categorical variables. To demonstrate the frequencies of anthropometric variables, absolute frequencies, percentages and their respective 95% confidence intervals were considered. Accuracy assessment was performed by calculating the sensitivity, specificity, positive predictive value, negative predictive value and the proportion of correct anthropometric measurements compared to FFMI for the diagnosis of malnutrition (muscular depletion) and the EBI %FM for the diagnosis of overweight. The formula A=(a+d)/(a+b+c+d) was used. Diagnoses of protein malnutrition and overweight with accuracy and sensitivity above 60% were considered acceptable. 20 For the concordance analysis, theobservedconcordance and the expected concordance for the Kappa statistic calculation were calculated at first. The Kappa index values ​were determined by the concordance classification according to Landis and Koch, 21 which establishes the following classifications: poor (Kappa equal to 0), light (Kappa between 0-0,20), acceptable (Kappa between 0.21-0.40), moderate (Kappa 0.41 to 0.60), substantial (Kappa 0.61 to 0.80) and excellent (Kappa 0.81 to 1.0). Statistical analyses were performed using the software IBM ® SPSS ® 13.0 and Epi InfoTM 7.2.2.2. Results The mean age of the individuals was 55.3 (13) years, predominantly males (70%), and diagnosis of HF with non-ischemic etiology (55%). Significant percentages of individuals with hypertension (70%) were found. Most individuals were in functional class I (76.7%), despite the diagnosis of 60% of the individuals with reduced ejection fraction heart failure (HFrEF). The BMI revealed that the study population presented average values ​indicative of overweight and total body water (TBW) percentage within normality (Table 1). It was found that protein malnutrition, when evaluated by cAMA (43.9%) and AMC (40%), was more frequent when compared to FFMI (13.2%), BMI (11.7%) and APMT (5.1). There was similarity in the diagnoses of excess fat indicators, where BFwas the one that identified a higher frequency of excess fat (63.5) (Table 2). Analyzing the indicators of accuracy of anthropometric measures related to protein malnutrition, it was found that BMI presented low sensitivity (43%) and accuracy (38.5%), but moderate concordance (0.50). On the other hand, AMC presented greater sensitivity (86%) and accuracy (66.4%) with acceptable concordance (0.36) compared to FFMI, being considered acceptable for the diagnosis of protein malnutrition. It was not possible to calculate the accuracy and concordance for the cAMAand APMT indicators because one of the categories presented a null value (Table 3). Regarding overweight, the anthropometric measures related to BF%revealed similarity between the percentage of sensitivity of the TST (60%) and BMI (67%) indicators, as well as low accuracy (20.3% and 22.2%, respectively). Considerable concordance values ​were found for both indicators, compared to BF% (Table 4). Discussion This pioneering study certifies the complexity of anthropometric assessment and body composition in individuals with HF. The results showed that there was a higher frequency of excess fat/weight than of protein malnutrition in our population diagnosed with HF. Among the methods of anthropometric evaluation reviewed, AMC was considered more sensitive for the diagnosis of protein malnutrition, despite the lower concordance with EBI. Indicators of excess fat/adiposity presented similar results and low sensitivity. Comorbidities diagnosed in individuals with HF may coexist or even be the cause of the disease, having a negative impact on the quality of life, increasing the frequency of hospitalization and mortality. 22 In this