Reliability of Anthropometry-Based Equations Compared to Dual Energy Absorptiometry for Assessing Body Composition in Predialysis Chronic Kidney Disease—A Longitudinal Study

Skinfold thickness measurements for assessing body composition are reported to have good reproducibility compared to the reference method of dual energy absorptiometry (DXA). In the current study, we compared the level of agreement between body composition measured with DXA and skinfold thickness (SFT) in CKD Stage 3 and 4, at 2 occasions, 6 months apart. Body composition was assessed in 177 Indian patients with CKD Stage 3 and 4 using DXA and anthropometry (SFT). The body fat mass obtained by the 2 methods was compared by paired t-test, intraclass correlation coefficients, regression analysis, and Bland-Altman plots. A linear regression analysis was done to identify the patient-related parameters which would account for the intermethod differences between DXA and SFT. Compared to DXA, SFT underestimated the fat mass at baseline as well as 6 months [DXA vs. SFT at entry: 15.85 kg (95% confidence interval, CI 15.07-16.65) vs. 13.71 kg (95% CI 13.21-14.32), P < .001; at 6 months: 16.13 (95% CI 15.33-16.93) vs. 13.85 (95% CI 13.25-14.45), P < .001]. The intraclass correlation coefficients at entry and 6 months were 0.894 (0.857-0.921) and 0.896 (0.860-0.923), respectively. The intermethod differences between DXA and SFT at baseline and 6 months were comparable: 2.08 kg (95% CI 1.66-2.5) at baseline versus 2.27 kg (95% CI 1.83-2.71) at 6 months, P = 0.200. Gender and body mass index turned out to be the significant predictors of intermethod differences at base line and exit (P < .001). SFT-based measurements show good reproducibility compared to DXA over a period of 6 months. However, SFT systematically underestimates the fat mass by 2 Kg compared to DXA.