Table 1.
Reliability and validity of body composition measurements in individuals with intellectual disabilities.
| Author(s) | Subjects mean (SD) | Measure | Regression equation | Reference method | Results: reliability and validity | Summary |
|---|---|---|---|---|---|---|
| Usera et al. (2005) [21] | 14 adults with DS; 38(11) yrs. | SKF AGM |
SKF: Jackson and Pollock [22]. AGM: Kelly et al. [23]; Lohman [24] |
ADP with BOD POD |
Validity: correlations with reference method and RMSE: Jackson et al. (r = .54), RMSE = 14.90; Lohman (r = .43), RMSE = 13.20; Kelly and Rimmer (r = .11), RMSE = 9.82. | Lack of validity in 3-field-based methods. New equations for DS recommended |
| Verstraelen et al. (2009) [18] | 76 adults with ID; 19–72 yrs. | BMI WC BIA SKF |
Jackson et al. [25]; Durnin and Womersley [26] |
— | Reliability: Cohen's kappa with 90% CI. Intertest agreement among BMI & WC (0.61). Agreements between BMI SKF & FFM index, WC to SKF & FFM and SKF to FFM (<0.6) | BIA & WC feasible measures. Lack of reliability and large noncompliance for both SKF (n = 5) and FFMI (BIA) (n = 14). |
| Waninge et al. (2009) [19] | 45 severe ID; 38(11) yrs. | BMI WC SKF Tibia length |
— | Reliability: ICC for all variables (95% CI) except SKF (>0.90). | Measuring tibia length possible. Noncompliance and low reliability noted for all SKF measurements |
|
| Temple et al. (2010) [17] | 46 adults mild to mod ID; 19–60 yrs. | BMI | — | DXA | Validity: BMI accounted for 68% of variance in %BF (r
2). Partial correlation of BMI with fat (r = 0.91) and fat-free mass (r = .12) |
BMI reasonable indicator of adiposity. |
| Rieken et al. (2011) [16] | 61 children w/neurol. disability and severe ID; 10(4) yrs. | SKF BIA Tibia length |
SKF: Gurka et al.[27]; Rieken et al. [16] BIA: Pencharz and Azcue [28]; Rieken et al. [16] |
Isotope dilution | Validity: ICC SKF-Gurka et al. [27] mean Difference = −9.2 ± 16.7; ICC = 0.51; SEE = 5.1 kg; R
2 = 0.27; Rieken et al. ICC = 0.59; SEE = 7.6 kg; R
2 =0.44; SEE = 2.2 kg; R
2 = 0.88; BIA- Pencharz and Azcue [28] Mean difference = 2.6 ± 4.4; ICC = 0.94; Rieken et al. ICC = 0.96; SEE 1.7 kg; R 2 = 0.92 |
SKF met with noncompliance (n = 12). Low reliability and validity for SKF compared to BIA. |
| Gonzalez-Aguero et al. (2011) [20] | 28 children with DS; 10–20 yrs. | SKF | Slaughter et al. [29]; Durnin and Womersley [26]; Johnston et al. [30]; Brook [31] |
ADP | Validity: Slaughter et al. [29] (r = 0.105 (P = 0.583)); mean difference = 0.69; 95% CI = 25.8; Durnin and Womersley [26] (r = 0.529 (P < 0.05)); mean difference = 2.34; 95% CI = 18.0; Johnston et al. [30] (r = 0.665 (P < 0.05)); mean difference = 2.73; 95% CI = 19.6; Brook (r = 0.389 (P < 0.05)); mean difference = −2.45; 95% CI = 22.3 | Slaughter's equation most accurate despite wide LOA. Other equations displayed substantial intermethods difference and under- or overestimation of % BF. |
ID: intellectual disability; DS: Down syndrome; AGM: anthropometric girth measurements; SKF: skinfold; BIA: bioelectrical impedance analysis, ADP: air displacement plethysmography; WC: waist circumference; FFMI (BIA): fat-free mass index derived by bioelectrical impedance analysis; %BF: percent body fat; ICC: intraclass correlation coefficients; CI: confidence interval; r: coefficient of correlation; LOA: limits of agreement.