Introduction

According to the World Health Organization, childhood obesity is one of the most emergent global public health concerns in the current century [1]. Prevalence of this crisis in children aged 2–19 years was 17% in United States between 2011 and 2014 2]. Moreover, the prevalence of childhood obesity in Kuwait (between 2012 and 2013) [3], Bangladesh (in 2016) [4] and Iran (in 2007) [5] was estimated to 30.5, 23 and 4.7%, respectively. The link of childhood obesity and adulthood chronic diseases such as type 2 diabetes, cardiovascular diseases and osteoarthritis has been demonstrated in previous studies [6,7,8].

Acid–base homeostasis has an important role in human health. Negative health effects of acidosis or increased acid loading include overweight, obesity, increased waist circumference (WC), infant growth retardation, unfavorable sport performance, increased atrophy and higher loss in muscle mass [9,10,11,12]. Diet composition might affect on the acid–base balance [13]. A diet rich in acidogenic foods including animal products (e.g. meat and fish), cheese, cereals and rice and low in alkaline foods (e.g. fruits and vegetables) induce endogenous acid production [14]. Poor dietary intake, impaired calcium-citrate balance, and cortisol-induced acidosis is related to the lipid profile abnormalities, development of obesity, and cardiovascular diseases (CVDs) [15]. It has been demonstrated that Western dietary pattern (high consumption of acidogenic foods including animal products as well as lower intake of alkaline foods such as fruit and vegetables are associated with higher risk of type 2 diabetes [16,17,18]; diet-induced acidosis may explain this association [19]. The association between imbalanced dietary acid–base load and defects in bone metabolism, hypertension (HTN) and chronic kidney diseases has been shown [20].

Two scores were developed as indicators of dietary acid load (DAL); the potential renal acid load (PRAL) score, which is based on protein, phosphorous, calcium, magnesium and potassium intake [9], and the net endogenous acid production (NEAP) score, that is based on total protein and potassium intake [9, 20]. The validity of both scores compared to a 24-h urinary acid load was examined and confirmed, previously [9]. In a large cross-sectional study, Iranian women with higher NEAP score had higher weight, WC and serum triglyceride (TG) levels [12]. Moreover, a study conducted on Japanese workers indicated that higher NEAP and PRAL scores were associated with insulin resistance [21]. Higher DAL score was associated with HTN in American [22] and young Japanese women [23], while no association was observed in older Swedish men [24] and Dutch adults [25].

To our knowledge, there is no information on the association of DAL and anthropometric indices in children and adolescents living in developing countries with various socioeconomic status (SES) and dietary habits, which may have different effects on DAL. Therefore, the present study was conducted to examine the association between DAL scores, assessed by both PRAL and NEAP, and anthropometric indices in a large representative sample of Iranian children and adolescents.

Materials and methods

Ethical standards

The present study was approved by the ethics committee of Isfahan and Alborz universities of Medical Sciences (Project code: 194049) and it was carried out based on the principles of the Declaration of Helsinki. The objectives and procedures of the study were explained to students and their parents. Eligible students and their parents who volunteered to participate in the study initially provided verbal consent and then signed a written informed consent at the base of the study.

Study design and population

Data for the present study were extracted from the complementary survey (the weight disorders survey) of the 4th phase of a national project entitled “Childhood and Adolescence Surveillance and Prevention of Adult Non-communicable disease (CASPIAN-IV)” (2011–2012). All protocol details have been reported previously [26]. Briefly, in the current national multi-center cross-sectional survey, 5326 students aged 6–18 years were recruited. Between 2011 and 2012, eligible subjects were recruited from urban and rural regions of 30 provinces of Iran using a multi-stage, cluster sampling method. Finally, 83 clusters (each contains 10 students) in each province of the country with equal sex ratio were selected. As a result, 14,400 students were examined. To assess dietary intake, about a quarter of clusters from any province, in total 6505 children and adolescents were selected. To perform statistical analysis, students with any special diet (n = 104), those with a history of chronic diseases such as type 1 diabetes, (CVDs) and cancer (n = 212), and those taking medications (n = 88) were excluded. In addition, students with an incomplete Food Frequency Questionnaire (FFQ) (775 subjects) were excluded as well. At last, data from a total of 5326 students were available to assess the association between DAL and anthropometric indices in these Iranian children and adolescent (Fig. 1).

Fig. 1
figure 1

Flow diagram of subjects’ enrollment

Full size image

Data collection

To collect necessary data, structured questionnaires with a face-to-face interview with students and their parents were used. The internal consistency of the questionnaire was tested previously and found to be excellent (overall Cronbach’s alpha = 0.93). Based on test–retest measurement in an earlier CASPIAN study [27], the reliability of the questionnaire was confirmed as 0.94. In children less than 10 years, their parents were asked; students, 10 years of age or older, completed the questionnaire independently. In the present study, demographic information, SES, dietary assessment, and physical activity (PA) were examined.

Demographic characteristics

A validated questionnaire was prepared according to the World Health Organization Global School-based Student Health Survey (WHO-GSHS) [26] and it was used to collect demographic information of participants. The demographic questionnaire contained age, sex, the family history of chronic diseases such as diabetes, obesity, dyslipidemia and hypertension; educational level of parents, place of residence, dietary behaviors, sleep hours, time spent watching TV and using computers.

SES assessment

The principal component analysis (PCA) method was applied to determine the SES of students. Questions included in this analysis contained parental education, type of house, parent’s occupation, having private car, having a personal computer, and type of school (public/private). The aforementioned parameters were incorporated as a unit index [28]. The obtained index was categorized into three grades of SES (low, moderate, and high). Questions about SES were asked to the student's parent for children < 10 years, while students ≥ 10 years old completed the questionnaire independently.

Anthropometric measurements

Body weight (Wt), height (Ht), WC, hip circumference (HC), neck circumference (NC) and wrist circumference were measured by trained staff in the selected schools. Each participant’s Wt was measured without shoes and wearing light clothing (precise: 0.1 kg). Ht was assessed without shoes to the nearest 0.1 cm with standard measurement against a stadiometer. All WC, HC and NC were measured using a non-elastic meter with the precise of 0.1 cm. WC was evaluated in the midpoint between the lowest rib and the iliac crest, while the students standing and breathing out. HC was also measured at the largest circumference of the ileac crest with no pressure to the body. The circumference below the Adam’s apple was measured at a comfortable status to determine NC.

BMI-z scores of the participants, by age and gender, were computed according to WHO growth reference for children aged between 6 and 18 years old. To calculate BMI for parents, Wt (kg) was divided to the square of the Ht (m2) [BMI = Wt (kg)/ Ht(m)2] [29, 30]. Waist-to-hip ratio (WHR) and waist-to-height ratio (WHtR) were calculated by dividing WC to HC and WC to Ht, respectively. Tri-ponderal mass index (TMI) was calculated as weight (kg) divided by the cubic of the Ht (m3) [31].

PA assessment and Screen time (ST)

To determine the student’s PA, a validated Physical Activity Questionnaire for Adolescents (PAQ-A) was used. The total score of PAQ-A questionnaire ranged between 1 and 5. Finally, PA was categorized as low (score between 1 and 1.9) and high (score between 2 and 5).

To estimate total ST during a week, an average number of hours/day for watching TV, personal computer (PC), and playing electronic games were considered. ST was categorized into two low and high score (less or more than 2 h/day) [32, 33].

Dietary assessment

Using a validated 168-item FFQ (Cronbach's alpha coefficient = 0.96), usual dietary intakes of the participants were assessed [27, 34, 35]. The FFQ for children < 10 years old was completed by the student’s parent; students ≥ 10 years old completed the form by themselves. The frequency and standard portion size of each food item intake through the previous year on daily, weekly, monthly, and yearly basis was reported. After converting total frequencies of each food to daily frequencies, the amount of food was computed using household scale guide. An adopted version of Nutritionist IV corrected for Iranian foods (First Databank; Hearst, San Bruno, CA, USA) including the United States Department of Agriculture (USDA) food composition data, was used. For other foods that were not included in this database, Iranian food composition table was used [36].

DAL calculations

Urinary net acid excretion is influenced by dietary intake. However, its direct measurement is difficult. Accordingly, two dietary indices have been developed to estimate acid production in body. In the present study, PRAL and NEAP were used to estimate DAL. PRAL was estimated based on the following formula [9]:

$$\begin{aligned} {\text{PRAL }}\left( {{\text{mEq}}/{\text{day}}} \right) = & \, 0.4888 \times {\text{protein}}\;{\text{intak}}e \, \left( {{\text{g}}/{\text{day}}} \right) \\ & & \; +\;0.0366 \times {\text{phosphorus }}\left( {{\text{mg}}/{\text{day}}} \right) \\ & \;{-}\;0.0205 \times {\text{potassium }}\left( {{\text{mg}}/{\text{day}}} \right) \, \\ & \;{-}\;0.0125 \times {\text{calcium }}\left( {{\text{mg}}/{\text{day}}} \right) \, \\ & \;{-}\;0.0263 \, \times {\text{magnesium }}\left( {{\text{mg}}/{\text{day}}} \right). \\ \end{aligned}$$

NEAP was also calculated based on the equation as follows [20]:

$${\text{NEAP }}\left( {{\text{mEq}}/{\text{day}}} \right) \, = \, \left[ {54.5 \, \times {\text{protein intake }}\left( {{\text{g}}/{\text{day}}} \right) \div {\text{ potassium intake }}\left( {{\text{mEq}}/{\text{day}}} \right)} \right] -\;10.2.$$

The validity of both aforementioned indices compared to a 24-h urinary acid load was examined and confirmed, previously [9].

Statistical analysis

Normal distribution of continuous variables was assessed using Kolmogrov–Smirnov test. To present continuous variables with and without normal distribution, means (standard deviations (SD)) and median (interquartile range (IQR)) were used, respectively. Categorical variables were presented as percentages. Mean and median comparison of continuous variables between sexes was assessed using Student’s Samples t test and Mann–Whitney U test, respectively. One-way ANOVA was used to compare the trend of mean values of the anthropometric indices across PRAL and NEAP quartiles. Frequency of categorical variables were compared using Pearson chi-square test. The association of anthropometric indices (dependent variables) with PRAL and NEAP quartiles (categorical and continuous independent variables) was assessed using crude (Model I) and adjusted (Model II) linear regression models. Age, sex, place of residence, ST and PA, SES and carbohydrate, protein and total fat intake were adjusted in the adjusted model. The first quartile of PRAL and NEAP was considered as reference group. Results of linear regression was presented as Beta coefficient (β) and standard error (SE). Cluster sampling procedure was applied in all statistical methods. P values were corrected using the Benjamini–Hochberg correction method to control the false discovery rate due to multiple comparison phenomenon [37]. The adjusted P value less than 0·05 was considered as statistically significant. Data analyses were performed using STATA® software (version 11).

Results

The characteristics of study population according to gender and age range (6–12 and 13–18 years old) are presented in Table 1. As it is shown, 5326 students (46.9% girls), with mean (±SD) age of 12.5 (3.1) years, completed the study (response rate: 98.1%). Most participants were from urban areas (% 71.2) and had high ST (% 79.8). SES in girls aged 13 to 18 years old was significantly higher than in boys in the same age category (36.3 vs 31.6%, respectively). The levels of PA in boys were significantly higher than in girls in both age categories (P < 0.001 in both).

Table 1 Characteristics of study population according to age group and sex in the CASPIAN-IV study, 2011–2012
Full size table

Overall, Wt, Ht, wrist circumference, NC, WC, WHR in boys were significantly greater than girls (P < 0.05 for all), while girls had higher BMI z-score, TMI, HC and parental BMI compared to all boys (P < 0.05 for all).

There was a significant difference in energy intake and percent energy from carbohydrate, protein and fat among boys and girls (P < 0.05 for all). Both PRAL and NEAP in girls aged 13–18 years old were lower than in boys of similar age category (P < 0.001 for both) (Table 2).

Table 2 Potential renal acid load, net endogenous acid production and intake of energy and macronutrients according to age group and sex in the CASPIAN-IV study, 2011–2012
Full size table

Anthropometric indices based on quartiles of the PRAL are shown in Table 3. There were no differences among various anthropometric indices across PRAL quartiles in boy students, while parental BMI in the lowest category of PRAL was significantly greater than other categories in the parents of boys (P = 0.004) and NC was higher in the greatest quartile of PRAL. Girls with lower PRAL significantly had greater WC (P = 0.02) and HC (P = 0.002). In total, students, individuals in the lowest category of PRAL had higher HC (P = 0.04) and parental BMI (P = 0.01) (Table 3).

Table 3 Anthropometric measures according to quartiles (Q) of the potential renal acid load in the CASPIAN- IV study, 2011–2012
Full size table

Across the different anthropometric indices, NC in boys who were in the top category of NEAP was significantly higher than other categories (P = 0.001). Moreover, boys with higher NEAP significantly had lower parental BMI (P = 0.005). HC in girl students with lower NEAP was greater than other NEAP categories (P = 0.04). Furthermore, an increasing trend was observed in NC of total students across the quartiles of NEAP (P = 0.002). Parental BMI in total students who were in the top category of NEAP was significantly lower than other categories (P = 0.003) (Table 4).

Table 4 Anthropometric measures according to quartiles (Q) of the net endogenous acid production in the CASPIAN-IV study, 2011–2012
Full size table

The association between PRAL and anthropometric indices in study population is provided in Table 5. Among the various anthropometric indices in crude model, PRAL showed the strongest significant inverse association with HC (β = − 1.27, standard error (SE) = 0.52). However, after age, sex, living area, ST, PA, SES and carbohydrate, protein and total fat intake adjustment, the association was no longer significant (β = − 0.26, SE = 0.41). In the adjusted model, there was a significant association between PRAL and NC in the highest category compared to the reference one (β = 0.48, SE = 0.17). Nevertheless, in both crude and adjusted models, an inverse association between PRAL and parental BMI remained significant in the all three top categories compared to the lowest one (reference group). There was no association between PRAL and other anthropometric indices.

Table 5 Beta estimates and standard errors for the association between the potential renal acid load and anthropometric measures in the CASPIAN-IV study, 2011–2012
Full size table

Students with higher NEAP (last quartile) had greater NC (β = 0.57, SE = 0.19); after adjusting for covariates, we found that all three top categories of NEAP were greater than the reference group (Table 6). In addition, in crude and adjusted models, there was an inverse association between NEAP and parental BMI in the highest quartile compared to the first one (β = − 0.56, SE = 0.20 and β = − 0.54, SE = 0.23, respectively). There was no association between NEAP and other anthropometric indices.

Table 6 Beta estimates and standard errors for the association between the net endogenous acid production and anthropometric measures in the CASPIAN- IV study, 2011–2012
Full size table

Discussion

The present study is one of the first studies on the association between food NEAP and PRAL with anthropometric measures in a large nationwide representative sample in Iran. We found that students with higher NEAP, had greater NC. Also, there was an inverse association between PRAL and NEAP with parental BMI. However, there was no significant association between PRAL and NEAP with other anthropometric indices.

As it was noted, we found a direct association between NEAP and NC. NC is a beneficial index measuring the distribution of upper subcutaneous adipose tissue which [38] can use to identify overweight and obese participants [39, 40]. In a cross-sectional study carried out on Brazilian children aged 8 and 9 years, higher adherence to Snacks dietary pattern was associated with increased NC [41]. Also, another observational study conducted on 290 overweight and obese women, indicated that higher adherence to Western dietary pattern (high intake of acidogenic foods) was associated with higher NC [42].

In this study, there was no association between PRAL and BMI z-score. Similar relationship was observed for NEAP. The relationship between DAL and anthropometric indices can be related to the effects of dietary acid–base load on muscle metabolism. Previous research indicated that acidogenic diets resulted in muscles loss through decreased synthesis, increased proteolysis, and amino-acid oxidation [43]. In line with the present study, a cross-sectional study conducted on women revealed no association between PRAL or NEAP and BMI [12]. Likewise, another observational study carried out on Japanese employees demonstrated no association between DAL and BMI [44]. One study in Japanese adults revealed a positive independent relation between the Pro: K ratio and BMI [23]. Moreover, Han et al. demonstrated a negative association between PRAL and BMI in Korean population [45]. In another cohort, study conducted on aged Sweden participants, results demonstrated a positive association between PRAL and BMI in both men and women [46]. These discrepancies in the findings of studies can be attributed to the observed differences in sample sizes, lifestyles, dietary habits, and races in study participants.

Moreover, in our study, there was no association between DAL and WC. In line with our findings, Han et al. in an observational study conducted on a random sample of 11,601 Korean men and women observed no association between PRAL and WC [45]. In another study carried out on patients with diabetic nephropathy, the Pro: K ratio was cross-sectionally associated with WC, but not PRAL [47]. Also, Mozaffari et al. found a positive association between NEAP and WC in Iranian women [12]. Moreover, in a study aimed to evaluate the association between DAL and cardiometabolic risk factors in Iranian adults, results demonstrated a significant relationship between PRAL/Pro: K and WC [15]. The observed discrepancies between the results of studies can be attributed to differences in sample sizes of studies and health status in study participants.

In our findings, there was an inverse association between PRAL/NEAP and parental BMI. In the investigation of 387 middle-class French families, the correlation between protein intake in parents and their offsprings was observed [48]. In another extensive study carried out among 1,283 Korean households aimed to assess the correlation of some nutrients intake between family members, the findings illustrated a positive correlation between calcium and phosphorus consumption in children/adolescents and their parents [49]. Davison et al. in an observational study of 401 child–parent pairs from New Zealand demonstrated that higher parental diet quality score was associated with lower children’s snack consumption [50]. These results presented the correlation between parent’s diet and their offspring’s diet.

To the best of our knowledge, the present study is the first attempt regarding this association in pediatric population. Results of this study provided further evidence about the possible association of NEAP and NC. Our findings suggest that the association between NEAP and NC is independent of potential confounders including demographic variables, PA, SES levels and carbohydrate, protein and total fat intake; however, because of the cross-sectional nature of our study, the causal relationship remained unclear. Also, some misclassifications of biases because of an FFQ usage to assess dietary intakes should be also taken into account. Thus, more possible associations need to be explored in further investigations. In addition, no information was available regarding renal function of participants, which has an important role in acid–base homeostasis. However, as our study population was apparently healthy children, it can be assumed that almost all of them had normal renal function. In addition, we did not measure PRAL and NEAP scores directly; they were estimated from dietary intake. However, dietary PRAL and NEAP values are widely used in epidemiological studies and showed strong correlation with acid load measured from 24-h urine [9]. The DAL might simply reflect differences in nutrients and in macronutrients of usual diet. In addition, the FFQ was not validated for the intake of phosphorus, potassium, calcium, magnesium that used for calculating PRAL and NEAP. Moreover, PRAL/NEAP are estimate values, derived from the FFQ and not calculated on the basis of a food diary or from a quantitative analysis of the meal. So less effort will have to be put into the conclusion.

Conclusions

In conclusion, our findings suggest limited association between diet-induced acid load and some anthropometric measures in pediatric age group. Further longitudinal studies are recommended to determine the relation of DAL and anthropometric measures, and to explore the underlying mechanisms.

What is already known on this subject?

It has been demonstrated that Western dietary pattern (high consumption of acidogenic foods including animal products as well as lower intake of alkaline foods such as fruit and vegetables) are associated with higher risk of type 2 diabetes; diet-induced acidosis may explain this association. The association between imbalanced dietary acid–base load and defects in bone metabolism, hypertension (HTN) and chronic kidney diseases has been shown. In a large cross-sectional study, Iranian women with higher net endogenous acid production (NEAP) score had higher weight, waist circumference and serum triglyceride levels. Moreover, a study conducted on Japanese workers indicated that higher NEAP and potential renal acid load (PRAL) scores were associated with insulin resistance. Higher dietary acid load (DAL) score was associated with HTN in American and young Japanese women, while no association was observed in older Swedish men and Dutch adults.

What does this study add?

To our knowledge, the present study is the first attempt regarding the association of DAL and anthropometric indices in pediatric population living in developing countries with various socioeconomic status and dietary habits, which may have different effects on DAL. We found that students with higher NEAP, had greater neck circumference. Also, there was an inverse association between PRAL and NEAP with parental body mass index. However, there was no significant association between PRAL and NEAP with other anthropometric indices.