The Association of Dietary Acid Load with Resting Metabolic Rate and Metabolic Components in Overweight and Obese Women: A Cross Sectional Study

Background: Several epidemiologic studies have reported that dietary acid load is associated with metabolic proles; however, to our knowledge, the relationship of this dietary pattern with resting metabolic rate (RMR) among obese and overweight females remains unreported. Therefore, this study aimed to evaluate the association of dietary acid load RMR and metabolic components among overweight and obese adult women. Methods: This cross-sectional study was conducted on 375 Iranian adults, aged 18–48 years. Dietary acid load indexes were calculated by using a validated 147-item semi-quantitative FFQ. Biochemical and anthropometric measures were assessed using standard methods. An impedance fat analyzer was used to obtain the body composition and an indirect calorimeter was used to assess the RMR. Result: It was observed that after correction for potential confounders, DBP and NEAP and PRAL scores were inversely associated (P<0.05). NEAP index was inversely associated with RMR (β= -0.25, 95% CI=-0.1.5 to 2.08, P=0.02), and positively associated with WC (β= 1.009, 95% CI=-1.43 to 3.45, P=0.05) and WHR (β= 0.01, 95% CI= -0.01 to0.04, P=0.03), such that subjects with higher scores in NEAP had lower RMR and higher WC and WHR. We also observed that DAL (β= -0.02, 95% CI= -0.08 to0.03, P=0.08) and PRAL (β= -0.037, 95% CI= -1.05 to 0.03 P=0.07) were marginally associated with RMR. Conclusion: The results of the present study suggested that higher dietary acid load scores may be negatively associated with lower RMR, while directly associated with greater WC, WHR, DBP, and HOMA-IR. test was used to assess the distribution of the data. Continuous variables were represented by mean ± Standard Deviations (SDs) and mean ± standard Errors (SEs), and categorical information was represented by percentage and number. To avoid classication errors, we calculated the energy-adjusted PRAL, NEAP, and DAL using the residual method, and then based on the median, dichotomized participants into low or high dietary acid load. Baseline characteristics of participants were compared by independent samples t test between the median of PRAL, NEAP, and DAL score and chi square(χ 2 ) tests for categorical variables. Also, we adjusted variables for confounders, such as age, energy intake, BMI, physical activity, and economic status, in addition further adjustment for free fat mass (FFM), using Analysis of covariance (ANCOVA). General linear regression (GLM) was used to assess the association of PRAL, NEAP, and DAL score with body composition, biochemical variables and RMR per kg. Results in crude and adjusted models were presented as beta(β) and 95% con dence intervals (CIs). Data were analyzed using IBM SPSS version 25.0(SPSS, Chicago, IL, USA) and p-values less than 0.05 were considered, a priori, to represent statistical signicance.


Introduction
The prevalence of overweight and obesity are increasing globally, and concerningly, it has been demonstrated that obesity is strongly related to several lifelimiting disorders, including cardiovascular diseases, diabetes, various types of cancer, and stroke (1). Obesity is regarded as a multi-factorial disorder (2), which may be in uenced by eating habits, low energy expenditure, and physical (in)activity (3).
The role of acid-base homeostasis in cardio-metabolic risk factors has recently been posited (4,5). Generally, animal protein such as sh and meat, as well as grain products, are in uential in a higher acid production in the body, while, conversely, vegetables, legumes, and fruits are known as alkaline foods (9,10).
It has been suggested that lower acid production is bene cial in human health (11). In recent years, dietary patterns have evolved towards a western-style diet, which is de ned by a higher intake of animal products, fast foods, and other food types that result in higher production of acid in body (5,12), and is not, generally, compensated for by fruits and vegetables, which can contribute to metabolic dysfunction.

Dietary assessment
Dietary intake was assessed using a validated semi quantitative food frequency questionnaire (FFQ) with 147-food items (37). All FFQs were completed by a trained dietitian, who asked participants to designate their intake frequency for each food item consumed during the past year, on a daily, weekly, or monthly basis. Finally, we converted portion sizes of foods to grams by using household measures (38). NUTRITIONIST-IV (version 7.0; N Squared Computing, Salem, OR, USA) was used to assess nutrient and energy intakes.
Medians of diet dependent acid load scores were used in statistical analysis.

Anthropometric assessments
Weight was determined on a digital scale, where participants wore light indoor clothing, without shoes, and was recorded to the nearest 100 g. Height was measured to the nearest 0.5 cm while participants were in the normal standing position, without shoes. Waist circumference (WC) an hip circumference were measured to the nearest 0.5cm, according to standard procedures. Subsequently, waist-to-hip ratio (WHR), waist-to-height ratio (WHtR), and BMI were calculated according to standard formulae. Overweight and obesity were de ned as 25≤BMI≤29.9 kg/m 2 and BMI≥30 kg/m 2 , respectively. Neck circumference (NC) was measured below the laryngeal prominence and perpendicular to the long axis of the neck, and the minimal circumference was recorded to the nearest 0.1 cm (41) Assessment of other variables A demographic questionnaire to discern information on age, marital status, education, and economic status was collected by researchers.
Blood pressure assessment Blood pressure measurements were taken using a standard mercury sphygmomanometer; where the participant sat for 10-15 min before two consecutive measurements were taken.

Physical activity
The Short form of the International Physical Activity Questionnaire (IPAQ) (42) was applied to evaluate physical activity and according to the frequency and time of common activities of daily life over the past year. Physical activity levels were expressed as metabolic equivalent hours per week (METs h/week) (43).

Measurement of biochemical parameters
Blood samples were collected after 10-12 hours fasting, and collected into tubes containing 0.1% Ethylenediaminetetraacetic acid (EDTA).Then, they were centrifuged for 10 min at 3000 rpm, aliquoted into 1 ml tubes, and stored at -70 o C until analysis. Samples were analyzed by using an autoanalyzer (Selectra 2; Vital Scienti c, Spankeren, Netherlands).
The GOD/PAP (glucose oxidase phenol 4-Aminoantipyrine Peroxidase) method was used for the measurements of fasting blood sugar (FBS) and triglyceride (TG) levels, and cholesterol levels were evaluated using the cholesterol oxidase Phenol 4-Aminoantipyrine Peroxidase (CHOD-PAP) method. Total cholesterol (TC) levels and direct high-density lipoprotein (HDL) were evaluated by the Immunoinhibition assay. Aspartate aminotransferase (AST) and Serum alanine aminotransferase (ALT) were speci ed by the International Federation of Clinical Chemistry and Laboratory Medicine method. Serum high-sensitivity Creactive protein (hs-CRP) was determined using a high-sensitivity immunoturbidimetric assay (Hitachi 902 analyser; Hitachi Ltd., Tokyo, Japan).
Resting metabolic rate assessment RMR measurement was performed by professional nutritionists using a standard protocol (33), and was measured by indirect calorimetry (spirometer METALYZERR 3B-R3, Cortex Biophysik GmbH, Leipzig, Germany). Based on the manufacturer's instructions, prior to each test, gas exchange and ventilation were calibrated. Fitmate is a desktop device designed for providing accurate RMR for all elds dealing with obesity and malnutrition with a mask to cover the nose and mouth. The device is designed to measure oxygen consumption and energy expenditure during rest and exercise. It uses a turbine ow meter to measure ventilation and a galvanic fuel cell oxygen sensor for analyzing the proportion of oxygen in expiration gases. Moreover, it uses a patent-pending sampling method that allows the analyzer to monitor the performance of a metabolic cart in a standard mixing chamber. RMR was calculated using information such as oxygen consumption, a xed respiratory quotient of 0.85, and estimated quantity of urinary nitrogen using a modi ed Weir equation below.
The RMR was evaluated in the morning, after a night's sleep and following 10 to 12 hours of overnight fasting. Participants were asked to avoid vigorous exercise and alcohol or caffeine consumption for a day before RMR assessment. The RMR was measured for 30 min in a quiet room and participants in xed, comfortable position. The respiratory exchange ratio and oxygen uptake (VO 2 ) were investigated within the average 30 min of the resting period. However, the rst 5 minutes were not included, and only the last 15 minutes were used to calculate RMR. Predictive RMR was determined using the Harris-Benedict equation, which considers age, weight, and height for individuals.

Complete body composition analysis
Body composition of subjects was evaluated via Body Composition Analyzer BC-418MA-In Body (United Kingdom), in accordance with manufacturer guidelines. To avoid possible discrepancies in the measured values, participants were asked not to exercise vigorously, to not use any electrical devices, and not consume excessive uid or food before evaluating their body composition.

Statistical analysis
The Kolmogorov-Smirnov test was used to assess the distribution of the data. Continuous variables were represented by mean ± Standard Deviations (SDs) and mean ± standard Errors (SEs), and categorical information was represented by percentage and number. To avoid classi cation errors, we calculated the energy-adjusted PRAL, NEAP, and DAL using the residual method, and then based on the median, dichotomized participants into low or high dietary acid load.
Baseline characteristics of participants were compared by independent samples t test between the median of PRAL, NEAP, and DAL score and chi square(χ 2 ) tests for categorical variables. Also, we adjusted variables for confounders, such as age, energy intake, BMI, physical activity, and economic status, in addition further adjustment for free fat mass (FFM), using Analysis of covariance (ANCOVA). General linear regression (GLM) was used to assess the association of PRAL, NEAP, and DAL score with body composition, biochemical variables and RMR per kg. Results in crude and adjusted models were presented as beta(β) and 95% con dence intervals (CIs). Data were analyzed using IBM SPSS version 25.0(SPSS, Chicago, IL, USA) and p-values less than 0.05 were considered, a priori, to represent statistical signi cance.

Study Population Characteristics
The mean of quantitative variables including; age, BMI, RMR, RMR normal, RMR per kg, and RMR per BSA of the participants in our study were; 36.67 years Association between quantitative and qualitative variables among PRAL, NEAP, and DAL medians The differences between the medians of PRAL, NEAP, and DAL and some quantitative and qualitative variables are shown in Table 1. After adjustment, physical activity was higher in high intake of PRAL, NEAP, DAL and also about age in high intake of NEAP (P < 0.05). FFM was higher in high intake of PRAL and DAL (P < 0.05) and normal RMR was higher in low intake of NEAP (P = 0.04).  Table 2. After adjustment for confounders, cereals consumption was higher in high intake of all three scores (PRAL, NEAP, DAL) and fruits consumption was higher in low intake of all three scores (P < 0.001); vegetables consumption was higher in low PRAL and DAL scores, and fat intake was higher in high PRAL, NEAP, and DAL scores. However, nuts and carbohydrate consumption was higher in low PRAL, NEAP, and DAL scores (P < 0.05); whilst, participants with high PRAL and low NEAP scores consumed higher amounts of total ber and calcium, respectively (P = 0.05). Magnesium, folate, potassium, and vitamins C, A, B6 were consumed in higher amounts in low PRAL, NEAP, and DAL scores (P < 0.05), whilst sodium and B12 consumption was higher in participants with high NEAP and DAL scores (P ≤ 0.05). Investigation of body composition, biochemical variables, and RMR among PRAL, NEAP, and DAL medians We made three models for investigation of body composition, blood parameters and type of RMR; crude model, model 1, and model 2 ( Table 3). In the crude model, WC and WHR were higher in of the high DAL group (P ≤ 0.05), and WHtR was incomparable across all three scores (P = 0.05). In model 1, we considered age, BMI, physical activity, and total energy intake as confounders, and in model 2, other variables, in addition to the aforementioned, were considered as confounders. In model 2: WC was higher in high DAL group (P ≤ 0.05); NC and WHR were higher in high DAL and NEAP scores, respectively (P < 0.05); WHtR was higher in high NEAP scores, whereas FBS and HOMA-IR were inversely higher with low NEAP (P < 0.05); TG was higher in high NEAP, and HDL was higher in low PRAL (P < 0.05); hs-CRP was higher in high PRAL and NEAP (P ≤ 0.05) and DBP was higher in low PRAL and DAL (P < 0.05). To investigate RMR types and RQ, we made two models; a crude model and model 1, which was adjusted for age, BMI, physical activity, and total energy intake (Table 3). In the crude model, RMR per kg was higher in low PRAL, NEAP, and DAL scores (P < 0.05) and RMR of carbohydrate was higher in low NEAP (P = 0.03). In model 1: RMR per kg was higher in low PRAL and NEAP (P < 0.05), but RMR of carbohydrate was higher in low DAL (P = 0.05); whilst RQ lost was no longer signi cant after adjustment in the Model 1. Relationship of body composition, blood parameters, RMR per kg, and high-low acid load scores of PRAL, NEAP, and DAL We made three models to investigate body composition, blood parameters, and RMR per kg: a crude model, model 1, and model 2 ( Table 4). In the crude model, we found a negative association between NEAP and FBS (P = 0.03). In model 1, we considered age, BMI, physical activity, total energy intake, and economic status as confounders, and in model 2, other variables, in addition to the aforementioned, were considered as confounders. In model 2 we observed a positive association for PRAL with WHtR (P = 0.09) and for NEAP with WC (P = 0.05), WHR (P = 0.03), TG (P = 0.01), and HOMA-IR (P = 0.08). Hs-CRP was positively associated with PRAL, NEAP, and DAL scores (P ≤ 0.08) and insulin with PRAL and DAL (P ≤ 0.08). In this model, some variables became negatively related with all three scores including: PRAL with HOMA-IR (P = 0.04) and HDL (P = 0.09), DAL with FBS (P = 0.07), and HOMA-IR (P = 0.007), and all three scores with DBP (P < 0.05). To investigate RMR per kg, we made four models; crude model, and model 1 (as above), model 3, that adjusted to age, BMI, physical activity, total energy intake, and fat free mass, and a nal model that adjusted for other confounders, in addition to those mentioned in model 3 (Table 4). In the crude model, a negative association between NEAP and RMR per kg was found (β= -0.66, 95% CI=-1.37 to -0.04, P = 0.05), and in the nal model this association became stronger (β= -0.25, 95%CI= -0.1.5 to 2.08, P = 0.02), and a marginal negative association for PRAL (β= -0.03, 95% CI= -1.05 to 0.03, P = 0.07) and DAL (β= -0.02, 95% CI=-0.08 to 0.03, P = 0.08) was found.

Discussion
The current study investigated the association of a diet with high acid load (characterized by a high DAL, a high PRAL or a high NEAP score) with anthropometric parameters, biochemical factors, and RMR among obese and overweight women. The principal ndings of this study were that, after consideration of potential confounders, we observed a direct association between blood pressure (DBP) for two scores, and a direct relationship between hs-CRP and DAL, and TG and NEAP. In addition, a favorable relationship between HOMA-IR and PRAL, as well as DAL, was observed. Findings pertaining to HOMA-IR and DBP seem to be more reliable than other biochemical markers that were signi cantly related to only one of the dietary acid load indicators.
Interestingly, the individuals with higher NEAP had lower RMR, while WC and WHR were higher in those subjects. We also observed that DAL and PRAL were marginally associated with RMR.
The load of acid production in body will be increased after higher intake of foods which are rich in animal protein, including meet, eggs, cheese, and grain. As highlighted in the present study, it was indicated that individuals with higher dietary acid load had higher intake of eggs and meat, as well as sodium, cereals, fat, and energy intake. These foods contain sulfur, amino acid, and phosphoric acid, and are strongly related with a higher production of hydrogen sulfate, which can increase the acid load in the body.
Diet acid load scores have been predicted and validated in people with kidney diseases (44), however, to date, studies have reported con icting ndings regarding the relationship of dietary acid load and cardio-metabolic risk factors (13,19,(44)(45)(46). Indeed, several studies suggested that higher acid load has a negative association (19,44,45), while others refuted this association (13,44). For instance, in an observational study on a cohort of 1136 females, it was suggested that PRAL was directly and independently associated with total cholesterol and HDL, although WC and BMI were substantially correlated with Pro:k (19).
The current study showed a signi cant association of NEAP and PRAL with DBP Previous studies have reported that in metabolic acidosis related to diet, blood pressure was increased (23,25). The Nurses' Health Study II, which was conducted on 1136 women aged 18 to 22, who had normal BMI, showed that a higher NEAP is connected with a greater risk of hypertension (44). However, this relationship is not universally supported; for instance, it has been reported that no association between blood pressure and dietary acid load was evident among participants that were higher than 55 years old (13). Confounding the equivocal reports, the mechanism underlying the proposed relationship have not been well de ned, although several pathways have been suggested, such as reduction in elimination of citrate, higher production of cortisol, and higher excretion of calcium (20,26,28).
The association of glycemic pro le and dietary acid load has been investigated in several studies. In the present study, we showed that subjects with higher dietary acid score had greater HOMA-IR. It was reported that all of metabolic acidosis markers, such as low serum bicarbonate, higher anion gap, and low urine pH are associated with insulin resistance (19,47); indeed, in the Atherosclerosis Risk in Communities (ARIC) cohort, it was also con rmed that the risk of type 2 diabetes incidence was increased with higher levels of serum lactate (47). Thus, concomitant to the Nurses' health study, evidence suggests that higher plasma bicarbonate may be linked to an decreased odds of type 2 diabetes incidence (48). , in their prospective research, suggested that subjects with a lower urinary pH had greater prevalence of glucose intolerance and type 2 diabetes in comparison with normal counterparts (17). On the other hand, no signi cant association was found for Hemoglobin A1C (HbA1c) and serum glucose levels in young Japanese women (44). Several studies have reported that an acidogenic diet may contribute to insulin resistance by constantly inducing a chronic acidosis condition in body (49,50). It is also proposed that insulin tendency to bind to its receptors is lessened in metabolic acidosis, which results in insulin resistance, albeit in animal studies (51). Moreover, insulin discharge is reduced whenever the body is faced with acid/base variations (52). Another possible mechanism may be related to the association of greater dietary acid load with higher discharge of magnesium, which decreases insulin sensitivity (53).
The associations of metabolic or urinary acidosis with biochemical parameters have been evaluated in several studies, yet, in spite of controversial results among such studies, we showed a positive association with dietary acid load and TG, WC, and WHR, and a negative relationship with RMR among the obese and overweight subjects in this study. The in uence of an acidogenic diet on RMR can be the result of higher insulin resistance (50). In a cohort of 782 adults, it was reported that RMR and HOMA-IR were signi cantly related (54), thus, it can be inferred that higher acid load, by increasing the insulin resistance, may reduce RMR. This nding may be suggested as an effective impact of a diet with a lower acid load in women with higher than normal BMI. Kentaro Murakami et al. found an independent and positive association of total and HDL-cholesterol with PRAL, and also reported that BMI and waist circumference are independently associated with Pro:K (44). However, van den Berg et al (2012) did not report any meaningful correlation between urinary acidosis and TG, HDL, and HbA1c, however, lower levels of hs-CRP and cholesterol were seen in subjects with higher acidosis (5).
The association of the dietary acid load and metabolic factors has not been fully discerned, therefore, the underlying mechanisms of how an acidogenic diet in uences metabolic outcomes have not been well explained. As mentioned above, animal protein in the diet may represent a cause of increasing acid load in body. In our study, the adverse effects of higher dietary acid score on TG, DBP, HOMA-IR and RMR was seen. The equivocal outcomes reported in studies can be explained by differences in target population, methodology, and speci c dietary behaviors; the current study focused on overweight and obese women. In addition, the validation of dietary acid load measurement methods among obese and overweight subjects is important; it has been shown that higher protein intake is not the only cause of higher acid production (6). Indeed, in the present study, we also observed that higher consumption of cereals, sodium, fat, and energy intake is related to higher acid load (6). There are several limitations in this study that should be considered when interpreting the results. The present study was conducted on only one gender, so cannot be extrapolated to males. The number of participants were relatively low, moreover, based on the crosssectional design of the study, we are not able to show any causal inference.

Conclusion
To date, no study has examined the relationship between dietary acid load and components of metabolic and RMR in obese and overweight females.
Accordingly, we found that dietary acid load scores were signi cantly related to lower RMR and higher WC, WHR, DBP and HOMA-IR in overweight and obese women. Further studies are warranted to con rm the clinical consequences of dietary acid load in obesity.

Consent for publication
Each participant was completely informed about the study protocol and provided a written and informed consent form before taking part in the study.

Availability of data and materials
The data that support the ndings of this study are available from Khadijeh Mirzaei but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Khadijeh Mirzaei.