Foods have many components other than the commonly known nutrients such as proteins, fats, carbohydrates, vitamins and minerals, many of which are associated with biological activities related to the reduced risk of several chronic diseases and cancer( Reference Craig 1 – Reference Hung 3 ). Flavonoids possess a wide range of biochemical and pharmacological effects, including anti-inflammatory and antioxidant effects( Reference Beecher 4 – Reference Prior and Cao 7 ). In addition, recent studies have suggested that flavonoid intake may contribute to body weight maintenance in the general female population( Reference Hughes, Arts and Ambergen 8 ). Flavonol and flavone contents are the highest in plant-based foods, and are concentrated mainly in leafy vegetables and some fruits.
Although flavonoids have various biological activities, their biological function depends mainly on their intakes and bioavailabilities. Databases of flavonoids in Chinese plant foods are currently limited and few data on the estimated flavonoid intake of the Chinese population have been published. Cao et al.( Reference Cao, Chen and Zhang 9 ) reported the major flavonols (quercetin, kaempferol, myricetin and isorhamntin) and flavones (apigenin and luteolin) found in 100 types of vegetables and fruits consumed by the Chinese population. Based on their findings, a preliminary database of flavonols and flavones was built.
Several international studies have reported flavonoid intakes of adults and a few studies have focused on the intake of minor populations, e.g. among the elderly and women( Reference Chun, Chung and Song 10 – Reference Otaki, Kimira and Katsumata 12 ), but few have focused on adolescent girls. Adolescence is an important period of growth and development. Given the special physiological and psychological factors that define adolescence among girls, as well as the phenomenon of blind weight loss, their health has received increasing attention among researchers in recent years( Reference Johnson, Wardle and Griffith 13 ). Suihua is located in northern China, where the economic condition is not as prosperous as in the southern parts of the country. Seasonal vegetables and fruits are still much less eaten, especially during winter. The flavonol and flavone intakes of female adolescents in this area remain unclear. Based on available data on flavonol and flavone contents, the present study was performed to estimate the intakes of two subgroups of flavonoids (flavonols and flavones) and to explain their major dietary sources among female adolescents in the Suihua area in northern China using an FFQ. To demonstrate the feasibility of the FFQ on the target population, the reproducibility and validity of the FFQ were also evaluated. This study would facilitate further investigation on the relationship between flavonoids and chronic diseases.
Methods
Ethics statement
The study was approved by the Research Ethics Committee of Harbin Medical University. Written informed consent was obtained from either the participants or their parents (for participants below 18 years old) before the participants were enrolled in the study.
Study population
Female adolescents aged 12–18 years were enrolled in the study using a stratified multistage cluster sampling method. The participants were randomly recruited from three middle schools in the Suihua area of Heilongjiang Province, China. Each school had six grades. A total of twenty-five resident (living in the countryside) and twenty-five non-resident (living in the city) female students were selected from each grade. A total of 887 adolescents from the 900 potential participants joined in the study, excluding those who did not satisfactorily complete the FFQ (n 13). Among the qualified participants, 120 girls were expected to participate in the 24 h dietary recalls (24-HR) and answer the FFQ again after one year.
Study design
The study began in March 2011 and continued into the following year. Four three-day 24-HR were quarterly collected from 120 participants, who also completed the second FFQ (FFQ2). The first FFQ (FFQ1, n 887) was administered during the first three-day 24-HR, whereas the FFQ2 (n 120) was administered in February 2012. The reproducibility of the FFQ was estimated using intra-class correlation coefficient (ICC) analysis between FFQ1 and FFQ2, whereas its validity was assessed by comparing FFQ1 with the average of the four three-day 24-HR using Spearman’s correlation coefficient.
FFQ and 24 h dietary recalls
The FFQ was developed according to the methodology proposed by Willett( Reference Willett 14 ). Revisions were made to ensure that the list of foods reflected the Chinese diet, according to the 2002 National Health and Dietary Survey( Reference Li, Rao and Kong 15 ). Further revisions were made to improve the estimates of flavonoid-rich food intakes. The FFQ collected consumption information across eight food groups and eighty-six food items (Table 1). Respondents were requested to recall the average consumption of a given amount of each food item during the previous 1-year period using a graded scale with seven levels, ranging from never to ≥3 times per day. The portion size was also considered for the survey. For the 24-HR, each participant was asked to choose from a set of colour photographs showing different-sized portions of twenty-three specific food items, whereas photographs of dishes, bowls and cups were used to represent the average portion size for the other sixty-three food items. Participants had to describe qualitatively and quantitatively all food consumed during the previous day using the correct pictures to allow the dietitian to estimate their food intakes. All FFQ data were double entered into a computer and any discrepancies were resolved by referring to the original forms. Two dietitians visited each participant four times to complete the three-day 24-HR and these data were used as a standard to measure the relative validity of the FFQ. The dietitians were instructed to ask the participants key questions for a better qualitative description of food items (e.g. green leafy vegetables and fruits).
Food grouping used in the present study
* Assigned portion sizes for all eighty-six food items used in the FFQ.
Intakes of dietary nutrients, flavonols and flavones
The nutrient calculator software Fei Hua V2·3 (Institute for Nutrition and Food Security, Chinese Center for Disease Control and Prevention, Beijing, China) was used to calculate the daily intakes of energy and nutrients. The database of flavonols and flavones established by our laboratory was used to calculate the daily flavonol and flavone intakes in our study.
Statistical analysis
Means and standard deviations of daily energy, flavonol and flavone intakes were calculated for the two FFQ and the average of the four three-day 24-HR, and analysed using the Wilcoxon signed-rank test. The residual method was used to help exclude the possibility of variations caused by energy intake( Reference Willett 14 ). Reproducibility between the two FFQ administrations was estimated using ICC( Reference Willett 14 ). The validity of the FFQ relative to the 24-HR was assessed using Spearman’s correlation coefficient. Considering within-person variations caused by day-to-day fluctuations and seasonal variations, Spearman’s correlation coefficients were ‘de-attenuated’ using the within- and between-person components of variation from the 24-HR( Reference Rosner and Willett 16 ). All statistical analyses were performed using the SPSS statistical software package version 13·0 and Microsoft® Excel 2003. Unless stated otherwise, P<0·05 was considered significant.
Results
The mean age and BMI of the participants (n 887) were 16·1 (sd 1·3) years and 23·9 (sd 3·5) kg/m2, respectively. The mean intakes of energy, protein, fat, carbohydrate, eight food groups, four flavonols and two flavones of 120 adolescent girls, who completed FFQ1 and four three-day 24-HR from four seasons and FFQ2 for over 1 year, were calculated using the questionnaires as bases, with the aid of the nutrient calculator software and flavonoid database. The aforementioned data are shown in Table 2. The intake of beans from FFQ2 and the intake of drinks from FFQ1 were significantly higher than those from the 24-HR (P<0·05), and the intake of drinks from FFQ2 was lower than that from FFQ1 (P<0·05). No differences were observed for the intakes of energy, protein, fat, carbohydrate and other food groups derived from the 24-HR, FFQ1 and FFQ2. The mean daily flavonoid intake derived from the 24-HR, FFQ1 and FFQ2 was 18·51 (sd 9·98) mg/d, 20·21 (sd 10·12) mg/d and 20·24 (sd 11·18) mg/d, respectively. No significant differences were detected for the mean total flavonoid intake (flavonols+flavones) and the individual flavonol and flavone intakes.
Daily mean intakes of energy, macronutrients, food groups, flavonols and flavones of female adolescents aged 12–18 years from the Suihua area of northern China (n 120), estimated using two FFQ and the average of four three-day 24 h dietary recalls (24-HR), March 2011–February 2012
* Significantly different from 24-HR, Wilcoxon signed-rank test, P<0·05.
† Significantly different from FFQ1, Wilcoxon signed-rank test, P<0·05.
Reproducibility and relative validity of FFQ
The crude- and energy-adjusted ICC for FFQ1 v. FFQ2 were calculated to assess the reproducibility of the FFQ (Table 3). All flavonols and flavones were moderately correlated (ICC=0·4–0·7). After adjusting for energy, all ICC decreased. The mean crude ICC was 0·52 from a range of 0·45 (isorhamntin) to 0·65 (quercetin), whereas the mean energy-adjusted ICC was 0·42 from a range of 0·35 (isorhamntin) to 0·50 (quercetin).
Reproducibility and validity of the FFQ designed for evaluation of the daily flavonol and flavone intakes of female adolescents aged 12–18 years from the Suihua area of northern China, March 2011–February 2012
* Intra-class correlation coefficient (ICC) was used to assess the reproducibility between FFQ1 and FFQ2 administrations (n 120).
† Spearman’s correlation analysis was used to assess the validity of FFQ1 relative to the average of four three-day 24 h dietary recalls (24-HR; n 120).
‡ All ICC values were significant (P<0·05).
§ All Spearman’s correlation coefficients were significant (P<0·05).
The crude, energy-adjusted and de-attenuated Spearman’s correlation coefficients for FFQ1 v. the average resulting from the four three-day 24-HR are also presented in Table 3. These values were used to assess the relative validity of the FFQ. The crude Spearman’s correlation coefficient between FFQ1 and 24-HR ranged from 0·53 (apigenin) to 0·77 (quercetin), with a mean value of 0·62. The energy-adjusted coefficient ranged from 0·40 (apigenin) to 0·59 (luteolin), with a mean of 0·48. The de-attenuated coefficients ranged from 0·55 (apigenin) to 0·78 (quercetin), with a mean of 0·64. All energy-adjusted Spearman’s correlation coefficients were lower than the crude ICC in terms of FFQ1 and FFQ2. However, de-attenuation to correct for intra-individual variability improved the Spearman’s correlation coefficients and led to an increase in the mean Spearman’s correlation coefficient for FFQ1 v. 24-HR (from 0·62 to 0·64).
Estimated flavonol and flavone intakes and major food sources
The mean daily dietary flavonol and flavone intakes were estimated from the FFQ (FFQ1, n 887) completed by female adolescents in the Suihua area of northern China. The total flavonoid intake (flavonols+flavones) was 20·60 (sd 14·12) mg/d, mainly from quercetin and kaempferol (26·8 % and 23·7 %, respectively), followed by luteolin (15·9 %), isorhamntin (14·6 %) and myricetin (11·1 %). The contribution of apigenin was relatively minimal (5·0 %). The flavonol intake was 16·29 (sd 11·91) mg/d (79·1 %) and the flavone intake was 4·31 (sd 2·21) mg/d (20·9 %; Table 4).
Daily flavone and flavonol intakes of female adolescents aged 12–18 years from the Suihua area of northern China, March 2011–February 2012
* Data from FFQ1 (n 887).
The primary food items contributing to the individual flavonol and flavone intakes are shown in Table 5. After colligating the daily dietary intakes, the results show that apple was the primary food item contributing to the total flavonol and flavone intakes (11·7 %), followed by potatoes (9·9 %) and lettuce (7·3 %). Other main contributors included oranges (7·0 %), Chinese cabbage (4·7 %), tomatoes (4·2 %), celery (4·2 %), soyabean sprouts (4·2 %), leeks (3·9 %) and aubergine (3·9 %). The sources of the total flavonoid intakes (flavonols+flavones) were vegetables (55·6 %), fruits (26·6 %) and others (17·8 %; e.g. cereals, tea, fruit juice and coffee).
Main food sources and their contribution (%) to the flavone and flavonol intakes of female adolescents aged 12–18 years from the Suihua area of northern China, March 2011–February 2012
* Percentage of individual flavonols or flavones.
† Percentage of total flavonols and flavones.
Discussion
Given that the FFQ method is inexpensive and reflects the long-term dietary conditions of participants, it is generally considered to be a good tool for conducting dietary nutritional epidemiological studies. The results from FFQ differ depending on cultural background, eating habits, food sources, season, climate, age, economy and other factors, even in the same country. Thus, a special FFQ should be designed for the target population while considering the differences in food items that contribute to the daily supply of nutrients or phytochemicals with biological activities( Reference Zhang and Ho 17 ). Suihua is located in north-east China, where winter lasts for more than five months and the economic condition is not as prosperous as that of the southern part of the country. In the current study, a new FFQ was designed for adolescent girls from Suihua, with the specific aim of selecting food groups that contribute significantly to flavonol and flavone intakes.
The reproducibility and relative validity of the FFQ were evaluated. The crude and energy-adjusted ICC for the mean daily flavonol and flavone intakes from FFQ1 v. FFQ2 were 0·52 and 0·42, respectively, for reproducibility. The crude, energy-adjusted and de-attenuated Spearman’s correlation coefficients for the mean daily flavonol and flavone intakes from FFQ1 v. 24-HR were 0·62, 0·48 and 0·64, respectively, for validity, indicating that better correlations were obtained. Our study population consisted of a group of female adolescents with similar lifestyles (i.e. students in three middle schools who frequently dine in the school canteen), which possibly contributed to the moderate correlations. The ability of an individual to deal with abstract concepts and form mental images of her dietary regimen that are as close as possible to the truth( Reference Yaroch and Resnicow 18 ), was another crucial factor in producing reliable estimates of habitual intake using the FFQ. Thus, colour photographs of foods taken in classrooms were used to make the task less tedious and encourage the interviewees to provide more accurate dietary information( Reference Tooze, Midthune and Dodd 19 ). When the 24-HR were used to estimate the individual dietary intakes, we used a method that provided better estimates of the intake of even episodically consumed foods by accounting for the correlation between the probability of consumption and the amount consumed, and incorporating covariate information( Reference Souverein, Dekkers and Geelen 20 , Reference Goodwin, Brule and Junkins 21 ). Adjusting for energy decreased the correlation in almost all subclasses of flavonoids, as high-energy foods may contain little to no amounts of flavonols and flavones( Reference Mullie, Clarys and Deriemaeker 22 ).
Despite numerous reports on the health-improving effects of flavonoids, a limited number of studies on flavonoid intake are available worldwide. The flavonol and flavone intakes from several countries are listed in Table 6, showing that the estimates varied widely among studies. In the current study, the mean flavonoid (four flavonols and two flavones) intake of 887 female adolescents was 20·6 mg/d. The total flavonol intake was 16·3 mg/d and the total flavone intake was 4·3 mg/d. These results are similar to those in studies conducted in Australia( Reference Johannot and Somerset 23 ), the Netherlands( Reference Hertog, Hollman and Katan 24 ), China( Reference Zhang, Li and Cao 25 ) and the USA( Reference Sampson, Rimm and Hollman 26 ), and are even higher than those in other studies reported from the USA( Reference Hertog, Kromhout and Aravanis 27 , Reference Chun, Floegel and Chung 28 ), Finland( Reference Ovaskainen, Torronen and Koponen 11 , Reference Knekt, Kumpulainen and Jarvinen 29 ) and Japan( Reference Kimira, Arai and Shimoi 30 , Reference Arai, Watanabe and Kimira 31 ). However, our results are lower than those found in studies conducted in Belgium( Reference Mullie, Clarys and Deriemaeker 22 ) and Spain( Reference Zamora-Ros, Andres-Lacueva and Lamuela-Raventos 32 ).
Estimated flavonol and flavone intakes in several countries
*M, males; F, females.
The major dietary sources of flavonol and flavone intakes reported in some of the aforementioned countries are onions and tea, followed by apples, broccoli and lettuce( Reference Hertog, Hollman and Katan 24 , Reference Sampson, Rimm and Hollman 26 , Reference Hertog, Feskens and Hollman 33 ). In the present study, apples, potatoes, lettuce, oranges, soyabean sprouts and leeks were the main food sources of flavonols, whereas tomatoes, aubergine, white radishes, celery and sweet potatoes were the main sources of flavones, similar to the results of a study on the Australian population with a sample population aged 16–18 years( Reference Johannot and Somerset 23 ). The participants in our study were adolescent girls with special dietary characteristics and much lower intakes of tea and onions compared with the respective populations of Western countries, leading to differences in both intake amounts and major sources. The contributions of cereals and cereal-containing foods, which the girls consumed in large amounts, to the flavonol and flavone intakes were minimal (7·2 %) in the present study. The differences in results could also be explained by the potato consumption. Our participants living in the Suihua area partly ate potatoes instead of the staple food, e.g. rice, making potato the primary food item that subsequently contributed to the total flavonol and flavone intake.
The current study had several limitations. First, only two subgroups of flavonoids were studied. However, other subclasses of flavonoids, such as isoflavones and isoflavanones, are also consumed in high amounts in China. This limitation could be ascribed to the flavonoid database, which has a limited number of food items, although the database is constantly being updated. Therefore, broader evaluations on other subclasses of flavonoids must be conducted. The second limitation was the time for the FFQ to be completed. The flavonoid contents of foods are influenced by region, season, sunlight and other factors( Reference Hertog, Hollman and Katan 24 ). For instance, the quercetin content of onions measured in our laboratory was 1·23 mg/100 g (fresh weight), which was much lower than that in the US flavonoid database (7·29–33·43 mg/100 g, fresh weight) and the contents reported by Arabbi et al. (38·3–93·6 mg/100 g, fresh weight)( Reference Arabbi, Genovese and Lajolo 34 ). Third, the content changes in the storage and processing of the foods were not considered. As previously mentioned, food storage temperature, processing and peeling can influence flavonoid content( Reference Srivastava, Akoh and Yi 35 ). Thus, more in-depth studies must be performed.
Conclusion
In conclusion, based on the proven reproducibility and validity of the FFQ, the present study estimated the dietary intakes and sources of flavonols and flavones among female adolescents. The dietary flavonoid intakes among female adolescents in the Suihua area were similar to those reported in previous studies, even though winter is longer and the economy is relatively poor in this area. The food sources showed significant differences because of the geographical position and dietary characteristics. This work provides data for epidemiological studies on the proposed relationship between these flavonoids and human health.
Acknowledgements
Financial support: This study was supported by a grant from the Science and Technology Agency of Heilongjiang Province, China (Youth Fund to W.X., QC06C059). The funder, W.X., contributed to the study design, conduct of the study, analysis of samples or data, interpretation of findings or the preparation of the manuscript. Conflict of interest: None. Authorship: W.X. conceived and designed the study, and is responsible for final editing and approval of the manuscript. C.S., H.W., D.Z. and Y.C. performed the investigations. C.S. and Y.Z. analysed the data and wrote the manuscript. All authors approved the final version of the paper for publication. Ethics of human subject participation: This study was approved by the Research Ethics Committee of Harbin Medical University. Written informed consent was obtained from either the participants or their parents before they were enrolled in the study.
