Adherence to a Low Carbohydrate Diet may Modify Resting Metabolic Rate among Overweight and Obese Women

Objective Resting metabolic rate (RMR) accounts for most of the daily energy expenditure. The low-carb diet attenuates decreases in RMR. This study aims to investigate the relationship between a low-carb diet and resting metabolic rate status. Methods We enrolled 304 overweight and obese women in this cross-sectional study. BMI, fat mass, fat-free mass, visceral fat, insulin level were assessed. RMR was measured using indirect calorimetry. A low carbohydrate diet score was measured using a validated semi-quantitative food frequency questionnaire (FFQ). Results Our results showed no relationship between LCDS and DNR even after adjust for confounders (Inc. RMR: OR: 0.97; 95% CI: 0.92–1.01, P = 0.20; Dec. RMR: OR: 0.97; 95% CI: 0.94-1.00, P = 0.14). Some components of LCDS had signicant differences with DNR, such as carbohydrate and Dec. RMR in adjusted model (OR: 1.62; 95% CI: 0.98–1.37, P = 0.08), MUFA and Dec. RMR in adjusted model (OR: 0.48; 95% CI: 0.21– 1.10, P = 0.08) and rened grain and Inc. RMR in crude model (OR: 0.87; 95% CI: 0.77–0.99, P = 0.04). Fasting blood glucose (FBG) levels were evaluated by a colorimetric method based on the GOD-PAP method, triglyceride (TG) was assessed by GPO-PAP, and low-density lipoprotein cholesterol (LDL-c) was evaluated by the direct method. An immunoinhibition assay was used for the measurement of high-density lipoprotein cholesterol (HDL-c) and total cholesterol levels. Pars Azmoon kit was used for all assessments Azmoon Inc. Iran) other than insulin. Serum insulin concentrations were analyzed by enzyme-linked immunosorbent assay (ELISA) insulin ELISA All measurements were taken


Background
Obesity, which is a serious, current health problem, affects 400 million adults worldwide [1]. Obesity is also a major public health problem in Iran, where 21.7% percent of the adult populations are obese [2].
Obesity is characterized as a chronic multifactorial disorder with a genetic basis, which is caused by surplus fat tissue accumulation. It leads to many severe comorbidities, such as insulin resistance, hypertension, development of dyslipidemia, and diabetes mellitus [1,[3][4][5]; in many countries, nutritional tendencies towards high fat and high energy foods, as well as low physical activity are among the main factors contributing the increase in the incidence of obesity [2]. The traditional treatment for obesity includes a combination of low-calorie diet therapy with enhanced physical activity and nutritional education [1]. Part of the solution might be to prescribe the most suitable diet for each subject, based on eating habits, desires, and patterns [6]. Although low-fat diets and energy-restricted diets are commonly advised diets for weight loss, low-carbohydrate diets (low-carb diet) are also a popular choice [7].
Low-carbohydrate diets restrain caloric intake by decreasing the consumption of carbohydrates to 20 to 60 g/day (typically less than 45% of the daily caloric intake) while enhancing protein and fat [6,8]. Lowcarbohydrate, high-fat and high-protein diets (referred to as low-carb diets) effectively improve weight loss, as well as providing notable improvements in lipid pro les and glycemic control [9,10]. The high protein content is satiating, and ketosis has an anorectic outcome, accounting for suppressed appetite.
Some studies are showing that low-carb diets result in rapid weight loss because of increased energy expenditure via ketogenesis, or simply by appetite repression because of the high protein content. Protein is more satiating than carbohydrates and fats, both in the long term and short term, and it seems to in uence thermogenesis, thereby in uencing satiety [1].
Several studies have recommended that low-carb diets (< 45% energy from carbohydrates) attenuate decreases in RMR, with proposed mechanisms including changed substrate availability and endocrinemediated in uences on anabolic and catabolic pathways. Furthermore, some studies have recommended that low-carb diets may support the preservation of fat-free mass (FFM) and preferential loss of fat mass, which would also attenuate decreases in RMR [12].
To the best of the researchers' knowledge, this is the rst study to investigate the relationship between resting metabolic rate status and low-carbohydrate-diet score (LCDS) in an adult population. Accordingly, this study was carried out to examine LCDS with the deviation of normal RMR (DNR) among a group of adult Iranian women.

Participants
This cross-sectional research included 304 adult women aged between 18 and 56, who had been referred to health centers in Tehran. Blood samples and anthropometric measurements were taken in the Nutrition and Biochemistry Laboratory of the School of Nutritional and Dietetics at Tehran University of Medical Sciences. Participants were in good general health, with a body mass index (BMI) in the range of 25-49 kg/m 2 . This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethics Commission of Tehran University of Medical Sciences (IR.TUMS.VCR.REC.1395.1597), and all participants signed written informed consent. The exclusion criteria for the study were as follows: regular use of medicine (including oral contraceptive pill), history of hypertension, cardiovascular disease, diabetes mellitus, and impaired renal and liver function, alcohol use, smoking, pregnancy, lactation period, and menopause. Furthermore, participants were excluded from a chronic disease affecting their diet, as well as those who had been following an arbitrary special dietary regimen, and also those with any signi cant body weight uctuations over the past 1 year.

Energy Expenditure Measurements
Resting metabolic rate was measured by indirect calorimetry (spirometer METALYZERR 3B-R3, Cortex Biophysik GmbH, Leipzig, Germany). According to the manufacturer's instructions, gas ventilation and exchange is calibrated before each test. Its high-resolution spiroergometric systems use an infrared sensor for CO 2 evaluation and an amperometric solid electrolyte sensor for O 2 evaluation, which is recorded continuously through breath-by-breath gas analysis. Utilizing an ergonomically designed mask, a small portion of breathed air is conducted through the volume ow sensor. The RMR is evaluated by measuring the amount of O 2 consumed and CO 2 produced. The RMR was assessed in the morning, after a comfortable night's sleep, and following a 10-12 hour fast. Participants were asked to avoid caffeine or alcohol consumption and severe exercise for a day before RMR measurements. After reclining in a steady-state and a supine position in a quiet room, the RMR was measured for 30 min. The respiratory exchange ratio and oxygen uptake (VO 2 ) were analyzed within the middle 20 min of the resting period.
Predictive RMR was determined using the Harris-Benedict equation, which considers the weight, height, and age of participants.

Body Composition Measurement
Body composition, including weight, BMI, fat mass, and FFM were acquired using a multi-frequency bioelectrical impedance analyzer InBody 770 scanner (Inbody Co., Seoul, Korea). This electrical impedance analyzer calculates the resistance of body tissues to the ow of an electrical signal sent through both hands and feet. The amount and proportion of bodily fat-free mass and fat mass can be measured as the current ows more e ciently through certain parts of the body. According to the manufacturer's instructions, after shoes, coats, and sweaters had been removed, subjects were required to stand on the balance scale in bare feet and grasp the handles of the machine. The measurements take approximately 20 seconds, and the output is printed.

Biochemical Parameters and Hormonal Assay
Venous blood samples were collected in the morning (8-10 A.M.) after 10-12 hours fasting by a quali ed phlebotomist. Within 30 to 45 min after each sample was collected, the blood was centrifuged for 15 min. Following separation, the serum was removed and frozen at -80ºC for later analysis. Fasting blood glucose (FBG) levels were evaluated by a colorimetric method based on the GOD-PAP method, triglyceride (TG) was assessed by GPO-PAP, and low-density lipoprotein cholesterol (LDL-c) was evaluated by the direct method. An immunoinhibition assay was used for the measurement of highdensity lipoprotein cholesterol (HDL-c) and total cholesterol levels. Pars Azmoon kit was used for all assessments (Pars Azmoon Inc. Tehran, Iran) other than insulin. Serum insulin concentrations were analyzed by enzyme-linked immunosorbent assay (ELISA) method (Human insulin ELISA kit, Monobind Inc., Lake Forest, USA). All measurements were taken at the Nutrition and Biochemistry Laboratory of the School of Nutritional Sciences and Dietetics.

Deviation of Normal RMR and Calculation Method
After examining the values of the body composition analysis, RMR components, and biochemical characters and comparing them with RMR status, the participants were categorized into 3 groups: increased RMR (Inc. RMR), normal RMR, and decreased RMR (Dec. RMR), based on the score of the deviation from normal RMR. Deviation of normal RMR was measured by indirect calorimetry (METALYZERR 3B-R3). The cutoff points for the groups were as follows: Inc. RMR (> 5% SD of normal RMR), normal RMR (-5% SD < normal RMR < 5% SD), and Dec. RMR (normal RMR < -5% SD).

Calculation of Low-Carbohydrate-Diet Score
A validated and reliable 168-item food frequency questionnaire (FFQ) was used to assess the dietary intake of participants. This semi-quantitative questionnaire consists of standard portion sizes for each food item and has been designed according to the Willett method. Participants were asked to determine the frequency of consumption of each food item during the previous year, based on serving sizes. The validity and reliability of the FFQ were determined previously [15]. Food intakes reported in household measures were then converted to grams of food per day using the nutritionist IV software. LCDS was measured for each subject. The participants included in the current study were divided into 11 strata based on their scores in the following seven categories: carbohydrates re ned grains, vegetable protein intake, monounsaturated fatty acids (MUFA), n3/n6 polyunsaturated fatty acids (PUFA) (expressed as a percentage of energy intake), as well as ber (gr/1000 Kcal), and glycemic load (GL). Dietary GL was estimated as (total glycemic index * total available carbohydrate)/100 [16] and expressed as gr/d. Women in the lowest stratum of re ned grains, carbohydrates and GL were given a score of 10, and those in the highest stratum were given a score of 0. For n3/n6 PUFA, MUFA, ber, and vegetable protein intake, the order of the strata were reversed. The points for the seven items were added together to create the overall score, named the "low carbohydrate diet score", which ranged from 0 (the highest carbohydrate intake and the lowest fat and protein intake) to 70 (the lowest carbohydrate intake and the highest protein and fat intake). Therefore, higher LCDS scores demonstrated closer adherence to low-carb diets [17].

Assessment of Other Variables
Physical activity levels were measured by a validated questionnaire (International Physical Activity Questionnaire-Short Form) that also included leisure, occupational commuting, and housework activities [18]. For height measurements, subjects were in a standing position without shoes, in contact with the wall with their head, shoulders, heels, and hips, and their height was recorded to the nearest 0.1 cm.

Statistical Analysis
All statistical analysis was performed using the IBM SPSS software version 22.0 (SPSS, Chicago, IL, USA). Normal distribution of data was checked by the Kolmogorov-Smirnov test. An independent sample t-test was used for assessed differences between groups with low and high adherence to low-carb diets. The differences between RMR status groups were assessed by one-way ANOVA and re-analyses by the general linear model (GLM) were performed to adjust for confounders' effects. Collinear variables did not enter into the model. Post-Hoc Multiple Comparison analysis resulted from the LCD procedure was used to demonstrate the signi cant differences between groups. Multinomial logistic regression was used to assess the association of DNR and LCDS and its components. Four models were constructed: Model 1 was adjusted for age, model 2 was adjusted for FFM, model 3 was adjusted for physical activity, and model 4 additionally adjusted for energy intake. Results were presented as odds ratios (ORs) and 95% con dence intervals (CIs) compared with the DNR groups.

Study Population Characteristics
Three hundred and four healthy obese women enrolled in this cross-sectional study. The mean age, height, weight, and BMI of the study participants were 36.49 years (SD=8.38), 161.38 cm (SD=5.90), 80.89 kg (SD=12.45), and 31.04 kg/m2 (SD=4.31), respectively ( Table 1). The mean body composition, RMR components, biochemical and anthropometric characteristics of subjects are shown in Table 1.

Study Participant Characteristics between High and Low Adherence of Low-Carb Diet
All participants were categorized based on the LCDS and divided into two groups ( Table 2). The differences between the low and high adherence of the low-carb diet groups were analyzed through independent sample t-test for RMR components, body composition analysis, and biochemical characteristics. As shown in Table 2, participants with high adherence to a low-carb diet had signi cantly higher LDL-c (P = 0.03). However, subjects with high adherence to a low-carb diet had higher total cholesterol (P = 0.22) and HDL-c (P = 0.10), compared to the low adherence low-carb diet group, but these ndings were not statistically signi cant. There were no signi cant differences in terms of height, weight, RMR, RMR/kg body weight, normal deviation, Respiratory Quotient, FFM, FBS, and TG between the two groups (P > 0.05) ( Table 2).

The Association of LCDS and RMR across Deviation of Normal RMR
Multivariate-adjusted models with 95% con dence intervals for the association between LCDS and RMR across DNR are presented in

Discussion
In an effort to address the issue of obesity and increased adipose tissue, the present study aimed to examine the mediatory role of a low-carb diet and its components on body weight. Several elds of study have suggested that overweight and obesity have both rapidly increased worldwide in recent decades [19]. One of the ways to control body weight is to increase RMR. However, a low-carb diet might be able to reduce the development of obesity. Previous studies had indicated links between low-carb diets and obesity [20]. Therefore, this paper sought to test the effect of LCDS on the possible link between obesity and deviation of normal RMR in overweight and obese women.
The ndings of the current study indicate that high adherence to a low-carb diet is associated with higher LDL-c. This result may be attributed to the replacement of carbohydrates with fats in a low-carb diet [21]. This nding was consistent with previous observations that fat intake results in an increase in LDL-c [22]. Also, this result showed the association between a low-carb diet and increases in total cholesterol and HDL-c. Based on previous studies, lower dietary intake of carbohydrates has been associated with higher concentrations of HDL-c [21,23].
The other nding of this research is that increased RMR is strongly associated with higher height, RMR/kg body weight, skeletal muscle mass, soft lean mass, and ISQUICKI. These ndings were in line with a previous nding which showed that, in women with up to 40% body fat, fat mass was associated with increased metabolic rate [24]. Low RMR is associated with increased fat mass and weight [25]. The excess fat mass has a signi cant in uence on metabolic function [26]. However, in obese and overweight individuals, the fat mass has a greater metabolic impact [27], both directly, by altering substrate oxidation and metabolic rate, and indirectly, by chronic changes in hormonal concentrations [26], with skeletal muscle being the most easily manipulated contributor to RMR. Lean mass, which includes both organ tissue and skeletal muscle, accounts for 60-70% and 20-30% of RMR, respectively [28]. Muscle mass speci cally is the main location for substrate oxidation and is correlated with enhanced health status, including improved insulin and glucose adjustment, but the correlation between body composition (speci cally metabolic function) and lean mass is still unclear [29].
This study found no signi cant association between LCDS and RMR status. This result is in line with previous studies which showed that low-carb diets failed to increase energy expenditure compared to lowfat diets [20]. The current study also investigated the relationship between the components of a low-carb diet and DNR. After adjustment for age, physical activity, FFM, and energy intake, a signi cant relationship was observed between carbohydrates, MUFA, and re ned grain on the one hand, and DNR on the other. Moreover, previous studies have suggested that dietary carbohydrates are among the factors thought to in uence metabolic adaptation [12]. However, the current ndings are in agreement with those ndings that proposed that reducing dietary carbohydrates may decrease reductions in RMR through mechanisms associated with substrate availability, and autonomic and hormonal activity [30].
Contrary to the results of this study, Gillingham et al. reported that there was no signi cant correlation between the consumption of MUFA and modulate resting or postprandial energy expenditure [31]. However, ndings from other studies have reported that dietary increases in MUFA [32,33], and the ratio of MUFA to saturated fatty acid or polyunsaturated fatty acid, increased the thermic effect of food, and/or fat oxidation [34,35]. More speci cally, MUFA is more powerful than saturated fatty acids in upregulating PPARα expression, inducing the transcription of genes associated with thermogenesis and fat oxidation, while suppressing the genes regulating fatty acid synthesis [31]. However, in the present study, there were no associations seen between vegetable protein intake, ber, n3/n6 ratio, and DNR.
Less is known about the potential in uence of basal blood hormones like insulin on RMR [36]. Astrup and colleagues [37] have reported a moderate correlation between insulin and RMR in females [36]. Moreover, previous studies revealed signi cant differences in RMR in individuals with insulin resistance [38,39]. More indirect support for the association between RMR and insulin comes from studies describing higher RMR in individuals with type 2 diabetes, contrasted with non-diabetics, which has been suggested, is due to insulin resistance [36]. Re ned grains had diminished insulin sensitivity, and one of the rst responses to alternations in insulin sensitivity is the change in hepatic insulin clearance rates [40]. The physiological mechanisms responsible for elevated RMR in individuals with insulin resistance are poorly understood. Several mechanisms have been suggested to explain the increased RMR, including futile substrate cycling, plasma glucagon, increases in protein turnover, and sympathetic nervous system activity [41]. The other proposed mechanism that revealed the correlation between RMR and insulin resistance was an increase in gluconeogenesis. It has been put forward that enhanced free fatty acid concentrations in individuals with insulin resistance contribute to increased hepatic glucose output and excessive rates of gluconeogenesis, depending upon the fatty acids oxidation and consequently the increased energy expenditure rate in these samples [41,42]. In support of this pathway, following improvements in glycemic control, a decrease in resting energy expenditure was reported.

Conclusion
In conclusion, this study's nuanced ndings do highlight that a low-carb diet has no signi cant correlation with DNR, but some of the components of this diet, like re ned grains, MUFA, and carbohydrates, revealed signi cant associations. This could lead to practical strategies to assist in the control or prevention of overweigh and obesity and related disorders in community.
To the researchers' knowledge, this study was the rst to assess the possible relationship between lowcarb diets and DNR in obese women. Studies of the possible link between this low-carb diet and DNR in obese people require more clinical trials, as well as further cohort research designs. The major limitation of our study was the relatively small number of participants and the same-sex sample. Also, due to the study type, causality is not able to be determined.

N=304
Data are indicated as Mean ± SD otherwise indicated