Table 3.
Comparison of univariable and multivariable regression models for the prediction of body composition (as determined by BIA).
| Visceral adipose tissue (kg) VAT | |||||||
|---|---|---|---|---|---|---|---|
| Men | Women | ||||||
| N | r2 | RMSE | N | r2 | RMSE | ||
| Age | Age | 1 | 0.22 | 1.48 | 1 | 0.52 | 0.48 |
| Lifestyle | Lifestyle | 3 | 0.27 | 1.46 | 3 | 0.10 | 0.67 |
| BMI | BMI | 1 | 0.59 | 1.06 | 1 | 0.43 | 0.53 |
| BMI + age | 2 | 0.74 | 0.84 | 2 | 0.68 | 0.40 | |
| BMI + age + lifestyle | 5 | 0.75 | 0.83 | 5 | 0.67 | 0.40 | |
| WC | WC | 1 | 0.84 | 0.67 | 1 | 0.74 | 0.63 |
| WC + age | 2 | 0.85 | 0.67 | 2 | 0.79 | 0.33 | |
| WC + age + lifestyle | 5 | 0.84 | 0.67 | 5 | 0.77 | 0.35 | |
| WHtR | WHtR | 1 | 0.77 | 0.80 | 1 | 0.70 | 0.39 |
| WHtR + age | 2 | 0.77 | 0.81 | 2 | 0.73 | 0.37 | |
| WHtR + age + lifestyle | 5 | 0.77 | 0.81 | 5 | 0.70 | 0.39 | |
| Scan variables | Scan variable | 5 | 0.86 | 0.62 | 8 | 0.76 | 0.35 |
| Scan variables + age | 6 | 0.89 | 0.55 | 9 | 0.81 | 0.32 | |
| Scan variables + age + lifestyle | 9 | 0.88 | 0.57 | 12 | 0.79 | 0.33 | |
| Relative fat mass (%) RFM | |||||||
| Age | Age | 1 | 0.21 | 6.28 | 1 | 0.47 | 5.12 |
| Lifestyle | Lifestyle | 3 | 0.24 | 6.14 | 3 | 0.11 | 6.79 |
| BMI | BMI | 1 | 0.46 | 5.13 | 1 | 0.55 | 4.72 |
| BMI + age | 2 | 0.61 | 4.35 | 2 | 0.75 | 3.48 | |
| BMI + age + lifestyle | 5 | 0.62 | 4.31 | 5 | 0.76 | 3.44 | |
| WC | WC | 1 | 0.69 | 3.95 | 1 | 0.75 | 3.48 |
| WC + age | 2 | 0.69 | 3.90 | 2 | 0.79 | 3.20 | |
| WC + age + lifestyle | 5 | 0.69 | 3.92 | 5 | 0.77 | 3.34 | |
| WHtR | WHtR | 1 | 0.63 | 4.27 | 1 | 0.75 | 3.48 |
| WHtR + age | 2 | 0.63 | 4.26 | 2 | 0.77 | 3.39 | |
| WHtR + age + lifestyle | 5 | 0.63 | 4.27 | 5 | 0.75 | 3.50 | |
| Scan variables | Scan variable | 10 | 0.75 | 3.53 | 12 | 0.81 | 3.07 |
| Scan variables + age | 11 | 0.76 | 3.41 | 13 | 0.82 | 2.94 | |
| Scan variables + age + lifestyle | 14 | 0.76 | 3.50 | 16 | 0.81 | 3.10 | |
| Skeletal muscle mass index (kg/m2) SMI | |||||||
| Age | Age | 1 | 0.23 | 1.03 | 1 | 0.20 | 0.71 |
| Lifestyle | Lifestyle | 3 | 0.05 | 1.18 | 3 | 0.07 | 0.78 |
| BMI | BMI | 1 | 0.51 | 0.81 | 1 | 0.23 | 0.70 |
| BMI + age | 2 | 0.76 | 0.56 | 2 | 0.59 | 0.51 | |
| BMI + age + lifestyle | 5 | 0.77 | 0.55 | 5 | 0.63 | 0.48 | |
| WC | WC | 1 | 0.14 | 1.10 | 1 | 0.02 | 0.78 |
| WC + age | 2 | 0.54 | 0.79 | 2 | 0.38 | 0.63 | |
| WC + age + lifestyle | 5 | 0.54 | 0.79 | 5 | 0.43 | 0.61 | |
| WHtR | WHtR | 1 | 0.12 | 1.10 | 1 | 0.02 | 0.79 |
| WHtR + age | 2 | 0.56 | 0.77 | 2 | 0.38 | 0.62 | |
| WHtR + age + lifestyle | 5 | 0.57 | 0.76 | 5 | 0.43 | 0.60 | |
| Girth forearm | Girth forearm | 1 | 0.64 | 0.70 | 1 | 0.28 | 0.67 |
| Girth forearm + age | 2 | 0.74 | 0.59 | 2 | 0.48 | 0.57 | |
| Girth forearm + age + lifestyle | 5 | 0.73 | 0.60 | 5 | 0.50 | 0.56 | |
| Girth Thigh | Girth Thigh | 1 | 0.55 | 0.78 | 1 | 0.38 | 0.62 |
| Girth Thigh + age | 2 | 0.55 | 0.78 | 2 | 0.43 | 0.60 | |
| Girth Thigh + age + lifestyle | 5 | 0.52 | 0.81 | 5 | 0.45 | 0.59 | |
| Scan variables | Scan variable | 8 | 0.83 | 0.48 | 6 | 0.60 | 0.50 |
| Scan variables + age | 9 | 0.85 | 0.44 | 7 | 0.64 | 0.48 | |
| Scan variables + age + lifestyle | 12 | 0.85 | 0.44 | 10 | 0.64 | 0.48 | |
The results are shown for the validated boot strapping models. In the scan variable models, the relevant predictors were selected from 30 scanned standard measurements using stepwise backward model. The fit of each model is given as explained variation (adjusted r2). “Lifestyle” includes 3 variables:education level, physical activity level and diet. “N” indicates the number of variables included in the model.
N numbers of variables included in the model, r2 explained variation, RMSE root mean square error.