Study design and participants
The Nutritionists’ Health Study (NutriHS) was conceived in the School of Public Health of the University of São Paulo, Brazil, aiming to investigate factors associated with health outcomes in nutritionists and undergraduates of the nutrition course [16]. This cross-sectional part of NutriHS was conducted at the University of Campinas, São Paulo, Brazil. Recruitment occurred between 2018 and 2019 in the metropolitan area of Campinas. Only women were included in this NutriHS-UNICAMP arm, since the vast majority of Brazilian nutritionists are female. The eligibility criteria were women aged between 19 and <45 and body mass index (BMI) between 18.5 and 40.0 kg/m². Women who were pregnant or lactating, using medication affecting glucose metabolism and/or body adiposity, who had used probiotics or antibiotics in the last three months and who had diabetes mellitus, heart, kidney, and liver diseases or other severe systemic diseases were excluded.
Ethical Approval
This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of each institution – University of Campinas (UNICAMP) and University of São Paulo (USP) (ethical approval number 79775817.4.1001.5404). All experiments were performed in accordance with relevant guidelines and regulations. All participants signed an electronic informed consent form, available at e-NutriHS, a specific web-based system developed for the study (www.fsp.usp.br/nutrihs) [17]. In total, 248 women answered the online questionnaires, 127 met the inclusion criteria and attended a clinical visit at our research laboratory, and 111 concluded the whole protocol.
Clinical Measurements
Body weight, height, and waist circumference - measured at the midpoint between the last rib and the iliac crest - were obtained. Body composition parameters, including total fat and visceral fat mass (in grams), were assessed by GE-Healthcare Lunar iDEXA using the automatic whole-body scan mode. All scans were performed and analyzed by trained researchers following a standard protocol. Instrument quality control was performed daily following the manufacturer’s instructions.
Blood pressure was measured in a sitting position using a mercury sphygmomanometer following recent guidelines [18]. Blood samples were obtained after a 12-h overnight fast.
Energy and glucose homeostasis assessment
The components of energy (resting energy expenditure – REE and substrate oxidation rate) and glucose (insulin resistance, insulin, and GLP-1 secretion) homeostasis were evaluated by the standardized mixed-meal tolerance test associated with indirect calorimetry in a 180-minute protocol, depicted in Figure 1. The standardized meal test was the choice for investigational purposes because it better reproduces real-life conditions among free-living individuals.
Indirect calorimetry canopy (VMAX N Encore) was performed according to the method proposed by Fulmer et al. [19]. The volunteers were instructed not to consume alcoholic beverages and not to engage in moderate/vigorous physical activity 48 hours before the test. On the previous night, the volunteers consumed a standardized meal containing 600 calories, 60% carbohydrate, 22% fat, 18% protein, and approximately 8 g of total fiber, since the standardization of the last meal guarantees accuracy in the evaluation of energy expenditure and the oxidation of energy substrates, avoiding any interference in the residual thermogenesis induced by diet [20, 21]. After this meal, the participants remained fasting for 12 hours until the first indirect calorimetry was performed. On the day of the examination, the calorimetry monitor remained on for 30 minutes before the start of the test for stabilization. The equipment was calibrated with a known gas concentration. The room was kept between 22 and 25 °C, with natural lighting and as little noise as possible. The volunteers arrived at the laboratory at 7:00 am and fasted for 12 hours to minimize any residual thermic effect of the diet. In addition, they were instructed to collect all urine produced in these 12 hours in a specific collector. After the initial rest, an evaluation of energy metabolism was performed for 30 minutes to determine the fasting REE and the oxidation rate of energy substrates. Subsequently, a venous catheter was inserted into an antecubital vein for blood collection. Next, the volunteers received a standardized liquid meal containing 524 kcal (Nutren 1.5®) in 340 mL, composed of 75 g of carbohydrates, 19 g of protein, 16.3 g of lipids, and 0 g of fiber, which was consumed in up to 5 minutes [22]. Blood samples were collected in dry tubes at -15, 0 (before food intake), 30, 60, 120, and 180 minutes (after food intake). For GLP-1 analysis, EDTA tubes containing 20 μL of dipeptidyl peptidase-IV inhibitor were used. For SCFA, lithium heparin tubes were used. Glucose and insulin were quantified at all times. GLP-1 was quantified at 0, 30, 60, 120, and 180 minutes. SCFA concentrations were quantified at 0, 60, 120, and 180 min. Indirect calorimetry was performed at the end of the 20 minutes preceding each blood collection. At the end of the test, the participants were instructed to empty their bladders in a specific collector. The total volume of urine was computed and aliquoted, and urinary nitrogen was analyzed as a marker of protein oxidation.
Biochemical analysis
Biochemical tests were performed to characterize the fasting lipid profile (LDL-cholesterol, HDL-c, triglycerides, and total cholesterol) using colorimetric and enzymatic methods. LDL-c was calculated by the Friedwald equation [23]. Glycated hemoglobin was measured with high-performance liquid chromatography (HPLC). Plasma glucose levels were promptly measured in the fasting state and during the dynamic test using a glucose analyzer (YSI 2700; YSI Life Sciences, Yellow Spring, OH, USA) with a coefficient of variation (CV) of 2%. Plasma insulin levels were analyzed using an automated two-site chemiluminescent immunometric assay (Immulite 1000 System; Siemens Health Diagnostics, USA). The intra-assay and interassay CVs were 5.2–6.4% and 5.9–8.0%, respectively. Total COOH-terminal amidated GLP-1 was measured using ELISA (cat. EGLP-35K Sigma Aldrich) with intra-assay and interassay CVs of 7% to 9% and <1% to 13%, respectively. Urinary nitrogen excretion was calculated by the UV enzymatic method by urinary urea dosage. Fecal and plasma concentrations of SCFA (acetic [C2:0], propionic [C3:0], and butyric [C4:0]) were quantified using the gas chromatography technique coupled to mass spectrometry. For stool samples, the protocol previously described by Fellows et al. [24] was applied, and for plasma samples, we used the protocol described by Wang et al. (25). The butyrate concentration was not detectable in plasma samples, and the concentrations of detectable SCFAs were summed to obtain the total concentration of SCFAs.
SCFA analysis
Fragments of feces or blood samples were harvested from volunteers under fasting or after meal conditions and used for measurement of SCFA, following a protocol similar to that used by Fellows et al. [24]. The day before the investigation, the subjects collected a fecal sample from their first sterile stool receptacles provided by the research team. The samples were directly stored on dry ice at home. Upon arrival in the laboratory, fecal samples were aliquoted and frozen at −80 °C until further analysis. After this, fecal samples were thawed, weighed, crushed, and homogenized in water. Sodium chloride, citric acid, hydrochloric acid, and butanol were added, and the mix was placed in a vortex and centrifuged at 4 °C. The supernatants were recovered. For blood samples, after 15 minutes of centrifugation of total blood to separate the serum, we followed the method proposed by Wang et al. [25]. Briefly, ethanol (P. A), n-hexane (P. A), and an internal standard (caprylic acid) was added to 200 mL of serum. All samples were vortexed and centrifuged at 13,000 rpm at 4 °C for 8 minutes. Immediately after, samples were transferred to specific vials, and p. H was adjusted to 4.0 using chloride acid (10%).
For all samples (feces and blood), a calibration curve with 0.015–0.1 mg/mL SCFA was used in the quantification. Chromatographic analyses were performed using a gas chromatograph-mass spectrometer (model GCMS-QP2010 Ultra; Shimadzu®) and a fused-silica capillary Stabilwax column (Restec Corporation, USA) with dimensions of 30 m × 0.25 mm internal diameter and coated with a 0.25-µm thick layer of polyethylene glycol. Samples (1 µL) were injected at 250 °C using a 25:1 split ratio for feces or splitless (1:1) for blood samples. High-grade pure helium (He) was used as the carrier gas with a constant flow rate of 1.0 mL/min. Mass conditions were as follows: ionization voltage, 70 eV; ion source temperature, 200 °C; full scan mode in the 35–500 mass range with 0.2 s/scan velocity. The runtime for each analysis was 11.95 min.
Mathematical calculations
Total resting energy expenditure (REE) (kcal/day) and REE adjusted by body weight (kcal/kg body weight) were calculated by the complete equation of Weir [26], considering the last 20 minutes, known as steady-state [27]. At 0, 30, 60, 120, and 180 minutes, fat and carbohydrate oxidation rates were calculated using standard equations for oxygen consumption, carbon dioxide production, and urinary nitrogen excretion [26, 28].
The HOMA-IR homeostasis model assessment of insulin resistance index [29] and PREDIM (predicted M value) [30] were calculated as markers of resistance and sensitivity to insulin, respectively. The insulinogenic index was calculated as a marker of insulin secretion considering the formula ΔT30glycemia/ΔT30glycemia. [31]. This index reflects the insulin response to the glucose stimulus related to the first phase of insulin secretion, which includes the prompt response of the beta-cell to the food stimulus. A decreased insulinogenic index is linked to the deterioration of beta-cell function.
The calculation of the areas under the curve (AUC) for the dosages of glucose, insulin, GLP-1, SCFA, and the lipid and carbohydrate oxidation rates were performed using the trapezoidal method and represent a complete picture of the postprandial state [32].
Definition of metabolic phenotypes
BMI was classified as normal if values ranged between 18.5 and 24.9 kg/m² and as overweight if values were ≥25.0 kg/m² [33]. Metabolic health was defined based on the strict criteria proposed in the BioShare Health Obese Project [34], which consider the absence of any one of the following alterations: systolic blood pressure ≥130 mmHg, diastolic blood pressure ≥85 mmHg, or use of antihypertensive medication; fasting plasma glucose ≥110 mg/dL or nonfasting plasma glucose ≥126 mg/dL or use of antidiabetic agent; HDL-cholesterol <50 mg/dL or triglycerides ≥150 mg/dL or use of lipid-lowering medication; and diagnosis of cardiovascular disease. Based on this definition, participants were classified into four phenotypes: normal-weight metabolically healthy (MHNW) or unhealthy (MUNW), obese/overweight metabolically healthy (MHO), or unhealthy (MUO).
Statistical analysis
Analyses were performed using IBM SPSS Statistics software. The Kolmogorov–Smirnov test was applied to test data normality. Median values and interquartile ranges are provided. Minimum and maximum values were presented for sample characterization variables to present the range of variation of some parameters. Tertiles of body adiposity measures were obtained for comparative analyses. Mann–Whitney and Kruskal–Wallis tests were applied for comparisons between two or ≥three independent groups, respectively. Spearman correlation was applied to evaluate the correlation between SCFA and variables of food consumption and glycemic and energetic homeostasis. Friedman's analysis of variance of two factors associated with multiple comparisons by pairs per rank was applied to compare the metabolic variables throughout the standard mixed-meal tolerance test. A significant p value was set as <0.05.