Introduction

Chronic diseases remain as significant issues in Canada, accounting for approximately 88% of all national deaths (World Health Organization, 2018). Although diet is one modifiable risk factor associated with chronic disease prevention, many Canadians are still not meeting current dietary guidelines for health (Hack et al., 2020, 2021). In North America, more than half of diet-related deaths were attributable to a low intake of fruit, vegetables, legumes, whole grains, and nuts and seeds (GBD 2017 Diet Collaborators, 2019). These plant-based foods, along with tea, coffee, and red wine, are significant sources of polyphenols, compounds long praised for their health-promoting properties, including chronic disease prevention.

Polyphenols are secondary metabolites distributed among plant foods that contribute to the bitterness, colour, flavour, odour, and oxidative stability in plant-based foods (Pandey & Rizvi, 2009). These compounds are believed to perform a plethora of biological actions including antioxidant, anti-inflammatory, anti-proliferative, and hormonal regulation capacities that may positively impact human health (Del Bo et al., 2020). A systematic review of observational studies reported a decreased risk of cardiovascular disease, type 2 diabetes, and mortality with higher polyphenol intake (Del Bo et al., 2020). Polyphenols exist as four main subclasses, including flavonoids, phenolic acids, lignans, and stilbenes; each of which can be further subclassified and is believed to convey different health benefits (Vetrani et al., 2018).

Although difficulties remain in establishing a recommended intake of polyphenols due to low bioavailability and extensive metabolism within the human body (Del Bo et al., 2020), the evidence regarding the observed health benefits of these compounds warrants further study. While several studies have assessed population-level polyphenol intake in countries across the globe (Carnauba et al., 2021; Zujko et al., 2012; Ziauddeen et al., 2019; Huang et al., 2020), no such data are available for Canadians. Therefore, the primary objective of this study was to estimate usual total daily polyphenol intake among Canadians aged 2 years and older using data from the 2015 Canadian Community Health Survey–Nutrition (CCHS-N). Additional analyses explored the major dietary contributors to total polyphenol, main class, and subclass intake; main class contributions to intake; and the potential impact of sociodemographic and lifestyle factors on intake.

Methods

Study population

The 2015 CCHS-N is a multi-stage, clustered, cross-sectional survey designed to collect dietary intake information from a nationally representative sample of 20,487 Canadians. To assess dietary intake, a 24-h recall method was used, which was adapted from the United States Department of Agriculture Automated Multiple-Pass Method. The recall was conducted on the total sample and repeated for a random sub-sample. Details regarding the survey design, sampling, weighting, and implementation of the 2015 CCHS-N can be found elsewhere (Health Canada, 2017).

For this study, respondents were excluded if missing dietary recall and/or demographic information, pregnant and/or breastfeeding, or less than 2 years old. A sensitivity analysis was performed to account for misreporting whereby the distributions of energy intakes on both recall days were analyzed and respondents with outlying energy intake on either day were removed. The final analytic sample included 94.74% of respondents with 37.02% having second dietary recall information (Fig. 1). Analysis of the 2015 CCHS-N Master Files was granted by Statistics Canada and conducted at the Statistics Canada Research Data Centre at Western University (Project No. 21-MAPA-UWO-6915).

Fig. 1
figure 1

Flow diagram of analytical sample derived from the 2015 Canadian Community Health Survey–Nutrition (n = 19,409)

Full size image

Estimation of polyphenol content of food and beverages

As the 2015 CCHS-N does not contain the polyphenol content of foods and beverages, the Phenol-Explorer Database (PED) version 3.6 (http://phenol-explorer.eu/) was used. The PED is an electronic database that compiles polyphenol contents of over 400 foods and beverages retrieved from articles published in peer-reviewed journals (Neveu et al., 2010).

Five analytical methods are used to determine the polyphenol composition of foods: chromatography, chromatography after hydrolysis, Folin assay, normal phase chromatography, and pH differential method (Neveu et al., 2010). For this study, preferential reporting of polyphenol content followed the approach of Huang et al. (2020). All animal-derived foods and some beverages (e.g., soft drinks, sport drinks) were excluded due to trace polyphenol content.

In the 2015 CCHS-N, each food features a unique food code, a description, and nutrient information per gram of food. Foods are further categorized into a food group system (e.g., Fowl) which classifies basic foods and complex recipes and allows researchers to categorize dietary information in the survey. Complex foods were broken down into individual components. Using standard recipes, ingredient weights were determined so the proportion of polyphenol-containing ingredients could be calculated from the total (e.g., the proportion of flour and chocolate in chocolate chip cookies). Items that directly matched foods and beverages in the PED were completed first. Values were extrapolated whenever possible for the remaining items. Foods and beverages with less than 2% weighted frequency of consumption and all dietary supplements (Huang et al., 2020) were eliminated from the analyses due to negligible polyphenol contribution.

Once polyphenol values were assigned, retention factors were applied to account for losses due to cooking practices (Rothwell et al., 2015). In addition, a correction factor for moisture content derived from the Canadian Nutrient File was applied to dried fruits and vegetables and instant tea and coffee (Huang et al., 2020). For all foods and beverages, total polyphenol, main class (flavonoids, phenolic acids, stilbenes, lignans, and other polyphenols), and subclass (flavanols, flavonols, flavonones, flavones, anthocyanins, isoflavonoids, hydroxybenzoic acids, and hydroxycinnamic acids) contents were calculated by multiplying the gram amount of each food or beverage item consumed with the polyphenol content (in milligrams (mg)) in one gram of that item. Respondents’ daily intakes of total polyphenols, main classes, and subclasses were calculated by summing the respective polyphenol content of all foods or beverages consumed each day (mg/day). Polyphenol intakes were adjusted by daily energy intake in kilocalories (kcal) (mg/1000 kcal/day) to account for variations in age and sex.

Dietary contributors to polyphenol intake

To determine the major food group and individual food contributors to total polyphenol, main class, and subclass intake, a total of 10 food groups were developed (Supplementary Table 1), adapted from the 7 food groups presented by Huang et al. (2020). Food groups included fruits and fruit juice (whole, dried, jams, and 100% juices); vegetable and vegetable juices (whole, dried, and 100% juices); non-alcoholic beverages (coffee and tea); alcoholic beverages (beer, wine, cider, and spirits); cocoa and cocoa products; non-whole-grain products; whole-grain products; legumes and soy products; nuts and seeds (whole and butters); and oils, vinegars, and spices.

Estimation of usual total polyphenol intake

The National Cancer Institute (NCI) method was used to determine usual intake of total polyphenols (Health Canada, 2017). While data from a single dietary recall provide an estimate of daily or ‘given-day’ intake of a nutrient, the NCI method uses data from repeat, non-consecutive dietary recalls to estimate usual or long-term average daily intake, accounting for intra-individual variation in intakes. Detailed instruction of the NCI method is described elsewhere (Herrick et al., 2018; National Cancer Institute, 2018). Decisions regarding model choice, stratification versus pooling, covariate inclusion, and exclusion of outliers were made based on previous work (Davis et al., 2019; Health Canada, 2021).

To estimate usual total polyphenol intake (both energy-adjusted and unadjusted) among Canadians, a stratified approach was employed by analyzing each age/sex group separately. Stratification is preferred over pooling to generate the most unbiased estimates (Davis et al., 2019). A pooled approach was then used to explore usual intake of total polyphenols across sociodemographic and lifestyle factors within children (2–18 years) and adults (≥ 19 years), as the stratified analysis revealed similar total polyphenol intakes and similar within-between variance ratios within child and adult age/sex groups (Davis et al., 2019).

Covariates

Covariates were chosen to align with previous work (Davis et al., 2019; Huang et al., 2020; National Cancer Institute, 2018). Covariates used in all models, along with age and sex, included body mass index (BMI); season of interview; household education level; ethnicity; coffee consumption; tea consumption; province; weekend effect (Friday–Sunday); and dietary recall number. Standard Dietary Reference Intakes categories were collapsed into 13 age/sex groups and used for stratification and covariate adjustment. BMI was classified as underweight, normal weight, overweight, obese, or having missing information (valid skip/not stated) according to age-specific cut points (Health Canada, 2017). Self-reported BMI was used where measured BMI was unavailable. Household education level was derived by asking all household members about their educational attainment, categorizing responses into five groups (‘Less than high school diploma’, ‘High school diploma’, ‘Trade or college diploma’, ‘University diploma or Bachelor’s degree’, or ‘University diploma above Bachelor’s degree’), and reporting the highest value within the household. Ethnicity was grouped across nine levels as per Public Health Ontario (2021). Respondents were classified as coffee or tea consumers depending on consumption indicated in the recall.

Statistical analyses

All statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). Values are represented in proportions (95% CI) and means (95% CI) where required. Energy-adjusted total polyphenol, main class, and subclass intake values are presented in the main text while unadjusted values can be found in Supplementary Tables 2, 3, and 4.

SAS SURVEY commands were used to account for the complexity of the survey designs and, as recommended by Statistics Canada, sample weights and bootstrap (n = 500) weights were applied to obtain population-level estimates and to calculate confidence intervals, respectively. A population proportion was determined for each demographic characteristic. Arithmetic mean intakes of total polyphenols, main classes, and subclasses were calculated using the first dietary recall only to generate given-day intakes. The mean proportion of total polyphenols contributed by each main class was determined by summing the intakes of individual polyphenol compounds (determined using chromatography) to derive a sum total polyphenol intake for each respondent’s first dietary recall. As in the study by Ziauddeen et al. (2019), this was required due to limitations of the PED regarding estimations of total polyphenol content using Folin assay versus individual polyphenol compounds using chromatography. Intakes and main class contributions were determined for the entire analytical sample and by age/sex groups.

To determine the contribution of individual foods and food groups to total polyphenol, main class, and subclass intakes, population proportions were generated using a method developed by Block et al. (1985) which calculates the contribution of each food and food group to the population’s total consumption of that nutrient by using the weighted sum of said nutrient in the amount of all foods and beverages reported on the first dietary recall.

To determine significant differences in usual total polyphenol intake between age/sex groups and between groups within sociodemographic and lifestyle variables, a confidence overlapping interval technique was used (Vatanparast et al., 2020). This technique was also used to explore differences in given-day contributions of main classes to sum total polyphenol intakes by age/sex group based on mean proportions.

Results

Demographic characteristics

A description of the analytical sample (n = 19,409) can be found in Table 1.

Table 1 Demographic characteristics of the analytical sample from the 2015 Canadian Community Health Survey–Nutrition (n = 19,409)
Full size table

Usual total polyphenol intake across age/sex groups

Table 2 displays the usual intake of total polyphenols by age/sex group. Intake increased with age and minimal sex differences were observed except for females 19–30 and  ≥ 71 years who had significantly greater intakes of polyphenols (899.2 mg/1000 kcal/day (95% CI: 808.3, 990.1) and 1302.2 mg/1000 kcal/day (95% CI: 1235.7, 1368.7)) than males of the same age (673.0 mg/1000 kcal/day (95% CI: 593.2, 752.7) and 1153.0 mg/1000 kcal/day (95% CI: 1083.7, 1222.3)).

Table 2 Usual intake of energy-adjusted total polyphenols (mg/1000 kcal/day) by age and sex for Canadians from the 2015 Canadian Community Health Survey–Nutrition (n = 19,409)
Full size table

Usual total polyphenol intake across sociodemographic and lifestyle factors

Adults

Usual intake of total polyphenols for adults was 1119.3 mg/1000 kcal/day (95% CI: 1090.5, 1148.1) (Table 3). Total polyphenol intakes yielded no significant differences by BMI category. Households with a university diploma above a bachelor’s degree had significantly higher total polyphenol intake than those with a high school diploma. Significant differences by ethnicity were found such that those of white ethnicity consumed significantly greater amounts than those of Black, East Asian, South Asian, and Southeast Asian ethnicities. Canadian adults who completed their dietary recall in winter consumed significantly higher amounts of total polyphenols than those interviewed in summer. Coffee consumers had a usual total polyphenol intake significantly greater than non-consumers. Similarly, those who consumed tea had significantly greater intakes than non-consumers.

Table 3 Usual intake of energy-adjusted total polyphenols (mg/1000 kcal/day) by sociodemographic and lifestyle factors for Canadian adults (≥ 19 years) from the 2015 Canadian Community Health Survey–Nutrition (n = 15,682)
Full size table

Children

Usual intake of total polyphenols for children was 473.0 mg/1000 kcal/day (95% CI: 454.9, 491.0) (Table 4). Children with a normal weight had significantly greater intakes than those who were underweight. Similar to the case for adults, children from households that had attained a university diploma above a bachelor’s degree had significantly higher intakes. Children of Middle Eastern ethnicity had significantly higher intakes than all ethnicities except Black, Latinx, or other, while white children had significantly higher intakes than children who were South Asian or Southeast Asian. In contrast to adults, there were no significant differences in total polyphenol intake by season of interview. Children who consumed coffee or tea had significantly greater intakes than non-consumers.

Table 4 Usual intake of energy-adjusted total polyphenols (mg/1000 kcal/day) by sociodemographic and lifestyle factors for Canadian children (2–18 years) from the 2015 Canadian Community Health Survey–Nutrition (n = 3727)
Full size table

Given-day mean total polyphenol, main class, and subclass intake

Supplementary Table 5 displays the given-day arithmetic mean intakes of total polyphenols, main classes, and subclasses for the entire Canadian population and by age/sex group. For all Canadians, given-day intake of total polyphenols was 1001.9 mg/1000 kcal/day (95% CI: 977.5, 1026.4). Mean given-day intakes of main classes were as follows: flavonoids 197.7 mg/1000 kcal/day (95% CI: 191.2, 204.1); phenolic acids 428.1 mg/1000 kcal/day (95% CI: 410.4, 445.9); stilbenes 0.4 mg/1000 kcal/day (95% CI: 0.4, 0.5); lignans 14.6 mg/1000 kcal/day (95% CI: 13.5, 15.6); other polyphenols 27.5 mg/1000 kcal/day (95% CI: 20.0, 35.0).

Given-day contributions of main classes to sum total polyphenol intakes

The contribution of each main class to given-day sum total polyphenol intakes based on mean proportions is found in Fig. 2 and Table 5. For all Canadians, flavonoids contributed 40.0% (95% CI: 39.3, 40.8), phenolic acids 49.8% (95% CI: 49.0, 50.7), stilbenes 0.1% (95% CI: 0.1, 0.1), lignans 3.7% (95% CI: 3.5, 3.8), and other polyphenols 6.4% (95% CI: 6.1, 6.7). Children obtained significantly more of their total polyphenols from flavonoids and other polyphenols while adults obtained significantly more from phenolic acids. Sex differences were identified such that females 31–70 years consumed a greater percentage of flavonoids and males 31–70 years a greater percentage of phenolic acids.

Fig. 2
figure 2

Contribution of each main class of phenolic compounds to sum total polyphenol intakes for all ages and by age/sex groups on the first dietary recall. Values are percentage based on mean proportions. M, males; F, females; Flavonoids, black with white dots; Phenolic acids, solid light grey; Stilbenes, solid black; Lignans, white with black dots; Other polyphenols, solid dark grey

Full size image
Table 5 (Supporting documentation for Fig. 2). Contribution of each main class of phenolic compounds to sum total polyphenol intakes by age and sex for Canadians based on given-day intakes from the 2015 Canadian Community Health Survey–Nutrition
Full size table

Given-day contributions of individual foods and food groups to polyphenol intake

Figure 3 and Table 6 report the contribution of each food group and individual food to total polyphenol intake based on population proportions. The main contributing food groups for the whole population were non-alcoholic beverages (35.6%); fruits and fruit juice (24.8%); vegetables and vegetable juice (10.8%); whole-grain products (10.2%); and cocoa products (5.9%). The top 5 main food contributors to total polyphenol intake were coffee (26.8%), tea (8.9%), fruit juice (4.2%), banana (4.1%), and apple (3.8%).

Fig. 3
figure 3

Contribution of each food group to the total polyphenol intake for a all ages, b female adults 31–50 years, and c male children 9–13 years on the first dietary recall. Values are percentage based on population proportions. These two age/sex groups were chosen to graphically illustrate the transition from childhood to adulthood in food group sources of total polyphenols

Full size image
Table 6 (Supporting documentation for Fig. 3). Top 5 individual food contributors and contribution of each food group to total polyphenol intake by age and sex for Canadians based on given-day intakes from the 2015 Canadian Community Health Survey–Nutrition
Full size table

For adults, non-alcoholic beverages were the greatest contributing food group to total polyphenol intake (range: 37.5–52.0%), followed by fruits and fruit juices, vegetables and vegetable juice, whole-grain products, and cocoa products. Coffee emerged as the top food contributor (range: 28.8–42.9%) for adults.

For children, fruits and fruit juice were the greatest contributing food group to total polyphenol intake (range: 27.7–49.3%), followed by whole-grain products, vegetables and vegetable juice, cocoa products, and non-whole-grain products. The main food contributor for those 2–13 years was fruit juice (range: 7.68–8.65%). Among older children 14–18 years, coffee (8.21%) for males and tea (8.94%) for females emerged as the top food contributors.

The contribution of each food group and individual food to main class and subclass intake for the entire Canadian population can be found in Supplementary Table 6. The main food group contributors to the main classes were as follows: fruits and fruit juices to flavonoids, non-alcoholic beverages to phenolic acids, fruits and fruit juices to stilbenes, vegetables and vegetable juices to lignans, and whole-grain products to other polyphenols.

Discussion

To our knowledge, this is the first comprehensive analysis of usual total polyphenol intake in a nationally representative sample of Canadians. Due to the lack of Canadian data, comparison to other countries is warranted; however, as noted by Huang et al. (2020), this can be challenging as studies use different dietary assessment methods and include different age ranges. Compared with studies using similar recall methods, adult Canadians’ total polyphenol intakes were higher (1119 mg/1000 kcal/day) when compared to adult populations in the United States (884 mg/1000 kcal/day), the United Kingdom (625–1021 mg/1000 kcal/day), and Brazil (205–225 mg/1000 kcal/day) (Huang et al., 2020; Ziauddeen et al., 2019; Carnauba et al., 2021). Furthermore, and similar to the case for the US, UK, and Brazil populations, energy-adjusted total polyphenol intakes of Canadians were higher in those who were white, older, female, or had higher educational attainment.

A closer look at the current dietary patterns of Canadian adults may reveal why differences in these sociodemographic subpopulations occur. As observed in US adults, females tend to consume more polyphenol-rich foods compared to males, leading to a higher intake of polyphenols (Huang et al., 2020). CCHS-N data from 2015 revealed that women tend to consume tea and whole fruits in greater quantities than men, while drinking similar quantities of coffee (Health Canada, 2018). The consumption of unsweetened beverages, including coffee and tea, was also highest among Canadians who were older and white (Jones et al., 2019). Finally, vegetarian and vegan dietary patterns rich in plants, and subsequently, polyphenols, are more prevalent among females and households with bachelor’s degrees or higher (Valdes et al., 2020).

For children, comparisons in total polyphenol intake can be made to a cohort of European adolescents (Wisnuwardani et al., 2019) and nationally representative samples in Brazil (Carnauba et al., 2021) and the UK (Ziauddeen et al., 2019). Canadian total polyphenol intakes (473 mg/1000 kcal/day) were higher than those reported by Brazil (173–187 mg/1000 kcal/day) and Europe (160 mg/1000 kcal/day) but were similar to intakes reported in the UK (486–558 mg/1000 kcal/day). Although girls were reported to have greater intakes than boys in Europe and the UK, no differences were observed in the Canadian population. In Canada, total polyphenol intake was higher in those from households of higher educational attainment, which is supported by findings in Canadian adults and from other studies (Carnauba et al., 2021; Wisnuwardani et al., 2019; Ziauddeen et al., 2019). Finally, Canadian children of Middle Eastern ethnicity were found to have the highest polyphenol intakes compared to other ethnicities.

Observed differences in children’s total polyphenol intake within sociodemographic subgroups may also be explained by reviewing the current dietary patterns of Canadian children. No intake differences were observed by sex, which is supported by 2015 CCHS-N data revealing that similar proportions of male and female children (9–13 years) consume fruit juice, whole fruits, and grain products (Health Canada, 2018). Among older children (14–18 years), more males consume fruit juice while more females consume tea (Health Canada, 2018), resulting in similar polyphenol intakes albeit from different sources. While household education status has not been shown to consistently impact fruit juice consumption among Canadian children (Warren et al. 2022), lower household education status is associated with higher consumption of unhealthy foods and lower consumption of healthy foods, including those rich in polyphenols, such as whole grains, fruits, and vegetables (Manyanga et al., 2017). Finally, higher polyphenol intakes among Canadian children of Middle Eastern ethnicity may be related to a greater proportion of these individuals following dietary patterns rich in fruits, vegetables, and whole grains (Keshavarz, 2022).

Differences between adults and children were observed with respect to food and beverage contributors to total, main class, and subclass polyphenol intake. While beverages were the greatest contributor to total polyphenol intake for all Canadians, for children, this beverage was fruit juice and for adults, coffee. This is consistent with children consuming significantly greater mean volumes of fruit juice and lower mean volumes of unsweetened coffee and tea than adults (Jones et al., 2019). Children also consumed a greater proportion of total polyphenols from flavonoids than adults, which is consistent with data showing children consuming more whole fruit, fruit juice, and chocolate (Health Canada, 2018), all of which are rich sources of flavonoids. Children obtained a greater percentage of polyphenols from both non-whole and whole-grain products, which again is unsurprising as they tend to consume greater average daily servings of grain products than adults (Tugault-Lafleur & Black, 2019) and explains why children have a higher contribution of ‘other polyphenols’ to total intake. Coffee emerged as the main beverage contributor to total polyphenol, phenolic acid, and hydroxycinnamic acid intake among Canadian adults, and likely explains why we observed higher phenolic acid intakes in Canadian adults as compared with children.

Strengths and limitations of the study

Strengths of the present study include the use of dietary recall data from a nationally representative sample of Canadians and the implementation of the NCI method which permitted an estimation of usual long-term dietary intake. Limitations of the analysis included bias related to misreporting on 24-h dietary recalls. While unavoidable, this was managed by removing unreliable reporters and by adjusting polyphenol intake by energy intake. Errors may have occurred in classifying food and beverages codes into food groups due to reliance on 2015 CCHS-N food descriptions, and in matching PED-derived foods and beverages to those reported in the survey due to limited detail on specific foods in the PED. To minimize these errors, two researchers independently matched food and beverage items to the PED and met to reach consensus on matching. Finally, a limitation of the PED is the incompleteness of data presented for each polyphenol class for any given food and the various analytical methods used to assess the polyphenol content of foods. These limitations resulted in discrepancies between estimates of total polyphenol content when determined using Folin assay versus when summing individual polyphenol class intakes determined by chromatography.

Conclusion

Usual intake of total polyphenols and percentage contributions of main classes and food groups to polyphenol intake differs primarily by age among Canadians. Non-alcoholic beverages, particularly coffee, tea, and fruit juice, are top contributors to total polyphenol intake across all ages. Further research is needed to explore associations between polyphenol intake and disease rates in the Canadian population, to help elucidate a beneficial level of intake for risk reduction.

Contributions to knowledge

What does this study add to existing knowledge?

  • This study is the first to estimate total polyphenol, main class, and subclass intake among Canadians using a nationally representative sample and to explore the potential sociodemographic and lifestyle factors that may impact intake.

  • It supports data highlighting better diet quality among those of higher educational attainment, particularly greater consumption of plant-based, polyphenol-rich foods.

  • It has revealed that non-alcoholic beverages, particularly coffee, tea, and fruit juices, are main contributors to total polyphenol intakes.

  • It highlights a shift in dietary patterns from childhood to adulthood and the resulting change in intake of various polyphenol main classes and subclasses.

What are the key implications for public health interventions, practice or policy?

  • Public health messaging has promoted water and unsweetened tea and coffee as healthy beverages for Canadians while recommending limited consumption of alcoholic beverages. These recommendations support current polyphenol intakes in that most polyphenols are consumed from tea and coffee with minor contributions from alcoholic beverages. An exception to current recommendations may include fruit juice consumption in children. The reduction in sugar intakes may be warranted in some groups; however, it is necessary to acknowledge the potential reduction in polyphenol intake that may occur within this subpopulation.

  • Evidently, greater understanding of the role polyphenols play in the Canadian diet is warranted.