Nutrient Intakes of Canadian Children and Adolescents: Results from the Canadian Community Health Survey (CCHS) 2015 – Nutrition Public Use Microdata Files

Background: Up-to-date and appropriate estimates of the usual intakes of key nutrients are crucial for monitoring the nutritional adequacy of the Canadian population. Comprehensive, nationally-representative nutrient estimates for Canadian children and adolescents are available using data from the Canadian Community Health Survey (CCHS) – Nutrition 2004, however results are scarce for the most-recent 2015 data. The objective of this research was to assess nutrient intakes of Canadian children and adolescents using data from the CCHS 2015 Public Use Microdata Files (PUMF) Methods: Participants’ rst 24-hr dietary recall, and the second-day recall from a subset of participants were used to estimate usual intakes of macronutrients, vitamins and minerals in children and adolescents (2-18 years). Usual intakes by Dietary Reference Intake (DRI) age-sex groups were estimated using the National Cancer Institute (NCI) method, adjusted for age, sex, misreporting status, weekend/weekday, and sequence of recall analyzed (rst/second), with consideration for outliers (N=5,493). Usual intakes from food and beverages were assessed for prevalence of inadequacy in relation to the DRI recommendations. Results: Children 2-3y consumed a percentage of total energy from protein above the Acceptable Macronutrient Distribution Range (AMDR). Among children and adolescents, 71% met the AMDR for carbohydrates and 63-71% met the AMDR for total fat. With the exception of calcium and vitamin D, a very low prevalence (<10%) of inadequate intakes was observed for select micronutrients among Canadian children 2-8y, while a substantial proportion of adolescents (>15%) did not meet requirements for vitamin A, vitamin C, vitamin D, calcium, magnesium and zinc. Conclusions: Canadian children and adolescents may not be meeting recommendations for short fall nutrients such as calcium, potassium, vitamin A and vitamin D. Fibre intake continues to remain low, while sodium intake exceeded recommendations. These ndings provide important, updated baseline estimates on the nutrient intakes of this subpopulation in Canada for continued monitoring of adherence to the 2019 Canada’s Food Guide and may be useful to inform future public health nutrition policies, programs and initiatives. This study presents results of analyses from the most-recent, nationally-representative health and nutrition survey of Canadians to estimate usual nutrient intakes of Canadian children and adolescents, and assesses the prevalence of nutrient inadequacy in relation to the DRIs. Results from this study showed that the majority of Canadian children and adolescents are meeting AMDR recommendations for macronutrients; however, many children may not be meeting requirements for vitamin D or calcium, while many adolescents may not be meeting requirements for vitamin A, vitamin C, vitamin D, calcium, magnesium and zinc. For nutrients with an AI, there is a concern that both children and adolescents may not be meeting their needs for potassium and bre, while consumption of sodium was in excess of international recommendations.


Introduction
Handling misreporting of dietary intake To account for over-and under-reporting of energy intake, reported energy intakes (EI) in 2015 CCHS were compared to each respondents' total energy expenditure (TEE) based on the method proposed by Garriguet (32). Under-and over-reporting were de ned as the ratio of EI:TEE < 0.7 and >1. 42, respectively; those in between were considered plausible reporters. TEE was calculated based on availability of age, sex, height, weight and physical activity levels (i.e. sedentary, low active, moderately active and highly active), as per the Institute of Medicine's energy expenditure equations (33). Body Mass Index (BMI) categories for children aged 18 years or younger were de ned by Cole et al. (34) and were calculated using measured height and weight for the subsample of respondents who had these measurements. For children where measured height and weight were not available, USDA equations were used to estimate TEE; these equations require age, sex and physical activity level only (35). If physical activity levels were unavailable, children younger than 14 years of age were assumed to be low active while teenagers aged 14 years or older were assumed to be sedentary based on the cut-offs proposed by Garriguet (32).

Assessing usual intakes and nutrient inadequacy
Since the previous 2004 CCHS-Nutrition, new statistical methods for the analysis of usual dietary intakes have been proposed (36), with the NCI method recommended by Health Canada for the analysis of 2015 CCHS data (26, 36). Advantages of using the NCI method include the ability to adjust for covariates of interest in the estimation of usual intake, and more precise estimations of the usual intakes of episodically consumed foods (37,38). In this study, the NCI method was applied using both days of 24-hr dietary recall where available to estimate all usual nutrient intakes across DRI age-sex groups (39).
Application of the NCI method to derive usual nutrient intakes, including considerations for the type of model to use (e.g. amount only or two-part), choice of covariates, strati cation vs. pooling of the data by speci c age-sex groups, and outlier removal were guided by recommendations from a joint Statistics Canada/Health Canada technical working group, detailed in the paper by Davis et al. (28,36). For micronutrients, the amount-only univariate NCI macros were used (26). For percent of calories from macronutrients, the amount-only bivariate macros were used (26). The following covariates were adjusted for in all NCI models (with the exception of energy, where misreporting status was not adjusted for): age, sex, misreporting status, sequence of recall analyzed ( rst/second); and weekend/weekday recall, with Friday considered a weekend.
Outliers with implausible nutrient intakes were removed from all analyses as proposed by Davis et al (36). All means, percentiles and proportion of the population above or below a DRI cut-off were generated using the NCI method (26). Further details on the development and application of the NCI method can be found elsewhere (37,38).
NCI-derived usual nutrient intakes were assessed for nutrient adequacy by comparing the estimated intakes to the age-and sex-speci c nutrient recommendations found in the DRIs (23,33,40). Since DRIs have not been de ned for saturated fat, intake of this nutrient was compared to the World Health Organization (WHO) cut-off of 10% of total energy (5). Based on the threshold proposed by Health Canada, low prevalence of inadequacy was de ned in this study as population estimated intakes <10% of the Estimated Average Requirement (EAR) (41,42). In the case of iron for females of menstruating age (aged 14-18y), the full probability approach was used to assess nutrient adequacy (43, 44), as the iron requirements distribution is known to be skewed for this age/sex group (24,39,43,44).

Statistical Analysis
Analyses were performed using R-Studio (v1.1.447) and SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Data manipulation was conducted using R-Studio, while analysis was conducted using SAS. Bootstrap balanced repeated replication (BRR) with 500 replicates was used to estimate con dence intervals, standard errors and coe cients of variation due to the complex sampling design of 2015-CCHS. Survey weights provided by Statistics Canada were applied to all analyses to ensure nationally-representative estimates (28). Descriptive statistics based on estimated usual dietary intakes (e.g., means, standard errors and percentiles) and prevalence of inadequacy for nutrients were calculated and presented by nutrient for each age-sex group. Data manipulation and analytic codes, for 2015-CCHS PUMF, are available from the authors upon request.

Percentage of Energy from Macronutrients
Mean percentages of total energy from carbohydrates, total fat and protein were 49%, 16%, 33%, respectively (Table 1, Figure 1a/1b, Table S2-S4). With the exception of 2-3y, estimated mean percentages of total energy from protein were within the Acceptable Macronutrient Distribution Range (AMDR), while a smaller proportion of Canadian children and adolescents had energy intakes from carbohydrates (~70%) and total fat within the AMDR (~64%). All intakes are based on food and beverage consumption only and exclude intakes from supplements. 2 The National Cancer Institute method (NCI method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI mode day of the week (weekend vs. weekday), and sequence of dietary recall analysed ( rst or second); dietary misreporting was not adjusted for to derive usual in were de ned and removed using the methodology reported in Davis et al (36). The NCI methodology was applied to each nutrient by DRI age-sex grouping se between sample sizes and point estimates. The number of respondents removed due to outlier methodology varied between 1-5 respondents for each nutrien population-level estimates using sampling survey weights provided by Statistics Canada. All reported SEs were bootstrapped using the 500 boot weights prov  1 All intakes are based on food and beverage consumption only and exclude intakes from supplements. 2 The National Cancer Institute method (NCI method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI models: age, sex, dietary misreporting status, day of the week (weekend vs. weekday), and sequence of dietary recall analysed ( rst or second); dietary misreporting was not adjusted for to derive usual intake of energy. Outliers for nutrient intake were de ned and removed using the methodology reported in Davis et al (36). The NCI methodology was applied to each nutrient by DRI age-sex grouping separately, hence small discrepancies between sample sizes and point estimates. The number of respondents removed due to outlier methodology varied between 1-5 respondents for each nutrient. All estimates were weighted for population-level estimates using sampling survey weights provided by Statistics Canada. All reported SEs were bootstrapped using the 500 boot weights provided by Statistics Canada. All intakes are based on food and beverage consumption only and exclude intakes from supplements. 2 The National Cancer Institute method (NCI method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI mode vs. weekday), and sequence of dietary recall analyzed ( rst or second); dietary misreporting was not adjusted for to derive usual intake of energy. Outliers for reported in Davis et al (36). The NCI methodology was applied to each nutrient by DRI age-sex grouping separately, hence small discrepancies between samp due to outlier methodology varied between 1-5 respondents for each nutrient. All estimates were weighted for population-level estimates using sampling surv bootstrapped using the 500 boot weights provided by Statistics Canada. All intakes are based on food and beverage consumption only and exclude intakes from supplements. 2 The National Cancer Institute method (NCI method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI mode dietary misreporting status, day of the week (weekend vs. weekday), and sequence of dietary recall analyzed ( rst or second); dietary misreporting was not ad derive usual intake of energy. Outliers for nutrient intake were de ned and removed using the methodology reported in Davis et al (36). The NCI methodology each nutrient by DRI age-sex grouping separately, hence small discrepancies between sample sizes and point estimates. The number of respondents removed methodology varied between 1-5 respondents for each nutrient. All estimates were weighted for population-level estimates using sampling survey weights pro Canada. All reported SEs were bootstrapped using the 500 boot weights provided by Statistics Canada. 3 For detailed intake distributions and comparisons to the Dietary Reference Intakes (DRIs) for each nutrient, see Supplementary Tables (S1-S27). 4 Sample size after outliers were removed.
With respect to fat-soluble vitamins, the prevalence of inadequate intakes was high for fat soluble vitamins (vitamin A, vitamin C and vitamin D). A substantial proportion of Canadian adolescents aged 9-18 y (males: 38-52%; females: 18-41%) consumed vitamin A in quantities below the EAR (Table 3, Figure 2, Table   S12). Although the prevalence of inadequate vitamin C intakes were low for adolescents aged 9-13y, >16% of adolescents 14-18y had inadequate intakes of this nutrient (Table 3, Figure 2, Table S13). Almost all Canadian adolescents (>92%) had a high prevalence of inadequate intake of vitamin D (Table 3, Figure 2, Table S14).
Inadequate intakes of trace elements were also observed for male and female adolescents. The prevalence of inadequacy for calcium ranged from 58-60% for male adolescents and 75-77% for female adolescents (Table 3, Figure 2, Table S21), whereas for magnesium, prevalence of inadequate intakes were slightly lower at 16-53% for male adolescents and 25-75% for female adolescents; for phosphorus, 18% of male adolescents and 31-37% of female adolescents had inadequate intakes, respectively (Table 3, Figure 2, Table S23-S24). Although a low prevalence of iron inadequacy (<5% below EAR) was found among adolescent males (9-18y) and females (9-13y), approximately 25% of females 14-18y consumed iron in amounts that fell below their EAR (Table   3, Figure 2, Table S22). A substantial proportion of Canadian adolescents had inadequate intake of zinc, ranging from 15-21% for males and 23-28% for females (Table 3, Figure 2, Table S25).
With the exception of vitamin A and zinc among children aged 2-3y, the proportion of intakes from food and beverages that were greater than the Tolerable Upper Intake Level (UL) was very low for most vitamin and minerals.

Intakes of Vitamins and Minerals with an Adequate Intake (AI) or Chronic Disease Risk Reduction (CDRR) level
The majority of Canadian children and adolescents (> 93%) had estimated bre intakes below the AI for their respective age-sex groups ( Table 2, Table S10).
The median potassium intakes of Canadian children and adolescents was below their respective AIs (2300-3000 mg/day), and thus, no assessment can be made regarding the prevalence of inadequacy of potassium (Table 3, Table S26).
More than half (66%) of the sample exceeded the CDRR for sodium. Males 14-18y had the highest estimated mean sodium intake at 3737 ± 84 mg/d, with

Discussion
This study presents results of analyses from the most-recent, nationally-representative health and nutrition survey of Canadians to estimate usual nutrient intakes of Canadian children and adolescents, and assesses the prevalence of nutrient inadequacy in relation to the DRIs. Results from this study showed that the majority of Canadian children and adolescents are meeting AMDR recommendations for macronutrients; however, many children may not be meeting requirements for vitamin D or calcium, while many adolescents may not be meeting requirements for vitamin A, vitamin C, vitamin D, calcium, magnesium and zinc. For nutrients with an AI, there is a concern that both children and adolescents may not be meeting their needs for potassium and bre, while consumption of sodium was in excess of international recommendations.
Although the available literature on nutrient intakes of children and adolescents is limited, results from this study are consistent with those published on macronutrient intakes of Canadian children and adolescents using one day of dietary recall from the 2015 CCHS-Nutrition, is consistent with results on Canadian children and adolescents using the 2004 CCHS-Nutrition (17,45), and is also consistent with results for US children and adolescents using the 2015-2016 National Health and Nutrition Examination Survey (NHANES) (46, 47). For example, inadequate intakes of vitamin A, vitamin D, magnesium and calcium is a concern for adolescents in both countries and within Canada between 2004 to 2015, and both children and adolescents may not be meeting their requirements for potassium and bre (17,18,45). Furthermore, the mean estimated energy intakes of Canadian children and adolescents reported in this study were similar to those found in previously published ndings using 2004 CCHS and 2015 CCHS-Nutrition -the only exception being for 1-3y, where we observed a mean energy intake of 1,967 kcal/day in our study compared to 1,585 kcal/day in 2004-CCHS and 1,308kcal/day in 2015-CCHS (48). However, this difference may be largely attributable to the exclusion of 1-year olds from our study (as they do not have a TEE equation), as well as methodological differences between this study and other published studies (48) (further discussed in Strengths and Limitations).
In this study, saturated fat contributed to ~10% of total energy in the diets of children and adolescents. A high-energy, high-saturated fat diet and low activity lifestyle that contributes to obesity and NCDs in adults may be having the same impact on children and adolescents (8). The Canadian Health Measures Survey found that among Canadian children and adolescents, 7% had either borderline or elevated blood pressure measurements, a risk factor for cardiovascular disease (49). Highly processed foods are a major source of saturated fat in the Canadian diet and can also be high in calories, sodium and free sugars (50). Ultra-processed foods contribute more than 50% of total daily energy in the diets of Canadian children and adolescents (51). Considering the mounting evidence that increased consumption of ultra-processed foods has a negative impact on diet quality and health, continued surveillance on nutrient intakes and interventions (e.g., food policies and regulations addressing the food environment, marketing to kids etc.) to reduce consumption of unhealthy processed foods will remain a priority.
An assessment of Canadian children and adolescents' bre intake with the recommendation is not possible given the limited usefulness of the AI in assessing nutrient adequacy of groups. Despite this limitation, the importance of an appropriate bre intake should still be promoted to young Canadians, as recent reports indicate decreasing intakes of fruits, vegetables and whole grains, which are an important source of bre and potassium (52).
Given its importance in the development of peak bone mass, low calcium intake is a signi cant health concern for children and adolescence and may play a role in the risk of fractures and developing osteoporosis later in life (53). Other than factors such as lactose intolerance or cultural/family practices and suboptimal intake of calcium may be linked to replacement of dairy products with sugar-sweetened beverages (46). Research has shown that soft-drink consumption is highest in adolescence while milk intake is lowest (54)(55)(56). In addition to calcium, milk is an important source of other shortfall nutrients such as vitamin D, phosphorus, magnesium, potassium, vitamin A, and zinc (8). Adolescent females are likely to drink less milk in order to cut kilocalories, favoring low-kilocalorie soft drinks (57).
Nearly 100% of Canadian children and adolescents had a high prevalence of inadequate intakes for vitamin D, similar to results from 2004-CCHS Nutrition (17,45). This is a likely a re ection of a signi cant decrease in uid milk consumption (58, 59), as all uid milk in Canada is forti ed with Vitamin D, and the change in DRI recommendations from an AI (5ug/day for children and adolescents) to a higher EAR (10ug/day) in 2011 (22). Estimates of the prevalence of inadequate intakes of vitamin D from food must be interpreted with caution as Vitamin D can also be synthesized by the body from sunlight (UV radiation) Alternatives" and "Meat and Alternatives" food groups now replaced with a "Protein Foods" group. Given these changes, it may be important to monitor both calcium and vitamin D intakes of children and adolescents in the coming years.
The largest prevalence of inadequate intakes of essential nutrients, such as vitamin B6, vitamin B12, folate, zinc and iron, were seen for adolescent girls 14-18 years. These nutrients are of particular importance for adolescent females as the commencement of menstruation during this age group demands for increased nutritional needs. Similar results were reported previously for female adolescents in the US population (67). Environmental and social factors (e.g. dieting, body image concern) can lead to disordered eating behaviours (68), possibly increasing the risk of inadequate intakes of essential nutrients within this speci c population group. A study of Canadian adolescents found that 4.5% of females met the criteria for an eating disorder (69). Considering these ndings, adolescent females may be a particularly vulnerable for malnutrition; highlighting the importance of understanding the intake practices of this group to promote better health and nutrition.
Although sodium intakes were lower in this study compared to 2004 (70), Canadian children and adolescents are still consuming too much sodium in comparison to the CDRR. This is a major concern as recent Canadian data indicates that 7% of Canadian children and adolescents now have either borderline or overt hypertension (49). Additionally, emerging evidence indicates an association between high sodium intake and being obese and risk of developing cardiovascular disease in children and youth (71)(72)(73). Moreover, these high intakes of sodium may be a speci c cause for concern for certain sub-populations of children and adolescents. For example, research has shown that children born with low birth weight may be at a higher risk for developing hypertension if they consume excess sodium across the lifespan (74). Due to an increased risk of cardiovascular disease associated with high sodium intakes, populationlevel sodium reduction interventions, such as the reformulation of speci c food products, have been recommended (23) and have resulted in effectively decreasing sodium levels across the food supply chain in Canada (75).
The Canadian government has initiated several policies since 2015 to improve the nutritional outcomes of children and adolescents. As part of Health Canada's Healthy Eating Strategy, speci c recommendations are being considered for food products speci cally designed for children and youth (76). As an example, nutrition labelling regulations have been updated to include the percentage daily value of sodium on the Nutrition Facts table on packaged foods marketed to children aged 1-4 years old (77). A promising policy initiative to restrict marketing of unhealthy foods and beverages via traditional and digital marketing advertising (e.g. television advertising) to children under the age of 13 years is also being considered (77). Introduction of such policies may be particularly impactful in reducing the consumption of processed foods in childhood and to prevent NCDs in adulthood. Findings from the current study may act as a benchmark to monitor progress in the quality of Canadian children and adolescents' dietary intakes given the government's proposed policy changes.

Strengths and Limitations
This is the rst study to utilize nationally representative data from 2015 CCHS -Nutrition to estimate nutrient intakes and the prevalence of nutrient inadequacy among Canadian children and adolescents, in addition to providing interpretation of the results and implications of these ndings. Additionally, caution should be exercised when making comparisons between published results on 2004 and 2015 CCHS-Nutrition due to methodological differences in data collection, data processing and data analysis between the two surveys (28). Some of the methodological differences include differences in sample size and response rates, updates to the nutrient databases, use of usual intake estimation methods (e.g., Software for Intake Distribution Estimation (SIDE) vs. NCI), handling of outliers and adjustment for certain covariates in estimation models. Further details on the differences between the two survey cycles can be found elsewhere (28).

Conclusions
Childhood and adolescence are important stages of human development with pronounced physical growth, reproductive maturation and cognitive transformation. The current study presents less-than-recommended intakes of some essential nutrient among Canadian children and adolescents; this indicates an urgent public health importance to improve the nutrient and diet quality of children and adolescents in Canada. Childhood and adolescent malnutrition presents a signi cant burden to health, and negatively impacts economic development and prosperity; the importance of adequate nutrition in children, especially among food insecure or marginalized communities, is unequivocal. These results provide insights into speci c nutrients for target age-sex groups that may be an important foci for public health interventions aimed at improving diet quality and nutrient adequacy in Canada.   The National Cancer Institute Method (NCI Method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI models: age, sex, dietary misreporting status, day of the week (weekend vs. weekday), and sequence of dietary recall analysed ( rst or second). Outliers for nutrient intake were de ned and removed using the methodology reported in Davis et al. (36). The number of respondents removed due to outlier methodology varied between 1-5 respondents for each nutrient. The analytical sample size before outlier removal was n=5,493.. All estimates were weighted for population-level estimates using sampling survey weights provided by Statistics Canada. All reported SEs were bootstrapped using the 500 boot weights provided by Statistics Canada. The SE represents the more conservative (i.e., larger) standard error estimate from the NCI output representing the proportion of the population meeting either the upper or lower AMDR bound.
b. Percentage of total energy from macronutrients, Canadian adolescents (9-18y). Data Source: Statistics Canada, Canadian Community Health Survey, Nutrition (2015) -Public Use Microdata File. All intakes are based on food and beverage consumption only and exclude intakes from vitamin or mineral supplements. The National Cancer Institute Method (NCI Method) for estimating usual dietary intake was used. The following covariates were adjusted for in all NCI models: age, sex, dietary misreporting status, day of the week (weekend vs. weekday), and sequence of dietary recall analysed ( rst or second). Outliers for nutrient intake were de ned and removed using the methodology reported in Davis et al. (36). The number of respondents removed due to outlier methodology varied between 1-5 respondents for each nutrient. The analytical sample size before outlier removal was n=5,493.. All estimates were weighted for population-level estimates using sampling survey weights provided by Statistics Canada. All reported SEs were bootstrapped using the 500 boot weights provided by Statistics Canada. The SE represents the more conservative (i.e., larger) standard error estimate from the NCI output representing the proportion of the population meeting either the upper or lower AMDR bound.