The doubly labeled water method produces highly reproducible longitudinal results in nutrition studies

Abstract

The doubly labeled water (DLW) method is considered the reference method for the measurement of energy expenditure under free-living conditions. However, the reproducibility of the DLW method in longitudinal studies is not well documented. This study was designed to evaluate the longitudinal reproducibility of the DLW method using 2 protocols developed and implemented in a multicenter clinical trial-the Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (CALERIE). To document the longitudinal reproducibility of the DLW method, 2 protocols, 1 based on repeated analysis of dose dilutions over the course of the clinical trial (dose-dilution protocol) and 1 based on repeated but blinded analysis of randomly selected DLW studies (test-retest protocol), were carried out. The dose-dilution protocol showed that the theoretical fractional turnover rates for (2)H and (18)O and the difference between the 2 fractional turnover rates were reproducible to within 1% and 5%, respectively, over 4.5 y. The Bland-Altman pair-wise comparisons of the results generated from 50 test-retest DLW studies showed that the fractional turnover rates and isotope dilution spaces for (2)H and (18)O, and total energy expenditure, were highly reproducible over 2.4 y. Our results show that the DLW method is reproducible in longitudinal studies and confirm the validity of this method to measure energy expenditure, define energy intake prescriptions, and monitor adherence and body composition changes over the period of 2.5-4.4 y. The 2 protocols can be adopted by other laboratories to document the longitudinal reproducibility of their measurements to ensure the long-term outcomes of interest are meaningful biologically. This trial was registered at clinicaltrials.gov as NCT00427193.

FIGURE 1

Reproducibility of the DLW method based on the dose-dilution data over 4.4 y (n = 79) in the caloric restriction clinical trial. Shown are the percentage difference of k_H with respect to a theoretical value of 0.10 (A); percentage difference of k_O with respect to a theoretical value of 0.13 (B); and percentage difference of k_O−k_H with respect to a theoretical value of 0.03 (C). The symbols within each panel represent the percentage difference for each monthly dose-dilution measurement. The solid line within each panel represents a mean difference of zero. The 2 dotted lines within each panel represent a 1% difference around zero (A and B), and a 5% difference around zero (C). DLW, doubly labeled water; k_H, fractional turnover rate of ²H; k_O, fractional turnover rate of ¹⁸O.

FIGURE 2

Reproducibility of the DLW method based on outcomes collected from 50 human studies randomly selected from the caloric restriction clinical trial for the test-retest protocol over 2.5 y (n = 50). Shown are the Bland-Altman pair-wise comparison between the retested k_H values and the originally calculated k_H values (A); Bland-Altman pair-wise comparison between the retested k_O values and the originally calculated k_O values (B); Bland-Altman pair-wise comparison between the retested N_H values and the originally calculated N_H values (C); Bland-Altman pair-wise comparison between the retested N_O values and the originally calculated N_O values (D); and Bland-Altman pair-wise comparison between the retested EE values and the originally calculated EE values (E). The solid line within each panel represents zero difference. The dotted line within each panel represents the bias or mean difference between the retest and original values. The 2 dashed lines within each panel represent the 95% CIs of the bias. The symbols within each panel represent the individual difference between the retest and the original values. DLW, doubly labeled water; EE, energy expenditure; k_H, fractional turnover rate of ²H; k_O, fractional turnover rate of ¹⁸O; N_H, isotope dilution space of ²H; N_O, isotope dilution space of ¹⁸O.