This present analysis reports on the secondary endpoints of an 8 week RCT in individuals at risk of the metabolic syndrome (19). The details of that study design, recruitment of participants, conduct of the dietary intervention and pharmacological compounding of Leu and placebo capsules has been previously described in detail (19). Briefly, we had conducted a double blind, placebo controlled, parallel, RCT over 8 weeks, that compared changes in body composition and glucose tolerance during caloric restriction between two groups i.e with Leu supplementation or with placebo (19). The trial was conducted in two phases that included 18 participants in the first phase and 19 in the second phase (19). The study was prospectively registered at Australian New Zealand Clinical Trial Registry (Trial Id: ACTRN12616001528448). Human ethics approval had been obtained from the institution’s ethics committee (HREC 108/2013), and all participants had provided written informed consent.
Energy Expenditure & Metabolic Flexibility
Participants attended the research centre at Curtin University (Bentley, Western Australia) following an overnight fast of a minimum 10 hours, 8 hours of sleep and abstinence from alcohol and vigorous exercise for at least 36 hours prior to the measurement. They were instructed not to shower in the morning to avoid any increase in metabolic rate. On arrival, they emptied their bladder and changed into a standard dressing gown, after which their weight and waist circumference were measured. They entered an insulated chamber measuring 5 x 3 meters with a volume of 57.75 m2 maintained at 25οC and rested in bed for 30 min in the supine position (20). Resting energy expenditure (REE) was measured via indirect calorimetry (Deltatrac II Metabolic Monitor; Datex-Ohmeda, Instrumentarium Corp, Helsinki, Finland). REE and respiratory quotient (RQ) were measured twice for 25 min each with a 10-minute rest in between. Fasting bloods were then drawn by a certified phlebotomist for blood chemistry and this was followed by an oral glucose tolerance test (OGTT) (75g glucose in 300 ml water, Carbo test, Australia). Measurements of REE and RQ were continued from minutes 15–30, 45–60, 75–90 and 105–120 following the OGTT in each person. At the conclusion of the OGTT test, a second blood sample was drawn at 2h. Blood samples were collected for fibroblast growth factor 21 (FGF21), glucose, insulin and liver function tests panel. Metabolic syndrome was determined by the presence of 3 or more metabolic risk factors from the following variables waist circumference, fasting blood sugar levels, fasting (HDL) levels, fasting triglycerides levels or Hypertension (21). Total and regional body composition including fat mass (FM, kg) and fat-free mass (FFM, kg) was measured by (DEXA, Prodigy™, Lunar Corp. Madison, USA). All participants were offered a beverage of choice and light refreshments before they left the premises. The entire protocol was repeated in each individual 8 weeks later, at the end of the weight loss trial.
Calculations:
REE was calculated from the average of last 25 min to min recordings of oxygen consumption (O2) and carbon dioxide (CO2) production using the Weir Eq. (22). RQ was calculated as the ratio of CO2 to O2. The area under the curve (AUC) for postprandial REE and RQ was determined using the trapezoid rule. MF was the calculated as the integrated AUC of RQ (iAUC) given by the difference between the total AUC from baseline AUC. In our laboratory, the within-subject CVs (including measurement error) were as follows: 2.1% for fasting RQ, 3.2% for REE, 1.6% for total area under the glucose curve (TAUC) for postprandial RQ and 4.3% for postprandial energy expenditure (23). Change variables were calculated as 2hr value minus fasting value. Stumvol’s insulin sensitivity index (ISI) (24) and Fatty liver index (FLI) (25) were calculated from their original equations.
Validation Study
The Deltatrac II machine used to measure REE broke down in phase 2 during post leucine supplementation measurements. This affected 12 of 19 participants in that phase. In these 12 participants, we used another available metabolic monitor, TrueOne metabolic cart (Parvo Medics, USA) where gas collection was made via nose-clip and mouthpiece mode as only that was available. The TrueOne shares the same make of oxygen and carbon dioxide analysers as the Deltatrac II. Following repairs to the Deltatrac II, we ran a validation study to determine the difference between the machines and the best predictive equation for converting TrueOne values to Deltatrac II values. In this validation trial, we tested the Deltatrac II in canopy mode to the TrueOne in mouthpiece and nose clip mode. The details of the crossover design and results of the conversion equations for O2 and CO2 from TrueOne to Deltatrac II(25), are presented here in supplementary file S1. (26).
Statistics
Unadjusted repeated measures ANOVA (RM-ANOVA) with treatment as between-subject factor were used to initially explore the change in outcome variables following weight loss. Analysis of covariance (ANCOVA) via General Linear Model (GLM) univariate approach was used to examine the effect of leucine supplementation vs placebo (Treatment) on fasting REE, RQ, and liver function tests and postprandial variables iRQ2, iREE2, GIT (%), FGF21 and Stumvol’s insulin sensitivity index (ISI). All outcomes at the end of the trial were adjusted for their respective pre-trial (baseline) values. Other covariates included were age, gender, phase, caloric deficit (kcal), change in FM, change in FFM, and change in waist circumference. In addition, all 2-way interactions were assessed and, if non-significant, were removed from the model. Normality was assessed and bootstrapping method with 1000 bootstrap samples was used on outcome variables for deriving robust estimates of standard errors, regression coefficients and corresponding 95% confidence intervals. Marginal means were estimated by substituting the mean values of covariates included, and mean value for the corresponding baseline variable.
We ran three sets of analyses for REE and MF that included:
(1) the measured data after machine inter-conversion (n = 37),
(2) a sensitivity analysis that excluded those with a machine failure (n = 25),
(3) an intention to treat analysis (ITT), where the last value was carried forward in those who had a machine failure (N = 37).
Finally, partial correlations assessed relationships between postprandial energy expenditure and RQ with various indices of liver function, ISI and number of metabolic syndrome components. Statistical significance was accepted when a p-value was less than 0.05. All the analyses were performed by using SPSS version 29 (SPSS Inc., Chicago, IL, USA).