Introduction

Adequate growth is considered one of the best indicators of physical well-being in children [1]. Important growth indices for children under age five include malnutrition (stunting, wasting, and underweight) and overweight [2, 3]. Stunting, or linear growth faltering, is a major public health problem most often occurring in lower-income regions such as Africa and Asia [4, 5], and that can lead to low academic performance [5]. In Asia and Oceania, wasting puts nearly one in ten children under five at increased risk of mortality, highlighting the need for preventive interventions and programs [6]. In terms of over-nutrition, children who are overweight tend to become obese in adulthood and are more likely to develop non-communicable diseases like diabetes and cardiovascular disease [7]. Along with a variety of health-related (e.g., infectious diseases, intrauterine growth retardation, low birth weight, insufficient exclusive breastfeeding) and social determinants (e.g., low socioeconomic status, low parental education, and poor health services), parental DP is a major factor influencing childhood growth [8,9,10,11]. Parents have the potential to play a major role in their child’s diets by limiting their access to unhealthy products such as salty and processed foods, sweets, and sugar-sweetened beverages and by introducing healthy foods at home [12, 13].

A large study of 2692 mother–child pairs revealed that dietary quality scores for mothers and their children were weakly to moderately correlated [14]. However, the similarities in the maternal and child’s DPs were stronger for total fat intake than they were for the overall dietary quality score [14]. Another study of 121 urban African-American adolescent-mothers pairs showed no correlation between nutrients and food groups for mothers and sons but a weak correlation in mother-daughter pairs [15]. A study of 1640 children showed a strong positive association between diet quality of 3-year-old children and maternal diet quality [16]. Ovaskainen et al. observed only a slight similarity between the dietary patterns of mothers and their pre-school aged children [17]. Regarding the association between child’s DP and child anthropometry, Mehranfar et al. found an inverse association between a white meat dietary pattern in children who are both underweight and wasted and also a positive association between white meat protein intake in children and risk of overweight [18]. Also, in Iran, child’s healthy beverages index (HBI) was associated with higher odds of wasting and underweight [19]. Moreover, high-fat milk intake was inversely associated with wasting among children [19]. Given these contrasting findings, more research is needed to understand how different dietary patterns are associated with child anthropometry. Moreover, most of the existing literature regarding the correlation between maternal and child’s dietary patterns has been from western countries, while the current study evaluates the correlation in a different less-studies population. Therefore, we aimed to study the association between maternal and child’s dietary patterns as well as how different dietary patterns are associated with child anthropometry in Tehran province, Iran.

Methods and materials

Study population

Data for this cross-sectional study were collected from 788 mother–child dyads in 2017–2018. Cluster random-sampling methods were used to select mother–child peers who had attended public health centers in southern Tehran, Iran, for the screening of primary school entering and vaccination of 6-year-old children. A random sample of 1200 mothers with 6-year-old girls was contacted, out of which (75%) 900 mothers consented and completed the study questionnaire. Out of 900 subjects, data was driven from 788 mother–child peers. The main reason for this sample drop (12.4%) was lack of parental interest during the study.

Mothers provided written informed consent regarding publishing their and their child’s data. The inclusion criteria of the research included: Six-year-old girls who did not have the prior chronic or malignant disease diagnoses were eligible for participation, female gender, Iranian, being in the age range of 20–50 years, women were not pregnant, breastfeeding or in post-menopausal status, women did not follow any specific diet (e.g., restriction diet due to allergies, celiac disease, weight loss, or weight gain diets), did not have any acute or chronic diseases (conditions such as diabetes, autoimmune diseases, cardiovascular disease, renal disease, liver disease, respiratory disease, hypertension, and allergy), and did not take medications for the mentioned diseases and medications for weight loss.

Children were excluded if their parents did not provide informed consent. This study was ethically approved (IR.NIMAD.REC.1398.211). Research reported in this publication was supported by Elite Researcher Grant Committee under award number 983073 from the National Institute for Medical Research Development (NIMAD), Tehran, Iran.

Dietary intake assessment

Dietary intakes were assessed with a reliable and valid 168-item semi-quantitative Food Frequency Questionnaire (FFQ) [20, 21]. To describe the relative validity and reliability of the FFQ, usual dietary intake was assessed twice using a 168-item semi-quantitative FFQ, 1 year apart (FFQ1 and FFQ2). Dietary data was also collected by twelve 24-h dietary recalls (24hDR) and biochemical markers to compare nutrient intakes from the FFQ and 24hDR in a total population of 132 subjects (sixty-one males and seventy-one females). All food groups reached acceptable validity [22]. Trained nutritionists asked the mothers about the frequency of their own and their children’s consumption of individual food items during the past year on a daily, weekly, or monthly basis. Portion sizes were converted to grams based on household measures. Individual dietary intake was analyzed using NUTRITIONIST 4 (First Data Bank, San Bruno, CA) software to estimate energy and nutrient intake.

Assessment of anthropometry, physical activity, and socioeconomic status

Anthropometric indices including weight and height were measured for all children and their mothers. Weight was measured on an electronic scale (Seca 753E; Seca Weighing and Measuring Systems, Hamburg, Germany) with a precision of 100 g. Children wore light clothes and no shoes or socks. Height was measured with 0.5-cm precision, while mothers and children stood with their shoulders in a normal position, without wearing shoes. Body mass index (BMI) was calculated by dividing weight (in kilograms) into height squared (in square meters). WHO criteria were used to define childhood growth indices. BMI-for-age and sex were categorized into obese (BMI-for-age ≥ 2SD), overweight (BMI-for-age ≥ 1SD), normal (BMI-for-age ≥  − 1SD and < 1SD), underweight (BMI-for-age <  − 1SD), and wasted (BMI-for-age <  − 2SD) [23]. Physical activity (PA) levels were measured for all children and their mothers using the International Physical Activity Questionnaire (IPAQ). IPAQ includes 16 questions about participation in three domains of physical activity (at work, traveling, and during recreational activities) as well as sedentary behavior. Scores were assigned based on the frequency, time, and intensity of activities (light, moderate, high, and very high-intensity activities) during the past year. For each participant, the amount of time for each PA was multiplied by their corresponding metabolic equivalent task value (MET-h/wk), and summed [24]. Socioeconomic status (SES) was evaluated using a valid and reliable scale developed to estimate SES and its relationship with health outcomes among Iranians [25]. This scale has several questions regarding occupation, academic education, vehicle ownership, home ownership, home appliances, number of people living in the household, number of rooms in the house, and domestic and international trips during the previous year which assess the socio-economic status of the individuals. All the question, answer options, and scores given to each answer are listed in Supplementary Table 1. After filling the information, the scored items of each question were scored based on logical idea. In this regard, the lowest value was scored for the lowest important item and the highest value was scored for the most important item. The total score can be assembled by summing up the scores of individuals for all items. Classification was based on tertile cutpoints. The first was considered as low SES, the second tertile as middle SES, and the third tertile as the high SES.

Statistical analysis

Statistical analyses were performed using SPSS software (version 23, SPSS Inc., Chicago, IL, USA). p < 0.05 was considered statistically significant. First, based on the similarity of foods consumed, the 168 daily consumed food items were categorized into 19 food groups (salty foods, soft drinks, hydrogenated fats, sweets and desserts, total grains, high-fat dairy products, low-fat dairy, starchy vegetables, red meats, egg, vegetable oils, poultry, legumes, fruits, nuts, olives, non-starchy vegetables, and fish). Then we used the residual method to adjust these food groups for total energy. Then, principal component analysis (PCA) was performed with rotated varimax to reduce the 19 food groups to three DPs. Eigenvalues > 1.00 and inflection in the scree plot were criteria used to select how many factors to retain from the factor analysis. A factor loading of > 0.3 was used as a criterion to define which food groups were dominant in each factor [26,27,28]. The Kolmogorov–Smirnov test was used to check the normality of the data for the original food groups. Then, factor scores for each participant and each DP were categorized into quartiles for further analysis. A comparison of the continuous variables across quartiles of the child’s DPs was performed using an analysis of variance (ANOVA) test. Analysis of covariance (ANCOVA) was used for estimating energy-adjusted children’s dietary intakes across quartiles of child’s DPs. Dietary macro-nutrient and micro-nutrient intakes across the tertiles of dietary patterns were compared using analysis of covariance while adjusting for daily energy intake. Correlations between maternal and child’s DPs were estimated using Pearson correlations. Finally, binary logistic regression was used to assess the associations between child’s DPs and child anthropometric indices in three different models: (1) a crude model; (2) a model adjusted for energy intake, child PA, and SES (Model 1); and (3) a model adjusted those variables in Model 1 plus mother’s age, mother’s BMI, and mother’s PA (Model 2).

Results

General characteristics of study participants (mother-daughter dyads) are presented in Table 1. The three major DPs labeled for children and their mothers are presented in Supplementary Table 2. The most relevant foods in the first child’s DP included salty foods, soft drinks, hydrogenated fats, sweets and desserts, total grains, high-fat dairy products, and starchy vegetables. The most dominant foods in the second pattern consisted of red meats, egg, vegetable oils, poultry, and legumes. The last child DP’s most relevant foods included fruits, nuts, olives, non-starchy vegetables, and fish. Non-starchy vegetables, legumes, red meats, egg, poultry, starchy vegetables, low-fat dairy products, salty foods, fruits, and olives were the major components in the maternal prudent DP. The most relevant foods in the Western DP included sweets and desserts, tea and coffee, soft drinks, and high-fat dairy products. Nuts, hydrogenated fats, and fish were the major components of the high fat DP.

Table 1 General characteristics of study participants (mother-daughter dyads)
Full size table

Mother’s characteristics across quartiles of child’s DPs are shown in Supplementary Table 3. Maternal age (p = 0.022) and maternal PA (p = 0.001) were strongly inversely associated with child’s Western DP. Maternal BMI (p = 0.008) and child energy intake (p < 0.001) were positively associated with child’s Western DP. Further, a strong inverse association was labeled between mothers’ age (p = 0.024) and child’s high-protein DP. Maternal BMI (p = 0.008), SES (p < 0.001), and child BMI-for-age z-score (p = 0.019) were positively associated with child’s high-protein DP. Child’s prudent DP was positively associated with SES (p < 0.001) and child energy intake (p < 0.001). Participant dietary intakes in each quartile of child’s DPs are presented in supplementary Table 4.

The correlations between maternal and their children’s DPs are summarized in Table 2. We observed a positive correlation between the child’s Western DP and all maternal DPs (correlation coefficient 0.201, 0.449, 0.253 and p < 0.001). Child’s high-protein diet was positively correlated with maternal prudent and Western DPs (correlation coefficient 0.576, 0.111 and p < 0.001). In addition, this pattern was negatively correlated with maternal high fat DP (correlation coefficient − 0.183 and p < 0.001). In addition, child’s prudent DP was positively correlated with maternal prudent and high fat DPs (correlation coefficient 0.426, 0.396 and p < 0.001).

Table 2 Correlation between maternal and child’s dietary patterns
Full size table

Table 3 displays the distribution of children with wasting, underweight, overweight, and obesity as well as odds ratios (OR) and 95% confidence intervals (CI) for crude and adjusted models of child anthropometric outcomes in relation to child’s DPs quartiles. A child being in the top compared to the lowest quartile of the high-protein DP was negatively associated with odds of underweight and wasting in the crude (OR 0.44; 95% CI 0.24–0.80) and adjusted (OR 0.43; 95% CI 0.23–0.80) models. No significant associations were found between child’s prudent DP and child anthropometric outcomes. The association between Western DP and child anthropometry was also non-significant.

Table 3 Multivariable odds ratio (OR) and 95% confidence intervals (CI) for risk of overweight/obesity and underweight/wasting in quartiles of child’s dietary patterns
Full size table

Discussion

This study is one of the first to investigate the associations between maternal and child’s DPs as well as relationships between child’s DPs and child anthropometric outcomes in Iran. We observed a significant positive correlation between the Western DP in children (salty foods, soft drinks, hydrogenated fat, sweets) and all three maternal DPs including prudent (non-starchy vegetables, legumes, red meat, eggs, poultry, low fat dairy, olives, fruits), Western (sweets, tea and coffee, soft drinks, high fat dairy), and high fat (nuts, hydrogenated fat) DPs. In addition, child’s high-protein DP (red meats, egg, poultry, and vegetable oils) was positively associated with maternal prudent and Western DPs. Child’s prudent DP (fruits, nuts, olives, non-starchy vegetables, fish) was also positively associated with maternal prudent and high fat DPs. However, there was a negative association between child’s high-protein DP and maternal high fat DP. Moreover, in both crude and adjusted models, we observed that children in the highest quartile of the high-protein DP had a lower likelihood of being underweight and wasted compared to children in the lowest quartiles.

It has been noted that parental eating behaviors shape child's dietary intake [29,30,31]. Findings from a sample of African-American families from a US Midwestern city showed that mothers’ intake of fruits, vegetables, and snack foods (i.e., cookies, cakes, chips, soda) were significantly correlated with their children's diets [32,33,34]. Results from Fisher et al. revealed that parents who consumed more fruits, vegetables, and milk had children with higher intake of these foods [29]. Klohe-Lehman and colleagues found that children of mothers who attended a weight loss program showed favorable dietary changes [35]. Longbottom et al. reported a significant correlation between fruits and chocolate intake in 5.5- to 8.5-year-old children and their mothers [36]. Several studies suggest that child’s fruit intake is significantly associated with maternal/parental fruit intake [37,38,39,40,41]. However, in a study by Gibson et al., maternal vegetable intake was not associated with child's vegetable intake [42]. The present study showed significant associations between SES and child’s prudent and high-protein DP, which is in line with previous studies. This finding may be explained by higher educational attainment in the high SES group, which has an important impact on mothers’ food choices [43, 44]. Also, the maternal Western diet, characterized by high consumption of foods such as sweets, soft drinks, hydrogenated fat, and sugar, was more common in children who had younger mothers, overweight mothers, and mothers who participated in less physical activity [33, 42, 45]. These findings are consistent with a study from the Avon Longitudinal Study of Pregnancy and Childhood that showed similar results [46]. In the present study, maternal Western DP was not associated with the prudent diet in their children. Conversely, we found that having a mother following the prudent diet was positively associated with children having a prudent diet at age six. This finding is consistent with earlier studies that have found similar associations between parents’ and young child’s diets [33, 47,48,49]. However, results of a review and meta-analysis by Wang et al. demonstrate that the influence of parent’s DP on young people’s dietary intake is not as strong as many people have assumed [50]. This weak association may be partially due to the difficulties of assessing the association between DPs of mothers and their children [50].

Our results showed an inverse association between child’s high-protein DP and odds of being underweight and wasted. This could be attributable to a significantly more favorable crude nitrogen balance caused by essential amino acids and high quality proteins [51]. Previous findings are inconsistent with our results; we found no association between the Western DP and the odds of being simultaneously underweight and wasted. Findings from a study by Sobhani et al. revealed that the modern DP (which is the same as the Western DP in our study) can lead to thinness in rural areas [52]. While our study adjusted for socio-economic factors, differences in our results and theirs may be due to the fact that some socio-economic factors were not controlled for in their study, like parental education and maternal age (which may affect the association between DPs and weight status). It should be noted that there are just a few studies documenting DPs and odds of underweight and wasting in children. Hence, more research may be needed to make firm conclusions regarding these results.

Results of the present study showed no significant association between DPs in children and odds of child obesity or overweight. Findings of Chong et al. and Norimah et al. revealed that the frequency of fruit and vegetable intake may not significantly influence children anthropometric status [53, 54]. However, a study by Matthews et al. in children aged 6.0–19.0 years in Southern California showed that higher intake of vegetables, but not fruits, was positively associated with a lower risk of overweight [55]. It should be noted that most of these studies had cross-sectional designs and failed to adequately adjust for potential confounders. Additionally, fruits and vegetables only account for a small component of the “prudent DP,” which was considered as a separate DP in our study. Our results are inconsistent with results of several studies that reported that the consumption of soft drinks (one of the main components of the Western DP) was positively associated with a higher total energy intake, leading to higher odds of obesity in children [56,57,58]. However, these studies evaluated the effect of only soft drinks, rather than the overall Western DP, on weight status.

Bahreynian et al. derived three major food patterns; “healthy,” “Western,” and “sweet-dairy” among 7- to 11-year-old children in Isfahan and found positive significant associations between the three DPs and obesity among girls [59]. These results are inconsistent with those of our study, as we found that children in the highest quartile of Western and prudent DPs did not tend to be overweight and obese compared to their counterparts in the lowest quartile. This difference in results may be due to differences in children’s ages. Another study by Johnson et al. showed that a high fat diet was positively associated with a higher risk of obesity and overweight in childhood [60]. However, one notable limitation of the mentioned study was that it had insufficient data on children’s physical activity. Also, the DP was considered a Western diet but only included some components of the Western diet. Contrary to our findings, Zhang et al. found that the modern DP (rich in milk, fast foods, eggs, other livestock meat, poultry, and cake) and the traditional north DP (rich in wheat flour-like noodles, other cereals, along with low consumption of rice, pork, and poultry) were positively associated with higher risk of obesity among 7- to 17-year-old children and adolescents [61]. The differences in the DPs reflect differences in types of foods consumed (including traditional foods) across different cultures. Moreover, dietary intake was assessed by using three consecutive 24-h dietary recalls which differed from our method of assessing dietary intake.

Literature has been inconsistent regarding the nature of the relationship between maternal and child’s DPs. Based on previous research, parents have a clear influence on their children's dietary intake [62]. However, reverse causation cannot be disregarded, i.e., that child’s diet will influence their mother’s dietary intake, especially in disadvantaged families where mothers may make sacrifices their diet to improve the adequacy of their children's diets [62]. The mechanism underlying the association between parents’ and their children’s dietary intake may be due to covert or overt types of food restrictions or incentives. Covert food restrictions involve keeping restricted foods (like unhealthy foods) out of children’s reach so that they cannot see or eat them [63]. The overt type of dietary restriction results from interactions between parents and children to decrease the available options and amounts of the restricted food items [64, 65]. Moreover, research has revealed that the opposite practice is also beneficial in stimulating the intake of healthy foods, i.e., when parents increase the portion sizes of healthy foods and enhance the diversity of healthy foods offered to children [66]. Experimental research has shown that parents can also promote children’s food preferences for healthy foods (such as some vegetables) by means of rewards, facial expressions, and repeated exposures [67,68,69,70,71].

Using a large sample size to evaluate of the relationship between major DPs and wasting among children along with the odds of overweight and obesity is one of the merits of the present study. A major limitation is that the dietary intake data were reported by the mother, since the children were not mature enough to provide accurate self-reports. Given this fact, these dietary data may reflect what the mothers imagine their children are eating rather than the children’s actual diets [72]. According to a study by Murakami et al. [73], subjects were identified as under or over-reporters of energy intake based on the ratio of reported energy intake to estimated energy requirement acceptable reporters were defined as having a ratio of 0.76 to 1.24. This ration was < 0.76 for under-reporters and ratio of > 1.24 for over-reporters. Based on these criteria, we did not have any misreporting of energy intake among children in the current study [74]. Also, we did not have any misreporting with energy intake of less than 500 and more than 3500 among mothers. In our study, given the mother reported both her own and her child’s food intake, the correlation between maternal and child dietary patterns may be overestimated. There is some evidence suggesting that what is reported as a child’s diet closely resembles the diet of the parent reporting their food intake [74]. Also, since our study only included girls, we are not able to generalize our findings to boys. Without longitudinal data, it is not possible to establish a true cause and effect relationship. Therefore, prospective studies are warranted to assess these associations over longer time periods. Results of the current study, regarding the association between child’s high-protein DP and decreased odds of children who are both underweight and wasted, suggest that better educating parents to assist their children to eat healthier foods such as enough high quality protein via a balanced diet should be encouraged. To improve our understanding of this issue, the larger picture should be considered, with attention to other influences on child’s eating patterns such as that of schools, the local food environment, peer influences, government’s policies that regulate school snacks and meals, and advertisements [75, 76].

Data transparency

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.