Associations between dietary acid load and obesity among Iranian women

Introduction: Diet-induced acid load may be associated with overweight and obesity as well as with diet quality. We aimed to study how dietary acid load is associated with overweight, obesity and diet quality indices in healthy women. Methods: We randomly selected 306 healthy 20 to 55 year-old women from health centers affiliated with Tehran University of Medical Science. They were enrolled in a cross-sectional study between June2016 and March 2017. Potential renal acid load (PRAL), net endogenous acid production (NEAP) and dietary acid load (DAL) were calculated for each person. Dietary quality index international (DQI-I),mean adequacy ratio (MAR), and energy density (ED) were estimated. Anthropometry was measured using standard protocols. Nutritional data were obtained from food frequency questionnaires (FFQ). We used multivariable logistic regression models to assess dietary acid load indices in relation to overweight, obesity and abdominal adiposity. Results: Participants had a mean age of 32.4 years. The number and percentage of women who were overweight, obese and who had abdominal obesity were 94(30.7), 38(12.4) and 126(41.2), respectively.The odds of obesity (adjusted odds ratio; Adj. OR = 2.41, 95% confidence interval; CI:1.01-5.74,P = 0.045) and abdominal adiposity (Adj. OR = 2.4, 95% CI:1.34-4.60, P = 0.004) increased significantly with tertile of DAL. Other dietary acid load indices (PRAL and NEAP) showed no significant association with obesity, overweight or abdominal obesity. As dietary acid load scores (PRAL, NEAP and DAL)increased, DQI-I and MAR significantly decreased whereas ED significantly increased across tertilesof dietary acid load indices (P < 0.001). Conclusion: Dietary acid load is associated with obesity and abdominal obesity and is also considered an indicator of diet quality.


Introduction
Overweight and obesity are serious public health problems that contribute to chronic diseases worldwide, especially cardiovascular disease. [1][2][3][4] The prevalence of these disorders has been increasing at an alarming rate across the globe. 5 A recent meta-analysis documented that the prevalence of overweight and obesity were about 41% and 13% in Iranian adults, respectively. 6 In another systematic review and meta-analysis, authors estimated that the prevalence of obesity in older adults was around 21% and that the proportion of overweight and obesity in women was higher than in men. 7 Among the myriad genetics, environmental, and behavioral determinants associated with the rise of overweight and obesity, nutrition is a critical factor. 8,9 Numerous studies have documented associations between the relative intake of different food groups (whole grains, dairy, fruits, and vegetables) and obesity. [10][11][12] A person's typical diet represents a mixture of food groups, in which foods and nutrients are consumed together. For that reason, an analysis of dietary patterns, rather than an analysis of food groups or individual foods, can contribute to our understanding of synergistic interactions between foods and nutrients. 13,14 The Western diet has a strong influence on global eating patterns. 15 In recent years, the possibility of changes in physiological acid-base balance due to the Western diet have attracted attention. Numerous studies show consumption of high-protein foods and insufficient fruit and vegetable intake can increase the body's hydrogen ion load. [16][17][18] These findings suggest that efforts to measure the net acid load from dietary intake may help refine our understanding of the effects of diet on human health. 19,20 Potential renal acid load (PRAL), net endogenous acid production (NEAP) and dietary acid load (DAL) are used to estimate the acidogenic potential of foods and are used as indices to assess dietary acid load. [21][22][23][24] Acid-base imbalance leads to metabolic changes accompanied by potential complications including insulin resistance, high cortisol levels, increased abdominal obesity and decreased insulin sensitivity. 3,[25][26][27] Recent studies have shown a positive correlation between high dietary acid load and increased risk of high blood pressure, cardiovascular disease, and diabetes mellitus. 21,[28][29][30] For example, a recent study by Han et al showed that individuals in the highest PRAL tertile were at increased risk of atherosclerotic cardiovascular disease (ASCVD) risk over a ten-year period and were at higher risk compared to those in the lowest PRAL tertile. 31 Since obesity and overweight are also risk factors for heart disease, they should not be ignored.
Food quality indices such as the diet quality index -international (DQI-I), Energy Density (ED) and the Mean Adequacy Ratio (MAR) are instruments to evaluate adherence to dietary guidelines. [32][33][34] The associations between some of these indices and obesity have also been investigated. 35,36 Given the variety of dietary patterns in Iran, the aim of our study was to investigate dietary acid load in relation to overweight, obesity and dietary quality indices among Iranian women.

Materials and Methods
The present study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for conducting and disseminating observational studies. Figure 1 displays a flow diagram of the study participants.

Study design, setting and participants
We conducted a descriptive and analytical cross-sectional investigation with 20-55 year old women who were referred to health centers affiliated with Tehran University of Medical Science (TUMS) between June 2016 and March 2017. These women were selected by using systematic cluster random sampling. To this end, we divided the health centers into 15 clusters and systematically enrolled approximately 20 participants in each cluster. In total, 306 participants agreed to participate who had no prior history of cardiovascular disease (CVD), diabetes, cancer, stroke, chronic or acute renal disease, and who were not taking medications that could affect their weight. Exclusion criteria included being foreign (non-Iranian), pregnant or lactating. In addition, we excluded participants who reported total daily energy intake outside the 800-4200 kcal range. Information regarding demographic characteristics, health status, history of smoking, current medications and supplement intake was collected. Also, we obtained self-reported information on socio-economic status, which included ownership of nine household items (home, car, dishwasher, washing machine, LCD TV, refrigerator, woven carpet, laptop computer and microwave). Having < 3, 4 to 6, and > 7 of those household possessions were considered low, moderate and high socioeconomic status, respectively. 37

Anthropometric indices
Body weight and height were measured using a tape measure and a standard scale with 100-gram accuracy. Anthropometry was assessed while participants wore light clothing and after removal of their shoes. Waist circumference (WC) was measured at the halfway point between the lower margin of the lowest rib and the top of the iliac crest and by using stretch-resistant tape. Participants stood in a relaxed position with their feet

Dietary intake
We collected dietary data using a 168-item food frequency questionnaire (FFQ) based on Willett et al. 42 The reproducibility and validity of the FFQ has been evaluated specifically in Iranian adults. 43 It includes a comprehensive list of diverse kinds of bread and grains, legumes, white and red meats, dairy products, fruits and vegetables, nuts, fats, and other foods. Food consumed per day, week, month, and year can be determined using this questionnaire. 44 After administering the FFQ, we then converted the reported intakes to grams per day using standard published guidelines. 45 51 Using residual adjustment, we adjusted PRAL, NEAP and DAL for energy intake in the statistical analyses. Nutrient adequacy was computed via the nutrient adequacy ratio (NAR). 52 We calculated NAR for energy and for ten nutrients including vitamin C, vitamin A, vitamin B1, vitamin B2, phosphorus, iron, calcium, magnesium, zinc, and potassium. The mean probability of adequacy across these nutrients was obtained using Dietary Reference Intakes (DRIs). The NAR for nutrients is the ratio of a participant's intake/current recommended allowance (RDA) for each age and sex category. Mean Adequacy Ratios (MAR) were also calculated with following formula: 53 MAR (Mean Adequacy Ratio) =

Statistical analysis
Statistical analysis was done with SPSS for Windows version 16 (SPSS Inc, Chicago, IL, USA). The Kolmogorov Smirnov test was used to assess whether the variables were normally distributed. Pearson correlations were used to evaluate correlations between dietary acid load indices. Cut-off points for quartiles of dietary acid load indices were calculated and categorised based on tertile cut-off points. Significant differences in general characteristics across tertile categories of dietary acid load indices were tested using ANOVA. If there was a significant main effect, Bonferroni tests were performed to detect pairwise differences. The chi-square test was used to detect any significant differences in the distribution of participants across tertile categories of dietary acid load indices for categorical variables (i.e. prevalence of obesity, overweight, abdominal obesity and socio-economic status). To assess dietary acid load indices in relation to overweight, obesity and abdominal adiposity, we used multivariable logistic regression models. Multivariable models controlled for age (years), energy intake (kJ/d), physical activity, socioeconomic status, family history of diabetes, stroke and consumption of estrogen or dietary supplements. In the second model for overweight, obesity and abdominal adiposity risk assessment, we also adjusted for consumption of carbohydrates, fat, refined grain, and sodium.

Main results
Energy-adjusted means for dietary variables (macronutrients, micronutrients and food groups) across quartile categories of dietary acid load indices are presented in Table 2. Energy intake significantly differed among PRAL and DAL tertiles. Protein, fats (cholesterol, SAFA, PUFA), refined grains, meats and sodium, were consumed in greater amounts by participants in both the high PRAL and DAL groups. Also, participants scoring in the highest tertile of dietary acid load consumed lower amounts of carbohydrates, fruits, vegetables, fiber, potassium and magnesium compared with participants in the lowest tertile of all three dietary acid load scores. Also, calcium and multivitamin supplementation did not significantly differ between tertiles of dietary acid load (results not shown).
Multivariate linear regression analyses showed that dietary acid load (DAL), was positively associated with women's weight and BMI (Table 3). PRAL, NEAP and DAL were strongly positively correlated with protein, meat sand meat products, and refined grains, whereas they were negatively correlated with fruits, vegetables, fiber, magnesium and potassium. Models controlled for calorie intake. With increasing dietary acid load scores (PRAL, NEAP and DAL), both of the dietary quality indices (DQI-I and MAR) significantly decreased (P < 0.001), whereas energy density significantly increased (P < 0.001) across tertile categories of dietary acid load indices. The results for dietary quality indices were similar for both crude and multivariate adjusted models (Table 4). Multivariate-adjusted odds ratios (OR) and 95% CIs for overweight, obesity and abdominal adiposity across tertile categories of dietary acid load scores are presented in Table 5. The odds of abdominal adiposity increased with tertile of DAL in the crude model (adjusted OR = 1.98, 95% CI = 1.13-3.49, P = 0.016). Abdominal obesity remained strongly associated with higher DAL even after controlling for confounding variables in Models 1 (adjusted OR = 2.77, 95% CI = 1.39-5.50, P = 0.004) and 2 (adjusted OR = 2.4, 95% CI = 1.21-4.50, P = 0.005). A similar relationship between the DAL score and obesity risk was found after adjustment for confounders (adjusted OR = 2.41, 95% CI = 1.01-5.74, and P = 0.045). Other dietary acid load indices (PRAL and NEAP) showed no significant associations with obesity, overweight or abdominal obesity.

Discussion
The findings of this cross-sectional study of Iranian women further the existing evidence regarding possible associations between dietary acid load with overweight and obesity. Our conclusions from this study are three-fold. First, our findings reveal an association between DAL and BMI, obesity and abdominal obesity. Second, in line with existing literature on this subject, we observed inconsistent relationships between overweight and specific measures of dietary acid load, noting no associations between PRAL and NEAP and weight. Third, we demonstrate, for the first time, that DAL may be used to measure dietary quality. We discovered robust associations between higher scores on 'quality of diet' measures DQI-I and MAR and lowered PRAL, NEAP, and DAL scores.
The positive associations between DAL with weight and      48 However, the association we observed between DAL and waist circumference was weaker than in that study. We also found that higher levels of dietary acid load, measured with DAL, were associated with increased odds of obesity and abdominal obesity after adjustment for confounding variables. Findings regarding the other dietary acid load indices (PRAL or NEAP) were in agreement with several previous studies i.e. they revealed no significant relationship with weight, waist circumference or BMI. [55][56][57][58] Differences in our results regarding anthropometric indices in relation to DAL (as opposed to the two other indices PRAL and NEAP) may be due to the formula used to calculate DAL, in which height and weight are directly PRAL, potential renal acid load; NEAP, net endogenous acid production; DAL, dietary acid load; DQI-I, diet quality index international; MAR, mean adequacy ratio; ED, energy density. * Data are mean ± SE and computed by using analysis of covariance in multivariable adjusted models. The crude model is not adjusted for any variables. Adjusted model: adjusted for age and total energy intake, physical activity, social economic status, family history of diabetes and stroke, estrogen drugs and supplements used. † P values resulted from ANCOVA test.
used to compute this score. Our results were not consistent with findings from other studies that suggested that higher PRAL score was associated with increased weight, waist circumference and/or BMI. 28,57 For example, one previous study found an independent positive association between BMI/WC and PRAL. In that study the mean BMI of that population was higher than in our study (BMI = 27-27.3 vs. 24.8 kg/ m 2 ). 28 In a study of Japanese students, higher levels of dietary acid load were significantly related to BMI/WC and pro:K (protein to potassium) ratio, but not with PRAL. 57 Likewise, in a study by Krupp et al conducted in healthy children and adolescents, findings for PRAL were dependent on participant BMI and weight. 59 To date, the association between dietary acid load and obesity or overweight has been examined in two cross-sectional 38,60 and two prospective studies. 21,61 In the cross sectional study, PRAL was not associated with the prevalence of overweight and obesity, 38 while in other studies higher PRAL scores were significantly associated with higher rates of overweight and obesity, 21,60,61 which was not confirmed by our study.
Differences in study results may be related to varied food patterns, different ways to measure dietary acid load, type or size of the study sample, variation in the assessment of dietary intake and to differences in which confounders were adjusted for. Moreover, in contrast to our study, prior studies were conducted in older adults (mean age = 32.4 y), 21,38,48  In all prior studies, older age was strongly correlated with higher dietary acid load scores. One explanation for this is that the capacity to excrete acid falls significantly with age because of decline in renal function. 21,28,38,48,62 Another possible explanation may be that older people are inclined to consume higher acid load diets. 55,60 Several possible mechanisms have been posited to explain the relationship between dietary acid load and obesity. First, higher dietary acid load leads to stimulation of glucocorticoid production and increased serum cortisol levels. 63 Acidbase status affects renal magnesium losses, irrespective of magnesium intake. 64 These processes may cause insulin resistance and subsequent overweight or obesity. [64][65][66] Second, diet-induced acidosis decreases the production of adipokines (leptin and adiponectin), which can inhibit appetite suppression. 67,68 Also, it has been demonstrated that higher dietary acid load reduces lean body mass among women and ultimately leads to higher body fat synthesis. 69 However, these mechanisms are not clearly understood and need further investigation.
Consistent with previous studies, 48,60,[70][71][72] the correlations among PRAL, NEAP and DAL were quite high. This suggests that, although not identical, these measures capture similar nutrient elements. Although PRAL, NEAP The crude model is not adjusted for any variables. ‡ Model 1 is adjusted for age and total energy intake, physical activity, social economic status, family history of diabetes and stroke, estrogen drugs and supplements used. § Model 2 is additionally controlled for significant dietary intakes in Table 2. We controlled for the effects of carbohydrates, fat, refined grains and sodium. || P values are P for trend based on the Mantel-Haenszel test. and DAL are considered reasonably valid indices of dietary acid load, more research is needed to determine which is a better indicator. 22,73 Similar to previous reserach in Iranian adults, 28 57 This may have been due to the dietary pattern observed in our study that included relatively higher quantities of fruits and vegetables and lower quantities of meat, in comparison with these studies.
Recent epidemiological studies have used PRAL, NEAP and DAL scores frequently to estimate dietary acid load. 48,55 PRAL score takes into account intestinal absorption rates nutrients that affect pH homeostasis (protein, potassium, calcium and magnesium) and the dissociation of phosphate at pH 7.4. A positive absorption rate reflected in a positive PRAL score represents acid-forming potential, whereas a negative PRAL value indicates base-(or alkaline) forming potential. 22,47 The NEAP score is calculated by taking the dietary intake ratio of protein to potassium consumed. 51 The DAL score combines PRAL and body surface area. 48 For all three dietary acid load scores, increasing values measure greater consumption of acid-inducing foods.
We observed an independent negative? association between all three dietary acid load scores and DQI-I and MAR. To our knowledge, this is the first study to assess how dietary acid load and dietary quality indices are associated. Our results were similar to findings from many studies that showed that higher intakes of meats, refined grains, fish, egg and fat groups were associated with higher PRAL, NEAP and DAL, and that the higher intakes of fruits and vegetables were related to lower scores. 56,57,61 Conversely, more than half of the points scored on DQI-I correspond to the consumption of fruits, vegetables, calcium and low intake of simple carbohydrates and fats (total fat, cholesterol and SAFA). These nutrients were significantly associated with dietary acid load scores (PRAL, NEAP and DAL) in the present study. Also, the MAR score was negatively associated with PRAL, NEAP and DAL. The MAR score typically includes two nutrient categories influence dietary acid load scores. The first includes potassium, calcium and magnesium, which are responsible for lower levels of dietary acid load. The second group of nutrients is positively associated with dietary acid load and includes protein and phosphorus. Although other nutrients included in the MAR score are not directly calculated in dietary acid load, they can indirectly contribute to the dietary acid load estimate because they represent protein, fruits and vegetable food sources. It appears that our study population consumed larger quantities of dietary sources containing acidreducing factors (fruits and vegetables) than food sources with protein and phosphorus. ED was the third dietary quality index that we investigated in this study. In contrast with DQI-I and MAR, ED had a positive relationship with dietary acid load scores. ED represents of the amount of calories in one gram of total daily food intake. 54 In a previous cross-sectional study, fat and protein intake appeared to be positively associated with energy intake when energy density was replaced by its nutrient correlates. 74 At the same time, fiber and water intakes were negatively associated with energy intake. Therefore, our study revealed that greater dietary acid load due to higher consumption of protein leads to increases in ED score. Several studies have reported that healthy dietary patterns, such as the Mediterranean diet and Dietary Approaches to Stop Hypertension (DASH) that are high in fruits and vegetables increase diet quality. 75,76

Conclusion
In conclusion, we assessed diet-induced acidosis using several methods: PRAL, NEAP and DAL scores. In our study, the correlations between PRAL, NEAP and DAL scores were high (Pearson correlation coefficient > 0.78, P < 0.001), which is consistent with previous studies (r = 0.95, P < 0.001). 62,77 A main limitation of our study is its cross-sectional design, which does not allow for analysis of cause and effect. Second, categorization of dietary acid load scores into tertiles could lead to data misclassification. Also, the semi-quantitative FFQ could be prone to recall bias suggesting that misclassification is another limitation.
We found that higher DAL was associated with higher risk of obesity and abdominal obesity in Iranian women after adjusting for possible confounding factors. Also, independent positive associations between PRAL, NEAP, DAL (more acidic dietary acid-base loads) and DQI-I and MAR were observed. Furthermore, diet-induced acidosis was associated with increased ED scores. Our findings show that these dietary acid load indices can be considered indicators of dietary quality. To confirm these results, the associations between dietary acid load and dietary quality indices should be tested in future prospective studies and clinical trials. Since this study was limited to healthy women, more research is needed to generalize these results to the whole population, These results can help guide clinical and policy guidelines for the adoption of appropriate dietary patterns to prevent and control weight loss and obesity.