Reliability of bioelectrical impedance analysis for body composition assessment: comparison with DEXA in patients with heart failure

Background

Body composition is increasingly recognized as a key prognostic factor, and its assessment is important for risk stratification in patients with heart failure (HF).

The standard reference method is Dual-energy X-ray absorptiometry (DEXA) but recently Bioelectrical impedance analysis (BIA) is widely used in routine clinical practice.

Purpose

To evaluate the reliability of BIA for the assessment of body composition in patients with HF in comparison with the DEXA method.

Methods

We have studied 33 patients with HF NYHA class III-IV, hospitalized in a single center. Information about demographics and disease status was gathered in all cases. Ejection fraction (EF) was estimated by 2D-echocardiography using Simpon’s method. Anthropometric measurements were performed using Tanita device and BMI was calculated. The waist circumference was evaluated at the level above crista iliaca. The total muscle mass, lean mass and fat mass was assessed in all patients by BIA and by DEXA. The significance of the difference in the measured variables was tested using paired sample T test and chi-square for categorical data. To test the agreement between the different methods for body composition assessment we used Spearman’s Correlation method and Intraclass Correlation Coefficient (ICC) with Cronbach’s alpha. P<0.05 was accepted as significant.

Results

The mean age of 33 patients with HF was 70±12.4 years, from 39 to 91. The male sex was dominant – 66.7%. The mean EF was 43.85±15.38% and 45.5% had reduced EF. The BIA systematically overestimated the lean mass and underestimated the total and fat mass compared to DEXA (Table 1). There was relatively good correlation between the two methods. The Spearman’s rho was 0.98 for BMI, 0.98 for total mass, 0.94 for total lean mass, 0.94 for lean mas of extremities, 0.88 for total fat and 0.78 for total fat percent, all significant at p<0.0001. The respective reliability as assessed by ICC was: 0.99 (95%CI 0.988-0.997) for BMI, 0.913 (0.832-0.956) for total mass, 0.864 (0.150-0.96) for lean mass, 0.784 (0.05-0.932) for extremities lean mass, 0.80 (-0.002—0.942) for total fat mass, and 0.634 (-0.061—0.87) for total fat percent, all p<0.0001.

Conclusion

Bioelectrical impedance analysis for assessing body composition tends to deviate from the golden standard DEXA. Nevertheless, the reliability of BIA method to measure the lean and fat body mass is relatively good and can be used for evaluation the body composition in patients with HF. The easier availability and lower cost make this method reasonable alternative to DEXA.For image description, please refer to the figure legend and surrounding text.