Beneficial nutrients are almost always consumed together with significant amounts of energy. Therefore, intake of energy and vitamin and minerals are often strongly correlated. Because energy intake is actively regulated by appetite and satiety, energy requirements may be satisfied well before requirements for vitamins and minerals are met, and this may lead to malnutrition (Backstrand, 2003). Despite our extensive understanding of human nutritional requirements for maintenance of health, malnutrition continues to be one of the main causes of morbidity and mortality in low- and middle-income countries, especially in young children (Black et al., 2008). Also, many children in developing countries suffer from the effects of food insecurity resulting in deficits in energy, protein, and micronutrients such as vitamin A, iron, and zinc. These deficits are associated with low weight- and low height- for- age as well as micronutrient deficiencies (Ramakrishnan and Huffman, 2008). The assessment of pediatric diet quality is of high value because food habits and behaviors that develop in childhood can track over time and predict adult diet-related disease (Craigie et al., Patterson et al., 2009). Thus, an investigation into the diet quality that contributes to healthier dietary intakes in children is a public health priority.
Various diet quality models have been developed to assess the overall nutritional quality of the total diet (Drewnowski, 2005, Nicklas et al., 2014). One of those models is the Nutrient-Rich Foods indices (NRF). The NRF index has been tested and validated against the quality of the total diet index such as the Healthy Eating Index (HEI) (Alkerwi, 2014, Fulgoni et al., 2009). The NRF9.3 index, based on nine nutrients to encourage and three nutrients to limit and per 100 kcal, explains the highest percentage of variation from HEI, index performance declined with the inclusion of additional vitamins and minerals. Previous studies confirmed that increasing the number of nutrients over 10 in a nutrient profile model provided little or no extra benefit in predicting overall diet quality (Drewnowski and Fulgoni, 2014).
Previous studies have not investigated the relation between NRF index, obesity and stunted growth in children. Hence, this study aimed to determine the cross-sectional association between the NRF index and malnutrition in girls in Tehran.
Materials and Methods
Participants: A school-based, cross-sectional study was conducted with 400 children (girls) aged 6.5-8.4 from 12 primary schools laid in Tehran, Iran, from September 2016 to May-2017. Multi-stage cluster random sampling was used to select the study sample. Since the Z-scores of malnutrition indices are defined separately for the girls and boys, the boys were excluded in this study. Written informed consent was obtained from the children's mothers. The children selected for this study were healthy female children without any diseases who were from 6.5 to 8.5 years old, and had not received any nutritional supplements during the latter year. The exclusion criteria were having a history of any diseases affecting growth, having dead or divorced parents, having energy intake of less than 1000 or more than 4000 kcal, and also those who provided incomplete questionnaires.
All the required confirmations were obtained from the Department of Education in Tehran. Students were interviewed together with their parents at the school. The study’s data has been collected by trained staff throughout two and a half months. Tehran is divided into five districts, and five regions were randomly selected from each district as follow: The northeast district: regions 1 and 4. The southeast district: region 14. The southwest district: region 19. The center district: region 7. Two primary schools were selected in each mentioned region and the parents of first grade students were invited for participation in this study. Most parents were willing to participate. A total of 400 students provided complete data and were included in the final analyses.  
Assessment of dietary intake: Dietary data for each child was collected from their parents by interviewers. Diet quality was assessed using the valid and reliable food frequency questionnaire (168-item) over the past month (Asghari G et al., 2012). If the child was present with an illness other than a simple cough or fever, and the mother reported the child had a poor appetite or diarrhea, the dietary data was collected in another time. The information was collected through a face-to-face interview.
Calculation of the NRF index scores: The NRF index has two components: (i) the nutrient-rich (NR) component which is based on a variable number of beneficial nutrients and (ii) the limiting nutrients (LIM) component. The NRF9.3 score is based on the sum of the percentage of reference daily values (DVs) for nine beneficial nutrients protein, dietary fiber, vitamin A, vitamin C, vitamin E, calcium, magnesium, iron and potassium minus the sum of the percentage of maximum DVs for three nutrients to limit saturated fat, total sugar, i.e. mono- and disaccharides, and sodium (Fulgoni et al., 2009) (Table 1). This index can be used as follows: Firstly, all foods consumed by each subject are scored using the NRF9.3 algorithms based on Drewnowski’s method (Drewnowski, 2009), and this resulted in a NRF9.3 score (per 100 kcal) for every food item, i.e., a NRF9.3 food score. The recommended daily allowances as set by the Institute of Medicine (IOM) of the National Academies (Dietary Reference Intake (DRI)) were used as reference DVs (Institute of Medicine, 2005, Institute of Medicine (US) Panel on Dietary Antioxidants and Related Compounds, 2000, Institute of Medicine (US) Standing Committee on the Scientific Evaluation of Dietary Reference Intakes and its Panel on Folate, 1998, Institute of Medicine Standing Committee on the Scientific Evaluation of Dietary Reference Intakes, 1998, 2001) (Table 2). The percentage of reference DV for each nutrient is capped at 100% DV to avoid overvaluing of food items that provide very large amounts of a single nutrient. Secondly, the NRF9.3 food scores were converted to individual NRF9.3 index scores by multiplying the amount of calorie consumed of each food item, in 100-kcal units, by the NRF9.3 food scores and then summing these scores for each subject. Thirdly, the NRF9.3 index scores were divided by the number of 100-kcal units of the participant’ total energy intake to provide a ‘weighted average’ diet quality score. Higher NRF9.3 index scores indicated higher nutrient density per 100 calories, and thus, the participants with a high NRF9.3 index score were considered to have a healthier dietary pattern than those with a low NRF9.3 index score.
Assessment of anthropometric measures and Demographic data: The researchers collected anthropometric data during the interview at school using validated, standardized methods. Weight measurements of parents and children were recorded to the nearest 0.5 kg using a SECA 761 medically approved flat mechanical scales. Height was recorded to the nearest millimeter using a Leicester portable height stick. Children were classified into body mass index (BMI) categories using the World Health Organization's child growth standards Z-scores. Measured parent BMI was classified according to the World Health Organization (WHO) classifications as normal weight (< 25 kg/m²), overweight (≥ 25 and <30 kg/m²) or obese (≥ 30 kg/m²)(WHO, 2000⁾.
Information on physical activity (PA), screen time, sleeping hours, parent's birthday date, parent's education, family dimension, geographical region were obtained from the parent questionnaire. Child's PA was measured from the physical activity questionnaire Baecke (Baecke Jos, 1982, Moghaddam et al., 2012). Child's screen time was derived from the time (in hours) spent during weekdays in (1) TV watching and (2) playing non-active video games or using the computer.
Parent’s education was used as a proxy of family socioeconomic status and was recoded into four categories. The four education categories were 1: primary school, category 2: guidance school, category 3: high school, category 4: diploma, and category 5: college/university or above. Information provided by the parents were about the child's usual bedtime, an estimate of how long they stay awake during the night and also the usual wakeup time. Also, the parents were asked about the duration of any naps during the day.
Ethical considerations: This study was approved by the Ethics Committee of Iran, University of Medical Sciences with the ethics code of IR.IUMS.REC.1395.9321680001.
Data analysis: The normality of the data was assessed using the Kolmogorov-Smirnov test. The baseline characteristics of the participants were compared between quintiles of the NRF9.3 index score using analysis of variance for continuous variables and the chi square test for categorical variables. Independent sample t-tests was used for normally-distributed data to compare the NRF index in different weight and height status. Linear regression analysis was used to study the association between the NRF9.3, NR9 and LIM3 index scores and malnutrition indicators like weight-, height- and BMI- for age. The multiple regression models were adjusted for liquid oil, legume, meat products, fruit, sugar, sweet sauce, snacks, fats, oil, and savory sauce intake. All statistical analyses were carried out using the Statistical Package for the Social Sciences version 22 (IBM, SPSS Inc., Chicago, USA). Two-sided P-values < 0.05 were considered statistically significant.
Results
Demographic Characteristics: Table 2 shows the distribution of demographic characteristics by the NRF9.3 index quintiles. The mean age of children was 7.25 ± 0.65 years. No significant difference was observed between children's age, sleep hours, screen time, physical activity, and parents' BMI in different quintiles of NRF index (P > 0.05). According to WHO's child growth standard Z-scores regarding BMI-for-age, 12.25% of our participants were overweight and obese (49 students), 34.50% were underweight (138 students), and 53.25% were normal (213 students). The mean weight was 25.50 ± 6.26 kg, and the mean height was 122.51 ± 7.08 cm. The mean of NRF9.3 index for the whole sample was 64.69 ± 30.98 (Table 2). The mean of total calories consumed was 1,411.95 ± 407.82 kcal.
Table 3 shows the difference in food items consumed among quintiles of the NRF9.3 index score. Participant in the highest quintile of the NRF9.3 index score consumed remarkably more liquid oil (P < 0.001), legume (P = 0.037), meat products (P = 0.005), fruit (P < 0.001), less sugar & sweet sauce (P < 0.001), snacks (P = 0.002), fats, oil and savory sauce (P = 0.002) comparing to those in the lowest quintile.
NRF9.3 index and anthropometric measures: In linear regression analysis, the NRF9.3 index score was positively associated with height-for-age after adjustment for confounding factors (from β = -0.03 (-0.17, 0.88, 95% CI) in the first quintile to β = 0.09 (0.27, 1.12, 95% CI) in the fifth quintile, P = 0.039). Whereas, the NR9 index score tended to increase height but it was not significant (P = 0.056). There was not a significant association between NRF9.3, NR9 and LIM3 index scores and weight- and BMI-for-age (Table 4). The mean of the NRF9.3 index for normal weight and overweight and obese was 64.00 ± 33.06 and 66.46 ± 25.28, respectively. There was no significant difference in the mean of the NRF9.3 index for overweight and obese children compared to normal weight (P = 0.47). Furthermore, the mean of the NRF9.3 index for normal and stunted children was 64.61 ± 31.16 and 71.54± 10.06, respectively. There was no significant difference in the mean of the NRF9.3 index for stunted children compared to the normal height (P = 0.619).