Hostname: page-component-7c8c6479df-ws8qp Total loading time: 0 Render date: 2024-03-28T16:58:45.565Z Has data issue: false hasContentIssue false

Beverage intake patterns of Canadian children and adolescents

Published online by Cambridge University Press:  23 June 2011

Adrienne D Danyliw
Affiliation:
Division of Nutrition and Dietetics, College of Pharmacy and Nutrition, University of Saskatchewan, 110 Science Place, Thorvaldson Building, Saskatoon, Saskatchewan S7N 5C9, Canada
Hassanali Vatanparast
Affiliation:
Division of Nutrition and Dietetics, College of Pharmacy and Nutrition, University of Saskatchewan, 110 Science Place, Thorvaldson Building, Saskatoon, Saskatchewan S7N 5C9, Canada
Nooshin Nikpartow
Affiliation:
Division of Nutrition and Dietetics, College of Pharmacy and Nutrition, University of Saskatchewan, 110 Science Place, Thorvaldson Building, Saskatoon, Saskatchewan S7N 5C9, Canada
Susan J Whiting*
Affiliation:
Division of Nutrition and Dietetics, College of Pharmacy and Nutrition, University of Saskatchewan, 110 Science Place, Thorvaldson Building, Saskatoon, Saskatchewan S7N 5C9, Canada
*
*Corresponding author: Email susan.whiting@usask.ca
Rights & Permissions [Opens in a new window]

Abstract

Objective

Little is known of the beverage intake patterns of Canadian children or of characteristics within these patterns. The objective was to determine beverage intake patterns among Canadian children and compare intakes of fourteen types of beverages, along with intakes of vitamin C and Ca, and sociodemographic factors across clusters.

Design

Dietary information was collected using one 24 h recall. Sociodemographic data were collected by interview. Cluster analysis was used to determine beverage intake patterns. Pearson's χ2 and 95 % CI were used to test differences across clusters.

Setting

Data from the Canadian Community Health Survey Cycle 2·2.

Subjects

Children aged 2–18 years with plausible energy intake and complete sociodemographic data (n 10 038) were grouped into the following categories: 2–5-year-old boys and girls, 6–11-year-old girls, 6–11-year-old boys, 12–18-year-old girls and 12–18-year-old boys.

Results

Five beverage clusters emerged for children aged 2–5 years, six clusters for children aged 6–11 years (both sexes) and four clusters for those aged 12–18 years (both sexes). Sweetened beverage clusters appeared in all age–sex groups. Intakes of sweetened beverages ranged from 553 to 1059 g/d and contributed between 2 % and 18 % of total energy intake. Girls 6–11 years of age in the ‘soft drink’ cluster had lower Ca intake compared with other clusters in that age–sex group. Age and ethnicity differed across clusters for most age–sex groups. Differences for household food security status and income were found; however, no pattern emerged.

Conclusions

Patterns in beverage intake among Canadian children include beverages that are predominantly sugar sweetened. Public health nutrition professionals can use knowledge about beverage patterns among children, as well as the characteristics of these groups, in the development of nutritional programmes and policies.

Type
Research paper
Copyright
Copyright © The Authors 2011

Expert panels have recommended limiting sweetened beverage intake for the general population in the USA and Mexico(Reference Popkin, Armstrong and Bray1, Reference Rivera, Munoz-Hernandez and Rosas-Peralta2). Canada's Food Guide recommends limiting sweetened beverage intake but does not provide details on the amount and frequency of consumption(3). Recommendations to limit sweetened beverages and juice have been made for children as they relate to childhood obesity or overnutrition(Reference Rao4, 5). Despite this, consumption of sweetened beverages has increased over time among American adults and children(Reference Popkin6, Reference Wang, Bleich and Gortmaker7). In Canada, no national trend data on consumption are available; however, data on beverage disappearance suggest that soft drink intake has doubled between 1980 and the 2000s, whereas milk intake has decreased(8). Food disappearance data do not provide consumption patterns of specific groups such as children. Thus, there is a need in Canada to identify and describe beverage consumption patterns, especially of children, in order to develop targeted policies and programmes that aim to reduce sweetened beverage consumption among the highest consumers.

The Canadian Community Health Survey (CCHS) is a national annual cross-sectional survey; however, to date, only one cycle, cycle 2·2 in 2004, has collected dietary data. This is the most recent national-level nutritional survey available in Canada in over 30 years. The objective of the present study was to use CCHS 2·2 to identify beverage patterns among Canadian children aged 2–18 years and to describe and compare beverage intake and sociodemographic characteristics across the resulting beverage patterns.

Experimental methods

Data from CCHS cycle 2·2 included dietary information collected by means of a 24 h recall using the Automated Multiple Pass Method (AMPM)(Reference Blanton, Moshfegh and Baer9). Data from a single 24 h recall were collected by trained interviewers using computers, following the five steps of the AMPM: quick list; forgotten foods and beverages; time and occasion; detailed information including amounts consumed and preparation method; and a final review(10). Participants could report the amount of beverages consumed as volume, container type, or choice of standard images of various sizes of glasses and mugs. Children aged ≥12 years provided their own dietary recall data, whereas children aged 6–11 years provided dietary information along with their parent or caregiver. Children under 6 years of age had a parent or caregiver to report their food intake(10). Weighting adjustments were applied to reduce seasonal and weekday variations in dietary recall information, where the proportion of those interviewed each season corresponded to one-quarter of the total population and the proportion interviewed on a weekday (Monday through Thursday) corresponded to four-sevenths of the population(11).

CCHS 2·2 used interviews to gather data on demographics and at socio-economic levels. Ethnicity was self-reported and, because of small cell size, collapsed into two categories: white and non-white. Household income was based on self-reported income before taxes and on the number of people per household. Because of small cell sizes, income has been reported as low and high. Low household income was considered ≤$CAN 15 000 if one to two people lived in the household, ≤$CAN 20 000 if three to four people lived in the household and ≤$CAN 30 000 if five or more people lived in the household. High income was the reverse of these categories. Household food security was based on responses to the Household Food Security Survey Module(10). Food insecure included both moderate and severe levels. Household education was the highest level of education attained by any member of the household. Individuals with a low level of education were defined as those who had obtained secondary-school graduation or lower. Individuals with high level of education were those who had attained at least some amount of post-secondary school education. Additional details on CCHS 2·2 can be found elsewhere(12).

The CCHS 2·2 comprised a final sample of 35 107 participants of all age groups. We used data from children aged 2–18 years (n 13 824) and used the age categories of 2–5, 6–11 and 12–18 years to compare with other studies on beverage consumption(Reference Popkin6, Reference LaRowe, Moeller and Adams13, Reference O'Connor, Yang and Nicklas14). The inclusion criteria to arrive at a sample of 10 038 were: valid 24 h recall data (as determined during the interview); complete data for household income, household education, household food security and ethnicity; sex-specific BMI-for-age that was greater than the 5th percentile, as in the study by LaRowe et al.(Reference LaRowe, Moeller and Adams13); feeding not only breast milk (for children aged 2–5 years); not pregnant or breast-feeding (for girls aged 12–18 years); and plausible energy intakes. Implausible energy intakes were either <3374 kJ (<800 kcal)/d or >11 297 kJ (>2700 kcal)/d(11) for children aged 2–5 years; <3766 kJ (<900 kcal)/d or >14 644 kJ (>3500 kcal)/d for children aged 6–11 years(Reference LaRowe, Moeller and Adams13); and <5021 kJ (<1200 kcal)/d or >20 502 kJ (>4900 kcal)/d for children aged 12–18 years, which is equal to 10 % less than the lowest calculated estimated energy requirement (EER) and 10 % more than the highest EER for that age group, rounded to the nearest 100 units.

Beverage pattern formation

Beverages consumed were identified and extracted from the CCHS 2·2 data files. These were categorized on the basis of energy and nutrient content (Table 1) and on Canadian recommendations, in which, for example, skimmed 1 % milk fat (MF) and 2 % MF milk are all considered low-fat milk(3). Beverage weight in grams was used in all analyses.

Table 1 Classification by category and type for beverages

Cluster analysis was used to classify children into clearly distinct groups on the basis of the dominant pattern of beverage intake. K-means cluster analysis was performed for each age–sex group using categories of beverages as described in Table 1. The FASTCLUS procedure categorized participants into non-overlapping groups in an iterative process by comparing Euclidean distances(15). As the procedure is sensitive to outliers, participants were removed if the intake of beverage group was ≥5 sd from the mean intake of the specific beverage category. These participants were returned to the data set once the optimal number of clusters was determined. With the K-means method, it was necessary to predefine the number of clusters. This procedure was commenced with forty clusters, and clusters with five or fewer members were temporarily removed.

The optimal number of clusters was determined using the cubic clustering criterion (CCC), pseudo-F statistic (PFS) and interpretability. The values of CCC and PFS for each cluster set (three to twenty cluster sets) were recorded and then assessed using a line graph to identify peaks in the values. The cluster set(s) at which CCC and PFS both peaked was (were) explored for a clear and unique pattern of intake for each cluster, and the best cluster solution was then determined.

Before finalizing the clusters for each category, an exploratory analysis was completed examining the impact of combining sweetened beverage categories, as well as the impact of water, on the clustering process. Combining sweetened beverage categories did not result in improved clustering. When water was included in the cluster analysis (fourteen beverage categories) as a separate beverage, the CCC and PFS did not produce early and positive peaks, as necessary for interpretation. A similar exploratory analysis was performed for all age–sex groups, and late and multiple peaks for CCC and PFS occurred repeatedly when water was included as a unique beverage group (data not shown). The decision was thus made to exclude water as a separate beverage category from the clustering process (resulting in thirteen beverage categories used for clustering) and to describe the intake of water within all clusters that emerged.

Statistical analysis

Means, sem and 95 % CI were calculated for beverage intake for each cluster across the age–sex groups. Data on percentage of energy from beverages, percentage of energy from sweetened beverages and percentage of energy from nutrient-based beverages, as well as vitamin C and Ca intakes – as they are key nutrients in beverages – by cluster were also examined. Sociodemographic characteristics were compared across clusters. To test differences in categorical variables across beverage patterns, the χ 2 test was used. For continuous variables, 95 % CI were examined and the results were considered to be statistically significant when no overlap existed(10).

All analyses were weighted to obtain estimates at a population level. Bootstrapping was used, as recommended by Statistics Canada, to account for the complex survey design(12). Alpha was set at 0·05. Statistical software packages used in various stages of analyses included SAS version 9·2 (SAS Institute Inc., Cary, NC, USA), STATA versions 10·0 and 11·0 (StataCorp LP, College Station, TX, USA), SPSS version 16·0 (SPSS Inc., Chicago, IL, USA) and Microsoft Excel 2007 (Microsoft Corporation, Redmond, WA, USA).

Results

Cluster analysis was completed separately for five age–sex groups: 2–5-year-olds; 6–11-year-old boys and girls; and 12–18-year-old boys and girls. For children aged 2–5 years, a five-cluster solution was determined to be the most appropriate. The most appropriate clustering patterns for older children were a six-cluster solution for both boys and girls aged 6–11 years and a four-cluster solution for boys and girls aged 12–18 years. The beverage clusters were named using the predominant beverage. All age–sex groups contained a cluster in which no dominant pattern of beverage intake existed, wherein these groups were classified as ‘moderate’ clusters. Intakes of beverages such as vegetable juice, diet drinks, tea and coffee were so low as to not contribute significantly to intake.

Beverage intakes by cluster

Among 2–5-year-olds within each beverage cluster, the defining beverage intake was the highest intake (Table 2). Total beverage intake was lowest among those in the ‘moderate’ cluster. Boys aged 6–11 years who drank mostly fruit drinks drank a relatively small amount of fruit juice (Table 3); however, boys who drank mostly fruit juice drank less fruit drink. Boys in the ‘milk’ cluster drank 859 g of milk, whereas boys in other clusters drank <250 g of milk (Table 3). There was no significant difference in water intake across clusters. Girls aged 6–11 years who drank mostly fruit drinks drank less fruit juice (Table 3). Similar to boys, girls who drank mostly fruit juice drank relatively less fruit drink. Girls who were in the ‘soft drink’ cluster drank 467 g of soft drinks on the previous day. Girls in the ‘milk’ cluster drank 722 g of milk, whereas consumption of milk among other clusters was <200 g (Table 3). Those who were in the moderate cluster had the lowest total beverage intake compared with all other clusters. There was no significant difference across clusters for water intake.

Table 2 Mean intake of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in children aged 2–5 years

%E, percentage of energy; SSB, sugar-sweetened beverages.

Mean values within rows with unlike superscript letters were significantly different from the amilk cluster, bmoderate cluster, chigh-fat milk cluster and djuice cluster.

*Significantly different from other clusters.

†Significantly different across beverage patterns using Pearson's χ 2 test (P < 0·05).

Table 3 Mean intake of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in boys and girls aged 6–11 years

%E, percentage of energy; SSB, sugar-sweetened beverages.

Mean values within rows with unlike superscript letters were significantly different from the amoderate cluster; bfruit drink cluster; csoft drink cluster; dall except the soft drink cluster; ehigh-fat milk, milk, moderate and fruit juice clusters; fall except the fruit juice cluster; gall except the fruit drink cluster; and hfruit drink, fruit juice and milk clusters within the same sex.

*Significantly different from all clusters within the same sex.

†Significantly different across clusters using Pearson's χ 2 test among boys (P < 0·05).

‡Significantly older than the milk cluster within the same sex (P < 0·05).

§Significantly younger than all except the fruit juice cluster within the same sex (P < 0·05).

∥Significantly older than all clusters within the same sex (P < 0·05).

Data for beverage intake by cluster of adolescent boys and girls aged 12–18 years (Table 4) showed that adolescent boys in the ‘milk’ cluster drank >1000 g of milk on the day of the recall, which was nearly half of their total beverage intake. Adolescent boys in the ‘soft drink’ cluster drank >1000g of soft drinks, and they drank significantly less fruit juice and water compared with boys in the ‘moderate’ cluster. Adolescent boys in the ‘fruit drink’ cluster drank the same amount of fruit juice as boys in the ‘milk’ cluster. Adolescent girls aged 12–18 years who were in the ‘fruit drink’ cluster drank significantly less milk compared with girls in the ‘moderate’ and ‘milk’ clusters, but not less than those in the ‘soft drink’ cluster. Girls in the ‘moderate’ cluster had the highest intake of fruit juice compared with all other clusters.

Table 4 Mean intakes of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in boys and girls aged 12–18 years

%E, percentage of energy; SSB, sugar-sweetened beverages.

Mean values within rows with unlike superscript letters were significantly different from the amoderate cluster and bmilk cluster within the same sex.

*Significantly different from all other clusters.

†Significantly different across clusters using Pearson's χ 2 test (P < 0·05).

‡Significantly older than all other clusters within the same sex (P < 0·05).

Nutrient and energy intakes by cluster

Within clusters we analysed both energy and selected key nutrients in beverages: percentage of energy from beverages, percentage of energy from sugar-sweetened beverages and percentage of energy from nutrient-based beverages and intakes of Ca and vitamin C.

Children aged 2–5 years consumed between 18 % (‘moderate’ cluster) and 34 % of energy from beverages (‘high-fat milk’ cluster; Table 2). Sugar-sweetened beverages contributed 16 % of total energy intake among those children in the ‘fruit drink’ cluster. Children aged 2–5 years in the ‘milk’ cluster and ‘high-fat milk’ cluster had higher Ca intake compared with all other clusters, although the intakes of the two milk clusters were not significantly different from each other. The ‘fruit juice’ and ‘fruit drink’ clusters had the highest intakes of vitamin C compared with all other clusters. Children in the ‘fruit juice’ cluster had significantly more vitamin C intake compared with those in the ‘fruit drink’ cluster.

Among children aged 6–11 years, percentage of energy from beverages was lowest among those in the ‘moderate’ clusters, at 15 % and 13 %, respectively (Table 3). Those in the ‘high-fat milk’ clusters had the highest percentage of energy from beverages: 28 % for both boys and girls. The percentage of energy from sugar-sweetened beverages was highest among those in the ‘fruit drink’ and ‘soft drink’ clusters. Among children aged 6–11 years, the total Ca intake was highest among boys in the ‘milk’ and ‘high-fat milk’ clusters. Girls in the ‘milk’ cluster had the highest total Ca intake, whereas girls in the ‘soft drink’ cluster has the lowest total Ca intake overall. For both boys and girls, those in the ‘fruit drink’ and ‘fruit juice’ clusters had the highest intakes of total vitamin C compared with other clusters.

The percentage of energy from beverages for adolescent boys aged 12–18 years was 17 % among those in the ‘moderate’ cluster and 29 % among those in the ‘milk’ cluster (Table 4). For girls aged 12–18 years, the percentage of energy from beverages was 17 % among those in the ‘moderate’ cluster and 27 % among those in the ‘fruit drink’ cluster. Percentage of energy from sugar-sweetened beverages was highest among those in the ‘fruit drink’ and ‘soft drink’ clusters (Table 4). Among 12–18-year-old boys, all clusters showed a mean Ca intake from food of >1000 mg/d, although the boys in the ‘milk’ cluster had significantly higher intake of Ca compared with those in other clusters. Boys in the ‘fruit drink’ cluster had significantly higher vitamin C intake compared with boys in other clusters. Adolescent girls aged 12–18 years who were in the ‘fruit drink’ cluster had significantly more vitamin C intake compared with all other clusters. Children in the ‘milk’ cluster had the highest intake of Ca compared with all other clusters.

Sociodemographic characteristics by cluster

Age was significantly different across clusters, except among adolescent girls aged 12–18 years (Tables 24). Among 2–5-year-olds, those in the ‘high-fat milk’ cluster were significantly younger than those in the ‘juice’ cluster, but not younger than those in other clusters. Boys aged 6–11 years who drank mostly high-fat milk were significantly younger than those in other clusters, but not younger than those in the ‘juice’ cluster. Girls aged 6–11 years and boys aged 12–18 years who were in ‘soft drink’ clusters were older than children in all other clusters of the same age–sex group. Boys aged 6–11 years who drank mostly soft drinks were older than boys of the same age who were in the ‘moderate’ cluster.

Ethnicity was significantly different across clusters for all age groups, except among girls aged 6–11 years (Tables 24). Household income was only significantly different across clusters for boys aged 6–11 years (P = 0·037). Food security was significantly different across clusters for adolescent boys (P = 0·038) and girls (P = 0·005). Household education did not differ across clusters for any age–sex group. Data related to physical activity and body weight by cluster will be provided in a separate publication.

Discussion

Cluster analysis is a data-driven method that categorizes individuals into non-overlapping groups. It identifies groups on the basis of food or beverage intake patterns that can be meaningful and interpretable(Reference Moeller, Reedy and Millen16) and goes beyond descriptive analysis of beverage intake. Cluster analysis has been reproduced across populations(Reference Moeller, Reedy and Millen16) and has been reproduced for beverage intake among American children(Reference LaRowe, Moeller and Adams13). We now report clustering of beverage intakes in Canadian children.

Our results showed clustering patterns similar to those reported by LaRowe et al.(Reference LaRowe, Moeller and Adams13) for 2–5- and 6–11-year-olds, despite slightly different beverage categories used to establish clusters. This may indicate reliability of the clustering process to examine beverage patterns among children. Similarly, the results of the present study were comparable to those found by LaRowe et al.(Reference LaRowe, Moeller and Adams13). Specifically, percentage of energy intake from beverages and total beverage intakes were comparable between Canadian and American children.

Other studies have examined beverage intake among children and adolescents. Garriguet(Reference Garriguet17), using CCHS 2·2 data, found that boys and girls aged 1–3 years consumed 28 % and 27 % of their energy from beverages, respectively, whereas boys and girls aged 4–8 years consumed 21 % and 18 % energy from beverages, respectively. In the same study, adolescents aged 9–13 years consumed 18 % of energy from beverages and 14–18-year-old boys and girls consumed 20 % and 19 %, respectively(Reference Garriguet17). Using age categories different from those used by Garriguet(Reference Garriguet17), we also found that 13–34 % of energy was from beverages, with the lowest percentage of energy being found among the ‘moderate’ clusters. The percentage of energy from sugar-sweetened beverages was higher among children in the ‘fruit drink’ and ‘soft drink’ clusters; children in these clusters are expected to consume high amounts of sweetened beverages and lower amounts of other energy-containing beverages.

In the USA, beverage intakes from the National Health and Nutrition Examination Survey (NHANES) 2005–2006 data showed that children aged 2–6, 7–12 and 13–18 years consumed 1075 ml, 1296 ml and 1927 ml, respectively(Reference Popkin6). Cluster analysis groups beverage consumption by type as well as by volume, thus identifying low and high patterns of beverage consumption. Our results show the ranges of total beverages consumed, which include the mean values by age for American children in 2005–2006, possibly indicating similar beverage consumption levels.

In Canada, sweetened beverage intake is a common pattern among children and adolescents aged 2–18 years. Pre-school-aged children in the fruit drink cluster consumed >500 g of fruit drink per day. Among boys aged 12–18 years in the soft drink cluster, a mean intake of >1000 g of soft drink was determined. These values are far above the recommendations set by US expert panels, which recommend limiting sweetened beverages to <225 g for adults(Reference Popkin, Armstrong and Bray1), and hence may pose health risks including obesity(Reference Sanigorski, Bell and Swinburn18Reference Libuda, Alexy and Sichert-Hellert20) and others(Reference Vartanian, Schwartz and Brownell21). Among American children aged 7–12 and 13–18 years in 2005–2006, the second most consumed beverage after water was soda/soft drinks (342 ml and 606 ml, respectively), followed by whole-fat milk (203 ml and 108 ml, respectively) and juice (79 ml and 107 ml, respectively)(Reference Popkin6). Longitudinal data from adolescents in Germany showed carbonated and uncarbonated sweetened beverage consumption to be 371 g for boys and 245 g for girls(Reference Libuda, Alexy and Sichert-Hellert20), levels lower than those in Canada or in the USA. This may be because of differences in study design and beverage grouping, or it may indicate higher sweetened beverage consumption in North America.

High intakes of sweetened beverages have also been associated with low intakes of other nutrients and low overall diet quality(Reference Rodriguez-Artalejo, Garcia and Gorgojo22). The current results show that young girls (aged 6–11 years) who drink mostly soft drinks have a significantly lower intake of Ca compared with all other clusters (661 mg/d). The Dietary Reference Intake values set by the Institute of Medicine include an Estimated Average Requirement (EAR) for Ca at 500 mg for children aged 1–3 years, 800 mg for those aged 4–8 years and 1100 mg for those aged 9–18 years(23). Other researchers have found that young girls who consume soft drinks may be at risk for low Ca intake(Reference Keller, Kirzner and Pietrobelli24) and for low bone mineral content(Reference Libuda, Alexy and Remer25).

Vitamin C is a key nutrient found in fruit juices and is often added to fruit drinks. Our results show that vitamin C intakes are above recommendations for all age–sex groups, regardless of cluster. The EAR for vitamin C is 13 mg for children aged 1–3 years, 22 mg for those aged 4–8 years, 39 mg for those aged 9–13 years, 63 mg for adolescent boys aged 14–18 years and 56 mg for adolescent girls aged 14–18 years(26). Children and adolescents in the ‘fruit drink’ and ‘fruit juice’ clusters had higher intakes of vitamin C compared with those in other clusters within the same age–sex groups. Among all cluster solutions, children and adolescents who consumed mostly fruit drinks had a lower intake of fruit juice and vice versa.

Data from the 2005–2006 NHANES indicated water as the most consumed beverage for 2–6-, 7–12- and 13–18-year-olds at 325 ml, 509 ml and 773 ml, respectively(Reference Popkin6). Similarly, our clustered data show that water was the most or second most consumed beverage within the cluster. Water consumption was also high in beverage clusters for 2–5- and 6–11-year-olds using US NHANES data(Reference LaRowe, Moeller and Adams13). Water consumption is encouraged in Canada as water is an energy-free beverage(3). When developing nutritional policy, supporting and enhancing current water consumption by encouraging replacement of sweetened beverages should be encouraged, as excess consumption is rare in normal circumstances(27).

By clustering data on beverage intake, we were able to examine the sociodemographic characteristics of children within beverage clusters. LaRowe et al. (Reference LaRowe, Moeller and Adams13) also examined selected sociodemographic characteristics across beverage clusters and also found that age, ethnicity and household income differed across clusters for children aged 2–5 and 6–11 years.

The present study has some important limitations. First, it is a cross-sectional study; therefore, we cannot determine temporality. Second, dietary information is presented here on the basis of one 24 h recall, which may result in misclassification. The AMPM is used to collect dietary information and has been reported to be accurate, especially among normal-weight individuals(Reference Moshfegh, Rhodes and Baer28). The AMPM attempts to reduce misclassification in the collection of detailed food and beverage information; however, some participants may incorrectly identify fruit beverage as fruit juice and vice versa. It is important to note that the mean value from 1 d intake is similar to that of the adjusted 2 d intake; only variation is reduced when data from 2 d intake are used and adjusted(10). Finally, cluster analysis is a data-driven method that required decisions and interpretations at various stages. We have documented how each decision was made and have made efforts to offer the most logical interpretations of the data. Despite its limitations, cluster analysis provides interpretable groups of children on the basis of their beverage intake, which goes beyond the traditional analysis of examining the mean intake. Cluster analysis allowed us to classify children into sufficiently large groups to assess group-level beverage intakes and sociodemographic factors.

Conclusion

Longitudinal data on beverage intake of Canadian children are needed. In the meantime, determining beverage intake patterns can help to understand the characteristics of high consumers of sweetened beverages. This information supports public health nutrition professionals to target intervention programmes and policies to decrease consumption of sweetened beverages.

Acknowledgements

The present study was supported by the Canadian Institutes for Health Research. The authors have no conflict of interest to declare. S.J.W., H.V. and A.D.D. designed the research; A.D.D. and N.N. conducted the research and analysed the data; A.D.D., S.J.W. and H.V. wrote the paper; A.D.D. had primary responsibility for the final content. All authors read and approved the final manuscript. The authors thank the Saskatchewan Research Data Centre for providing access to the data.

References

1.Popkin, BM, Armstrong, LE, Bray, GM et al. (2006) A new proposed guidance system for beverage consumption in the United States. Am J Clin Nutr 83, 529542.CrossRefGoogle ScholarPubMed
2.Rivera, JA, Munoz-Hernandez, O, Rosas-Peralta, M et al. (2008) Beverage consumption for a healthy life: recommendations for the Mexican population. Salud Publica Mex 50, 173195.CrossRefGoogle ScholarPubMed
3.Health Canada (2007) Eating Well With Canada's Food Guide. http://www.hc-sc.gc.ca/fn-an/food-guide-aliment/index-eng.php (accessed January 2009).Google Scholar
4.Rao, G (2008) Childhood obesity: highlights of AMA Expert Committee recommendations. Am Fam Physician 78, 5663.Google ScholarPubMed
5.Committee on Nutrition (2001) American Academy of Pediatrics: the use and misuse of fruit juice in pediatrics. Pediatrics 107, 12101213.CrossRefGoogle Scholar
6.Popkin, BM (2010) Patterns of beverage use across the lifecycle. Physiol Behav 100, 49.CrossRefGoogle ScholarPubMed
7.Wang, YC, Bleich, SN & Gortmaker, SL (2008) Increasing caloric contribution from sugar-sweetened beverages and 100% fruit juices among US children and adolescents, 1988–2004. Pediatrics 121, e1604e1614.CrossRefGoogle ScholarPubMed
8.Statistics Canada (2008) Food Statistics. http://www.statcan.gc.ca/pub/21-020-x/21-020-x2008001-eng.htm (accessed July 2009).Google Scholar
9.Blanton, CA, Moshfegh, AJ, Baer, DJ et al. (2006) The USDA automated multiple-pass method accurately estimates group total energy and nutrient intake. J Nutr 136, 25942599.CrossRefGoogle ScholarPubMed
10.Health Canada (2006) Canadian Community Health Survey Cycle 2·2 Nutrition (2004) – a guide to accessing and interpreting the data. http://www.hc-sc.gc.ca/fn-an/surveill/nutrition/commun/cchs_guide_escc-eng.php (accessed October 2010).Google Scholar
11.Statistics Canada (2008) Canadian Community Health Survey (CCHS) – Cycle 2.2 (2004): Nutrition – General Health (including vitamin & mineral supplements) & 24-hour dietary recall components: User Guide. http://www.statcan.gc.ca/imdb-bmdi/document/5049_D24_T9_V1-eng.pdf (accessed January 2011).Google Scholar
12.Statistics Canada (2009) Canadian Community Health Survey – Nutrition (CCHS): Detailed information for 2004 (Cycle 2·2). http://www.statcan.gc.ca/cgi-bin/imdb/p2SV.pl?Function=getSurvey&SDDS=5049&lang=en&db=imdb&adm=8&dis=22 (accessed June 2010).Google Scholar
13.LaRowe, TL, Moeller, SM & Adams, AK (2007) Beverage patterns, diet quality, and body mass index of us preschool and school-aged children. J Am Diet Assoc 107, 11241133.CrossRefGoogle ScholarPubMed
14.O'Connor, TM, Yang, S & Nicklas, TA (2006) Beverage intake among preschool children and its effect on weight status. Pediatrics 118, e1010e1018.CrossRefGoogle ScholarPubMed
15.SAS Institute Inc. (2009) SAS/STAT(R) 9·2 User's Guide. http://support.sas.com/documentation/cdl/en/statug/59654/HTML/default/fastclus_toc.htm (accessed April 2009).Google Scholar
16.Moeller, SM, Reedy, J, Millen, AE et al. (2007) Dietary patterns: challenges and opportunities in dietary patterns research an experimental biology workshop. J Am Diet Assoc 107, 12331239.CrossRefGoogle ScholarPubMed
17.Garriguet, D (2008) Beverage consumption of children and teens. Health Reports 19, 16.Google ScholarPubMed
18.Sanigorski, AM, Bell, AC & Swinburn, BA (2007) Association of key foods and beverages with obesity in Australian schoolchildren. Public Health Nutr 10, 152157.CrossRefGoogle ScholarPubMed
19.Libuda, L & Kersting, M (2009) Soft drinks and body weight development in childhood: is there a relationship? Curr Opin Clin Nutr Metab Care 12, 596600.CrossRefGoogle ScholarPubMed
20.Libuda, L, Alexy, U, Sichert-Hellert, W et al. (2008) Pattern of beverage consumption and long-term association with body-weight status in German adolescents-results from the DONALD study. Br J Nutr 99, 13701379.CrossRefGoogle ScholarPubMed
21.Vartanian, LR, Schwartz, MB & Brownell, KD (2007) Effects of soft drink consumption on nutrition and health: a systematic review and meta-analysis. Am J Public Health 97, 667675.CrossRefGoogle ScholarPubMed
22.Rodriguez-Artalejo, F, Garcia, EL, Gorgojo, L et al. (2003) Consumption of bakery products, sweetened soft drinks and yogurt among children aged 6–7 years: association with nutrient intake and overall diet quality. Br J Nutr 89, 419429.CrossRefGoogle ScholarPubMed
23.Institute of Medicine (2011) Dietary Reference Intakes for Calcium and Vitamin D. Washington, DC: National Academies Press; available at http://www.nap.eduGoogle Scholar
24.Keller, KL, Kirzner, J, Pietrobelli, A et al. (2009) Increased sweetened beverage intake is associated with reduced milk and calcium intake in 3- to 7-year-old children at multi-item laboratory lunches. J Am Diet Assoc 109, 497501.CrossRefGoogle ScholarPubMed
25.Libuda, L, Alexy, U, Remer, T et al. (2008) Association between long-term consumption of soft drinks and variables of bone modeling and remodeling in a sample of healthy German children and adolescents. Am J Clin Nutr 88, 16701677.CrossRefGoogle Scholar
26.Institute of Medicine (2000) Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, Carotenoids. Washington, DC: National Academies Press; available at http://www.nap.eduGoogle Scholar
27.Institute of Medicine (2005) Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate. Washington, DC: National Academies Press; available at http://www.nap.eduGoogle Scholar
28.Moshfegh, AJ, Rhodes, DG, Baer, DJ et al. (2008) The US Department of Agriculture Automated Multiple-Pass Method reduces bias in the collection of energy intakes. Am J Clin Nutr 88, 324332.CrossRefGoogle ScholarPubMed
Figure 0

Table 1 Classification by category and type for beverages

Figure 1

Table 2 Mean intake of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in children aged 2–5 years

Figure 2

Table 3 Mean intake of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in boys and girls aged 6–11 years

Figure 3

Table 4 Mean intakes of selected beverage groups, selected nutrients and energy, as well as sociodemographic characteristics, by beverage cluster in boys and girls aged 12–18 years