Study subjects and design
The methods have beedn described in detail previously [14]. Breinfly: This randomised crossover study group included 20 men with T2D, 20 obese men and 20 healthy men. All individuals consumed two energy- and macronutrient-matched test meals in random order. The general metabolic characteristics of each group are given in Table 1. Written informed consent was obtained from all participants prior to enrolment in the study. The study was approved by the ethics committees of Thomayer Hospital and the Institute for Clinical and Experimental Medicine, Prague, Czech Republic (protocol identification number G14-08-42). The study was prospectively registed at ClinicalTrials.gov (Identifier: NCT02474147).
All participants were male and of Czech nationality. Men with T2D and at least three hallmarks of metabolic syndrome (30-65 years of age; BMI, 25-45 kg/m2; HbA1c, 42-105 mmol/mol) were treated by lifestyle alone or with oral hypoglycaemic agents (metformin and/or sulfonylureas) for at least one year. The obese men were BMI- and age-matched to men with T2D; the healthy men comprised age-matched controls with normal BMI (between 19 and 25 kg/m2) and normal glucose tolerance. Exclusion criteria were thyroid, liver or kidney disease, drug or alcohol abuse, unstable drug therapy, or significant weight loss of more than 5% body weight in the preceding three months.
Randomization and interventions
The participants were randomly assigned in a 1:1 ratio a vegan meal or an energy- and marconutrient-matched conventional meat containing meal based on a computer-generated randomisation protocol. The randomization protocol was designed so as not to be accessed beforehand, with the interventions unmasked. Outcome assessors were blinded to the interventions.
All participants fasted for 10 to 12 hours overnight. Diabetic patients avoided diabetes medication the evening before, and on the morning of, the assessment. The meal consisted of either a conventional meat burger (M-meal) or a vegan burger (V-meal). The composition of the test meals is given in Table 2. Tap water was allowed ad libitum. Unaware of the sequence of interventions, participants arrived at the laboratory in the morning to be assigned one of the randomized test meals. The participants were checking in for their meal assessments between 7-8:30 am. They usually finished their meal in 15-20 mins. The whole meal test took 3 hours from finishing their meal. After a washout period of one week, participants returned to complete the opposite test meal. The participants were instructed not to change their usual dietary habits or physical activity during the study. All participants found both meals acceptable and nobody complained about any particular meal.
Analytical methods
Anthropometric measurements and blood pressure: Height and weight were measured using a stadiometer with a calibrated scale accurate to 0.1 kg. Ensuring participants had been in a seated position be forehand, blood pressure was measured using the M6 Comfort digital monitor on three occasions at 2-minute intervals. A mean value was calculated from the last two measurements. Blood samples were drawn in the fasting state and then 30, 60, 120 and 180 minutes after the standard meal. After that, the samples were centrifuged and aliquots of plasma/serum were stored at -80°C for analysis. Plasma glucose was analysed using the Beckman Analyzer glucose-oxidase method (Beckman Instruments, Inc., Fullerton, CA, USA) and glycated haemoglobin using the VARIANT II Hemoglobin Testing System (Bio-Rad Laboratories GmbH, Munich, Germany). Plasma lipids were measured using enzymatic methods (Roche, Basel, Switzerland).
Inflammatory markers and appetite hormones: Concentrations of TNFa, MCP-1, leptin and ghrelin were determined by multiplex immunoanalysis based on xMAP technology using the MILLIPLEX MAP Human Metabolic Hormone Magnetic Bead Panel (HMHEMAG-34K) (Millipore, Billerica, MA, USA) and the Luminex 100 IS instrument (Luminex Corporation, Austin, USA).
Oxidative stress markers: The whole blood level of reduced and oxidised forms of glutathione were determined using the Glutathione in Whole Blood – HPLC diagnostic kit (Chromsystems, Munich, Germany). The activity of glutathione peroxidase (GPx) was analysed using a glutathione peroxidase assay kit (Cayman Chemical, MI, USA). The serum level of ascorbic acid was measured using a spectrophotometric method as previously described [6].
Dicarbonyl stress markers: The concentration of methylglyoxal was determined after derivatisation with 1,2-diaminobenzene using HPLC and fluorescence detection according to Fleming and Bierhaus as previously described [15]. A registered dietitian analyzed both meals, using a countr-speficif food database and software [16].
Statistical analysis
Sample size was estimated based on a power analysis with an alpha of 0.05 and a power of 0.80 to detect between intervention differences in thalamus perfusion (primary outcome) and serum concentrations of GLP-1 and GSSH (secondary outcomes), using the PASS 16.0 Power analysis and sample size software, 2018 (NCSS, Kaysville, UT, USA). Based on our preliminary data, after data transformation to achieve normal distribution, to have 80% power to detect a difference between the two meals would require 10 subjects in each group for the primary outcome and 14 subjects in each group were required for the secondary outcomes. Assuming an attrition of 25%, the expected sample size is 18 in each group.
Intention to treat analysis was performed, using repeated-measures ANOVA. Group, subject and time factors were all included in the model as follows: inter-individual (T2D vs. obese vs. controls); intra-individual (time taken to complete the meal test); and interaction between factors (divergence degree between the time profiles of each group).
To eliminate skewed data distribution and heteroscedasticity, the original data was transformed to a Gaussian distribution to attain symmetric distribution in both predictors and dependent variables and, at the same time, to stabilize the variance (attaining homoscedasticity), by a power transformation using the statistical software Statgraphics Centurion, version XV from Statpoint Inc. (Herndon, Virginia, USA), as descibed in detail previously [17].
Spearman’s correlations were calculated based on the relationship between postprandial changes in oxidative/dicarbonyl stress markers and alterations in inflammatory markers and appetite hormones. The fasting period in addition to any changes (120-0, and 180-120 min after ingestion of a standard breakfast) were calculated as follows: for each separate period, and then for all 5 values combined. Analysis was carried out using PASS 2005 statistical software (Number Cruncher Statistical Systems, USA); the statistician was blinded to the analysis. Data are presented as the mean with a 95% confidence interval (CI).