Sugar sweetened beverage consumption by Australian children: Implications for public health strategy

Transcription

1 RESEARCH ARTICLE Open Aess Sugar sweetened beverage onsumption by Australian hildren: Impliations for publi health strategy Katherine Hafekost *, Franis Mitrou, David Lawrene and Stephen R Zubrik Abstrat Bakground: High onsumption of sugar sweetened beverages (SSBs) has been linked to unhealthy weight gain and nutrition related hroni disease. Intake of SSB among hildren remains high in spite of publi health efforts to redue onsumption, inluding restritions on marketing to hildren and limitations on the sale of these produts in many shools. Muh extant literature on Australian SSB onsumption is out-dated and laks information on several key issues. We sought to address this using a ontemporary Australian dataset to examine purhase soure, onsumption pattern, dietary fators, and demographi profile of SSB onsumption in hildren. Methods: Data were from the 2007 Australian National Children s Nutrition and Physial Ativity Survey, a representative random sample of 4,834 Australian hildren aged 2-16 years. Mean SSB intake by type, loation and soure was alulated and logisti regression models were fitted to determine fators assoiated with different levels of onsumption. Results: SSB onsumption was high and age-assoiated differenes in patterns of onsumption were evident. Over 77% of SSB onsumed was purhased via supermarkets and 60% of all SSB was onsumed in the home environment. Less than 17% of SSB was soured from shool anteens and fast food establishments. Children whose parents had lower levels of eduation onsumed more SSB on average, while hildren whose parents had higher eduation levels were more likely to favour sweetened juies and flavoured milks. Conlusions: SSB intake by Australian hildren remains high and warrants ontinued publi health attention. Evidene based and age-targeted interventions, whih also reognise supermarkets as the primary soure of SSB, are reommended to redue SSB onsumption among hildren. Additionally, eduation of parents and hildren regarding the health onsequenes of high onsumption of both arbonated and non-arbonated SSBs is required. Bakground There has been a well-doumented rise in the prevalene of overweight, obesity and lifestyle related diseases in the Australian population over reent deades [1]. Although reent evidene suggests that the prevalene of overweight and obesity in Australian hildren has plateaued [2], within the global ontext the urrent levels remain disturbingly high. Additionally, the signifiant assoiated osts to both the individual and ommunity warrant ontinued publi health fous and development of effetive interventions. * Correspondene: khafekost@ihr.uwa.edu.au Telethon Institute for Child Health Researh, Centre for Child Health Researh, The University of Western Australia, PO Box 855, West Perth WA 6872, Australia Despite muh researh, and signifiant publi health and media attention, prevention and effetive long term treatment of exess weight gain and lifestyle related onditions remains elusive. Although it is likely that these onditions are due to a ombination of fators, the rapid inrease in their prevalene suggests hanging environmental fators may play a signifiant role. One potential fator, whih is temporally assoiated with the rise in overweight and obesity, is an inreased onsumption of refined arbohydrate [3]. Sugar sweetened beverages (SSBs) are a signifiant soure of refined arbohydrate in the diet of developed nations suh as the United States and Australia [4-7]. Current researh has attempted to examine the ontribution of SSB onsumption towards exess weight gain 2011 Hafekost et al; liensee BioMed Central Ltd. This is an Open Aess artile distributed under the terms of the Creative Commons Attribution Liense ( whih permits unrestrited use, distribution, and reprodution in any medium, provided the original work is properly ited.

2 Page 2 of 10 and disease. Notwithstanding differenes in methodology, partiipants, definitions of SSBs and funding soures [8], well-designed longitudinal and prospetive studies typially report positive assoiations between higher and more regular SSB onsumption and detrimental health outomes [9]. While the exat biologial mehanisms whih link SSB onsumption and weight gain remain unknown, a number of plausible hypotheses have been proposed whih explain how energy from SSB may bypass the homeostati regulatory systems that ontrol appetite and energy intake resulting in inreased hunger, redued satiety and exessive energy onsumption. For example, inomplete ompensation for alories onsumed as liquids, in omparison to isoalori solids, mayontributetoexessenergyintakeandinreased risk of weight gain and the high glyemi load typial of SSBs may redue satiety and inrease risk of disease. Additionally, the unique metaboli pathway of frutose mayinreasetheriskofanumberofshortandlong term negative health outomes, and some evidene suggests high SSB onsumption is linked with other poor dietary patterns whih may further inrease the risk of weight gain and disease [10-12]. In reognition of the link between SSB and exess weight gain, the Australian Dietary Guidelines [13] reommend limiting the onsumption of soft drink and ordial. Health advoay groups, suh as the Australian Medial Assoiation [14] and Publi Health Assoiation of Australia [15] have alled for a number of additional interventions to redue SSB intake. Proposed measures inlude the taxation of nutritionally poor produts suh as soft drinks with generated revenue to be hannelled into preventive programs and healthare, restrition of the sale of soft drinks in shools, and limitations on the promotion and advertising of these produts to hildren. These strategies mirror international efforts to redue SSB onsumption by hildren. Despite the inreased publi health attention, additional restritions on marketing, and limitations on the sale of these produts in many shools, intake of SSBs by hildren remains high. Moreover, the extant literature laks omprehensive desriptions of the nature of SSB onsumption in Australia with respet to the soure, pattern, demographi and other dietary fators assoiated with intake. We aim to address this by desribing the major soures of SSB, the patterns of onsumption, and demographi and dietary fators assoiated with high and regular onsumption in a representative random sample of Australian hildren. Additionally, we sought to identify potential target areas for future publi health intervention. Methods Data Data were from the 2007 Australian National Children s Nutrition and Physial Ativity Survey (NCNPAS), funded jointly by the Australian Government Department of Health and Aging, Australian Government Department of Agriulture, Fisheries and Forestry and industry group the Australian Food and Groery Counil. The NCNPAS was designed to assess the nutrient intake and physial ativity levels of a representative sample of Australian hildren. Demographi, dietary, physial ativity and basi anthropometri data were olleted from 4,487 hildren aged between 2 and 16 years. Interviews were onduted from February through August of Food, beverage and supplement intake was olleted from partiipants and aregivers, for hildren 8 years and younger who required assistane, in two standardised multiple-pass 24 hour realls. Additional information regarding usual dietary habits, inluding self- or aregiver reported usual daily intake of fruit and vegetables (number of serves), type of milk onsumed, whether salt was added to ooking or meals, and type of salt used (iodised or non-iodised) was olleted. This dietary intake information, in addition to soio-demographi data, was olleted during a omputer assisted personal interview (CAPI) and a seond, follow up omputer assisted telephone interview (CATI) was ompleted within 7 to 21 days. Height, weight, and waist irumferene data were olleted using reognised protools [16]. The final sample, who provided soio-demographi information and a minimum of one day of dietary reall data, was 4,834. This represented 40% of eligible households. As demographi information was olleted at the time of the CAPI, for families who were initially reruited but did not omplete the personal interview due to study quotas being filled (n = 1,450) or withdrew part way (n = 502), no demographi information was available. Additionally, as random digit dialling was used in reruitment, there is no information about the harateristis of non-respondents. Further information on study methodology an be found in the NCNPAS User Guide [17]. Data treatment Beverage lassifiation For the purpose of this analysis, beverages were lassified as follows: water, non-flavoured milk, sugar sweetened beverages (SSB), pure fruit juie (no added sugar), hot tea and offee (regardless of added sugar), alohol, artifiially sweetened diet beverages, and other (e.g. alohol and vegetable juie). SSBs were further ategorised into arbonated produts (inluding energy drinks), juies with added sugar, ordial (defined as flavoured drink onentrate), sports drinks, milkshakes/smoothies, and flavoured milk. Flavoured milks inluded both lowand high-fat produts as both have substantial sugar ontent and sugar levels were generally omparable.

3 Page 3 of 10 Loation Loation of beverage onsumption was provided on the NCNPAS dataset as: home or any other residene, plae of purhase, institution (inluding shool and hildare), during transport, leisure ativity and other. Soure The soure of beverage referred to where the produt was purhased and inluded the following ategories: fast food outlets inluding shool anteens, pakaged beverages purhased from supermarkets, and fresh produts suh as homemade juies, milkshakes or smoothies. As the ontribution from fresh produts was small, for analyti purposes this ategory was ombined with produts from unknown soures. Levels of onsumption Children were lassified as high, low-moderate or nononsumer of SSB based on their level of intake of SSB during the study period. Children whose SSB intake ontributed more than one third of their total daily intake of beverages by volume were lassified as high onsumers. As the average volume of beverage onsumed by hildren was approximately 1.4 litres per day, this ut-point was equivalent to slightly less than two glasses of SSB and approximately 50 grams of sugar per day. Those who reported no SSB onsumption during the study were labelled non-onsumers, and the remaining hildren were onsidered low-moderate onsumers. Additionally, high onsumption of eah subgroup of SSB was lassified separately. Children whose onsumption was in the 90th perentile or above for eah of the three main ategories, arbonateddrinks,sweetenedjuieorflavouredmilk,were onsidered high onsumers. In order to investigate fators assoiated with regular intake, onsumption on both days of dietary data olletion was used as a proxy measure of regular SSB onsumption. This analysis only inluded hildren who provided 2 days of ompleted dietary reall data (n = 4,633, 95.8%). Children who reported onsuming SSB on both reall days were lassified as regular onsumers (n = 2,331). This lassifiation was used as an additional measure of intake and it was unrelated to overall level of onsumption. The majority of regular onsumers (74%) were not onsidered high SSB onsumers. Demographi variables Household variables Household variables inluded the total annual household inome (before tax) at the time of the survey and number of hildren living in the household. Family type was derived and ategorised into original family, step or blended family, one parent family, and other (e.g. multi-generational families). The highest level of eduation ahieved by either parent or aregiver within the household was ategorised into university degree, voational eduation (TAFE ertifiate or diploma), year shooling, and year 10 shooling or below. Maternal variables Maternal age at the birth of the study hild and breast feeding duration, as realled by the mother, were inluded in modelling. Child variables Study age groups were seleted to align with the Nutrient Referene Values for Australia and New Zealand age bands [18]. Birth weight in grams was ategorised into low ( 2500 g), normal (male > 2500 g g, female > 2500 g-4000 g) and high (male > 4200 g, female > 4000 g), and urrent body mass index was derived and ompared to age and sex relevant riteria whih allowed lassifiation of partiipants as underweight, normal weight, overweight or obese [19]. Dietary variables Dietary variables were based on the 2-day dietary reall data. Added sugar intake, exluding sugars from beverages and fruit, and saturated fat onsumption were alulated as a perentage of total daily energy intakes from dietary reall data. Average alium intake was alulated from dietary reall data and ompared to age and sex based reommended dietary intakes referenes. Average daily affeineintakewassimilarlyalulated and ompared to age relevant ut-offs. As no Australian reommendations exist for intake of affeine by hildren, values were ompared to age-based Canadian reommendations [20]. In addition, parent and hild reported usual daily intake of fruit and vegetables, whih were olleted at the time of the CATI, were inluded. Statistial analysis All statistial analyses were arried out using SAS Version 9.2 (SAS Institute, In., Cary, NC) taking into aount population weights and the omplex survey design. Population level average intake was alulated usingsurveyweightsandassoiatedstandarderrors were alulated using expansion in Taylor series and the ultimate luster variane estimate tehnique [21]. Mean intakes were alulated for beverage ategory, SSB type, SSB soure and loation of SSB onsumption. T-tests were used to determine the signifiane of omparisons inluded in this report. Six logisti regression models were fitted, adjusting for the omplex sample design, to evaluate the assoiations between demographi and dietary fators and high, no and regular SSB onsumption. Fators assoiated with high onsumption of the most ommonly onsumed sub-groups of SSBs, arbonated beverages, sweetened juies and flavoured milk, were determined. Sports drinks, smoothies/milkshakes and ordials ontributed little to overall onsumption, and therefore these ategories of SSB were not inluded in sub-group analyses. Variables were eliminated from the final models if

4 Page 4 of 10 statistially non-signifiant, as determined by Wald s test (alpha = 0.05), and the most parsimonious model reported. Sensitivity analysis was undertaken to investigate the effet of altering the ut-point used to define high SSB onsumption. Logisti regression models were run using alternative ut-points (40, 25 and 20 per ent of daily beverage intake) in order to assess the impat on the strength and diretion of results. The average differene between the results of these models was less than 10 per ent, and the use of different ut-points did not hange the diretion of any results, not did it affet the interpretation of the final models. Results With respet to the general harateristis of the sample, omparisons with the national ensus figures [22] revealed differenes between the study partiipants and the general Australian population (Table 1). Households partiipating in the NCNPAS typially had higher levels of eduation and inome. Approximately 14% of partiipants were onsidered high SSB onsumers, 66% were low-moderate onsumers and approximately 20% reported no SSB onsumption on either study day (Table 1). The ontribution of SSB to total daily energy intake is displayed in Table 2. The proportion of total energy intake from SSB for high onsumers was double that of low-moderate onsumers (high onsumers 14.2%, 95% CI = , Low-moderate 6.0%, 95% CI = , p < 0.001). Age related differenes in the average ontribution of SSB to total energy intake were evident. Older hildren onsumed a signifiantly greater proportion of energy from SSB than younger hildren (2-3 year olds 4.0%, 95% CI = , year olds 7.5%, 95% CI = , p < 0.001). Age-assoiated differenes in the ontribution of SSB sub-ategories to total intake were evident (Table 3). ThemainsoureofSSBinthedietofyoungerhildren was sweetened juie (2-3 year olds 39.9%, 95% CI = ). For eah suessive age group we observed a greater mean ontribution from arbonated beverages Table 1 Partiipant Charateristis NCNPAS 2006 Census N Per ent Per ent Age group 2-3 years years years years Sex Male Female SSB onsumption High Low-moderate No BMI ategory Underweight Normal weight Overweight Obese Highest Level of Household Eduation a University qualifiation Voational qualifiation Year 12 or below Other Household Inome ($AUS per week) b $1-$ $500-$ $1000-$ $ Other Nil/negative a Census data inludes families with hildren aged 0-17 years b Census data inludes families with hildren aged 15 years and below, and families with dependent hildren aged years

5 Page 5 of 10 Table 2 Contribution of SSB to Total Daily Energy Intake, by Level of Consumption, Age Group and Sex Level of SSB Consumption 2-3 years (n = 1191) 4-8 years (n = 1262) 9-13 years (n = 1219) years (n = 1161) Sex Per ent 95% CI Per ent 95% CI Per ent 95% CI Per ent 95% CI No SSB Male 0.0 ( ) 0.0 ( ) 0.0 ( ) 0.0 ( ) Female 0.0 ( ) 0.0 ( ) 0.0 ( ) 0.0 ( ) Low-moderate Male 5.2 ( ) 5.9 ( ) 6.2 ( ) 6.6 ( ) Female 4.6 ( ) 5.5 ( ) 6.4 ( ) 6.8 ( ) High a Male 15.9 ( ) 13.5 ( ) 13.3 ( ) 14.9 ( ) Female 13.3 ( ) 13.1 ( ) 14.1 ( ) 16.4 ( ) Total Male 4.2 ( ) 5.6 ( ) 6.7 ( ) 7.6 ( ) Female 3.8 ( ) 5.1 ( ) 6.7 ( ) 7.4 ( ) a High onsumption: > 1/3 of average daily beverage intake from SSB and, in hildren over 9 years of age, arbonated drinks were the primary type of SSB onsumed (14-16 year olds 41.7%, 95% CI = , p < 0.001). There was no redution in the volume of non-arbonated SSBs onsumed among the older age groups to ompensate for the higher arbonated SSB onsumption (Table 4). The majority of SSB was onsumed at home (2-3 year olds 81.1%, 95% CI = , 4-8 year olds 67.2%, 95% CI = , 9-13 year olds 66.1%, 95% CI = , year olds 63.1%, 95% CI = ). The proportion of SSB onsumed at the plae of purhase, in institutions and during leisure ativities and transport was higher in older age groups. However, even amongst the oldest hildren who reported the greatest variation in the loation of onsumption, over 60 per ent of SSB was still onsumed in the home (Table 5). Supermarket purhased produts were the main soure of SSB in all age groups (2-3 year olds 87.9%, 95% CI = , 4-8 year olds 77.2%, 95% CI = , 9-13 year olds 76.2%, 95% CI = , year olds 74.38%, 95% CI = ). Although the average ontribution of SSB from fast food soures was higher in suessive age groups, it remained relatively small (2-3 year olds 9.3%, 95% CI = , year olds 21.1%, 95% CI = , p < 0.001) (Table 6). Of note, as a result of diffiulty in determining whether beverages onsumed in institutions were supermarket purhasedandbroughtfrom home or purhased at the institution, many SSBs onsumed in this loation were oded as from an unknown soure (52%). High onsumption of SSB, defined as onsumption whih was greater than a third of daily beverage intake, was assoiated with older age groups, male gender, lower levels of household eduation and a number of markers of poor dietary patterns inluding lower intake of fresh fruit and vegetables, and high affeine onsumption (Table 7). Conversely, no SSB onsumption was assoiated with younger age groups, higher levels of household eduation and markers of healthy dietary patterns. Both high and non-onsumers were more likely to be below the reommended dietary intake of alium than low-moderate SSB onsumers. Regular intake of SSB, defined by onsumption of SSB on both study days, was assoiated with older age groups, male gender, lower levels of household eduation, and low daily intake of fruit (Table 7). There was no signifiant relationship between regular onsumption and vegetable or affeine intake. Additionally, regular SSB onsumers were less likely to be below the reommended dietary intake for alium than irregular onsumers. Fators assoiated with high onsumption of the three main sub-groups of SSB differed (Table 8). Whilst high intake of arbonated SSB was signifiantly assoiated with lower levels of household eduation, intakes of sweetened juie and flavoured milk did not vary with Table 3 Mean Contribution of SSB Subategories to Total SSB Intake, by Age Group Beverage Classifiation (SSBs) 2-3 years (n = 836) 4-8 years (n = 1012) 9-13 years (n = 1055) years (n = 972) Per ent 95% CI Per ent 95% CI Per ent 95% CI Per ent 95% CI Carbonated soft drink 15.0 ( ) 23.4 ( ) 38.4 ( ) 41.7 ( ) Juie - added sugar 39.9 ( ) 34.7 ( ) 23.9 ( ) 20.9 ( ) Cordial 19.9 ( ) 12.8 ( ) 11.6 ( ) 10.0 ( ) Sports drink 0.3 ( ) 0.9 ( ) 3.1 ( ) 3.5 ( ) Flavoured milk 22.1 ( ) 24.1 ( ) 19.6 ( ) 21.2 ( ) Milkshake/Smoothie 3.8 ( ) 4.1 ( ) 3.5 ( ) 2.8 ( )

6 Page 6 of 10 Table 4 Mean Volume (mls) of SSB Subategories Consumed, by Age Group Beverage Classifiation (SSBs) 2-3 years (n = 836) 4-8 years (n = 1012) 9-13 years (n = 1055) years (n = 972) Volume(mls) 95% CI Volume (mls) 95% CI Volume (mls) 95% CI Volume (mls) 95% CI Carbonated soft drink 56.5 ( ) 134 ( ) 305 ( ) 426 ( ) Juie - added sugar 145 ( ) 176 ( ) 157 ( ) 175 ( ) Cordial 27.7 ( ) 28.7 ( ) 40.3 ( ) 38.4 ( ) Sports drink 1.2 ( ) 8.4 ( ) 30.1 ( ) 38.2 ( ) Flavoured milk 38.8 ( ) 74.3 ( ) 84.3 ( ) 131 ( ) Milkshake/Smoothie 12.9 ( ) 21.8 ( ) 28.9 ( ) 25.2 ( ) levels of household eduation. Similarly suggestive of an assoiation between parental eduation and dietary hoies, we observed that high onsumption of arbonated drinks was assoiated with lower intake of fruit, but high onsumption of flavoured milk and sweetened juie were not. Disussion SSBs ontributed a substantial amount of energy to the diet of Australian hildren with mean intakes ranging from 4 per ent in hildren 2-3 years old to 7.5 per ent in 14 to 16 year olds. These values are lower than those reported for US hildren (2-18 years) whose SSB intake is approximately 10 per ent of their energy intake [23]. However, given the growing evidene whih suggests biologial mehanisms linking high and regular onsumption of SSBs with negative short and long term health outomes [8], effetive publi health intervention to redue levels of intake by Australian hildren is reommended. The majority of SSB in the diet of hildren in all age groups was soured from supermarkets and onsumed at home. Only a small perentage of SSBs were purhased from fast food soures. Even among the year age group, who typially would have a higher level of independene from parents and some disretionary spending ompared with younger hildren, around 75 per ent of their SSB intake was purhased from supermarkets. These findings are similar to those of Wang and olleagues who reported, in a study of US hildren, 55 to 70 per ent of the alories from SSBs were onsumed in the home environment [24]. Although publi health strategies and interventions have traditionally foused on fast food soures of SSBs the results of this study suggest that in order to signifiantly redue levels of intake by hildren aged 2-16 years, future strategies should shift their fous to supermarket purhased SSBs and at home onsumption. The variation in the pattern of SSB onsumption by age is noteworthy. Of partiular interest was the high onsumption of sweetened juie by the 2-3 year old hildren. Similarly, Wang and olleagues reported, in US hildren, the main type of SSB onsumed by hildren aged 2-5 years was sweetened fruit punh and fruit juie [24]. During a ritial period of growth and developmentthehighintakeofsugaryprodutsisonerning. In addition to the early inreased risk of weight gain and assoiated hroni disease, intake of SSBs at a young age has been linked to lower intake of milk and, as a result, lower intake of alium, riboflavin, vitamin A and phosphorus [25]. Further, high intake of SSBs has been assoiated with dental aries [26], poor growth [27] and digestive issues in very young hildren [28]. In addition to the diret health onsequenes, there is some evidene to suggest that early food experienes influene later on-going food preferenes and dietary patterns [29,30]. At a young age when parents are likely to have almost omplete ontrol over a hild s diet, replaing SSBs with unsweetened milk or water may be relatively easy. Further, the development of publi health strategies to eduate parents about the health impliations of high onsumption of sweetened fruit juie should be a priority. The differenes in patterns of SSB onsumption between age groups suggests that foused, Table 5 Proportion of SSB Consumed at Home, Plae of Purhase, Institution and Other Loations Loation 2-3 years (n = 836) 4-8 years (n = 1012) 9-13 years (n = 1055) years (n = 972) Per ent 95% CI Per ent 95% CI Per ent 95% CI Per ent 95% CI Home 81.1 ( ) 67.2 ( ) 66.1 ( ) 63.1 ( ) Plae of Purhase 9.4 ( ) 11.5 ( ) 11.4 ( ) 13.0 ( ) Institution 3.5 ( ) 13.6 ( ) 10.6 ( ) 11.4 ( ) Other 5.3 ( ) 7.4 ( ) 11.8 ( ) 12.4 ( )

7 Page 7 of 10 Table 6 Proportion of SSB from Supermarket, Fast Food and Unknown Plae of Purhase Soure 2-3 years (n = 836) 4-8 years (n = 1012) 9-13 years (n = 1055) years (n = 972) All (n = 3875) Per ent 95% CI Per ent 95% CI Per ent 95% CI Per ent 95% CI Per ent 95% CI Supermarket 87.9 ( ) 77.2 ( ) 76.2 ( ) 74.4 ( ) 77.4 ( ) Fast Food 9.3 ( ) 14.5 ( ) 18.3 ( ) 21.1 ( ) 16.7 ( ) Unknown 3.2 ( ) 8.9 ( ) 5.8 ( ) 4.9 ( ) 6.3 ( ) age relevant interventions to target the main types, soures and loations of SSB intake is likely to improve the effetiveness of strategies. The demographi and dietary fators assoiated with high onsumption of arbonated SSBs and non-arbonated SSBs differed. While high onsumption of arbonated drinks was related to a number of markers of unhealthy dietary patterns and lower levels of household eduation, non-arbonated SSBs were not. These findings are similar to those of Ranjit and olleagues who reported that high arbonated beverage onsumption was assoiated with poor dietary and physial Table 7 Fators Assoiated with High, No and Regular SSB Consumption High SSB Consumption a No SSB Consumption b Regular SSB Consumption OR 95% CI OR 95% CI OR 95% CI Age group 2-3 years 0.31 ( ) 3.32 ( ) 0.41 ( ) 4-8 years 0.65 ( ) 1.82 ( ) 0.69 ( ) 9-13 years d years 1.18 ( ) 1.18 ( ) 0.90 ( ) Sex Female d 1.00 e Male 1.22 ( ) 1.21 ( ) Highest Level of Household Eduation University qualifiation d 1.00 Voational qualifiation 1.68 ( ) 0.57 ( ) 1.38 ( ) Year ( ) 0.68 ( ) 1.36 ( ) Year ( ) 0.47 ( ) 1.84 ( ) Unknown 1.96 ( ) 0.52 ( ) 1.81 ( ) Number of Children in Household 4 d 1.00 e ( ) ( ) ( ) Daily Vegetable Intake 4 serves 0.78 ( ) 1.49 ( ) e 2-3 serves d 1.00 < 2 serves 1.29 ( ) 0.93 ( ) Daily Fruit Intake > 2 serves 0.75 ( ) 1.13 ( ) 1.41 ( ) 2 serves d 1.00 < 2 serves 1.19 ( ) 0.78 ( ) 0.85 ( ) Calium Intake f RDI d 1.00 < RDI 1.32 ( ) 1.35 ( ) 0.70 ( ) Caffeine Intake g RDI* 1.00 e > RDI 2.32 ( ) a High (> 1/3 of average daily beverage intake) vs. low-medium and no onsumers b No vs. low-medium and high onsumers Regular (Consumed SSB on both study days) vs. irregular (SSB onsumption on 1 or no days) d Referene ategory e Non-signifiant variables tested but eliminated from the final models (p > 0.05). Additional variable tested but eliminated from all models inluded loation (rural or metropolitan), mother born overseas, birth weight, breastfeeding duration and BMI ategory. f RDI ut-offs for alium: 1-3 years < 500 mg/day, 4-8 years < 700 mg/day, 9-11 years < 1,000 mg/day, years < 1,300 mg/day, years < 1,300 mg/day g High affeine: 2-6 years > 45 mg per day, 7-9 years > 62.5 mg per day, years > 85 mg per day, years > 95 mg per day

8 Page 8 of 10 Table 8 Fators Assoiated with High Consumption of Carbonated Drinks, Sugar Sweetened Juie and Flavoured Milk Carbonated Drinks a Sweetened Juie a Flavoured Milk a OR 95% CI OR 95% CI OR 95% CI Age group 2-3 years 0.04 ( ) 0.48 ( ) 0.12 ( ) 4-8 years 0.20 ( ) 0.79 ( ) 0.45 ( ) 9-13 years b years 1.25 ( ) 1.22 ( ) 1.37 ( ) Sex Female b 1.00 Male 1.92 ( ) 1.46 ( ) Highest Level of Household Eduation University qualifiation b 1.00 Voational eduation 1.80 ( ) Year 11 or ( ) Year 10 or below 2.21 ( ) Don t know 2.24 ( ) Daily Fruit Intake > 2 serves 0.70 ( ) 2 serves b 1.00 < 2 serves 1.28 ( ) Calium Intake d RDI b 1.00 < RDI 0.37 ( ) Caffeine Intake e No b 1.00 Yes 3.15 ( ) 2.32 ( ) a High onsumption: 90 th perentile of intake Referene ategory Non-signifiant variables tested but eliminated from the final models (p > 0.05). Additional variable tested but eliminated from all models inluded loation (rural or metropolitan), mother born overseas, birth weight, breastfeeding duration and BMI ategory. d RDI ut-offs for alium: 1-3 years < 500 mg/day, 4-8 years < 700 mg/day, 9-11 years < 1,000 mg/day, years < 1,300 mg/day, years < 1,300 mg/day e High affeine: 2-6 years > 45 mg per day, 7-9 years > 62.5 mg per day, years > 85 mg per day, years > 95 mg per day ativity patterns in US hildren, while onsumption of non-arbonated drinks was assoiated with positive health behaviours [31]. The variation in fators assoiated with arbonated and non-arbonated SSBs suggests differenes in publi pereption. Non-arbonated SSBs are typially marketed based on health-related laims, or as funtional beverages whih ontribute to a healthy balaned diet and ative lifestyle. In ontrast, arbonated SSBs have been heavilytargetedbypubli health advoates. These easily identifiable produts are well reognised as junk foods with little nutritional value and onsumption of these produts is likely to be limited in health onsious families. Additionally, the Australian Dietary Reommendations [13] and Australian Guide to Healthy Eating [32], whih mirror the Dietary Guidelines for Amerians, suggest limiting intakes of soft drinks, ordials and/or sweetened drinks [13] but do not make expliit whether flavoured milks and sugar sweetened juies are inluded within this reommendation. Therefore a key step in reduing the intake of SSB appears to be eduation of the population regarding the health onsequenes of high and regular onsumption of both arbonated SSBs and non-arbonated SSBs, and the role of these produts in a healthy and balaned diet. Future publi health guidelines should make expliit appropriate intakes and serving sizes for hildren and make lear reommendations for both arbonated and non-arbonated produts. Using the various definitions of high and regular intake we onsistently found SSB intake to be assoiated with markers of poor dietary patterns. For example, high intake of affeine, and lower intake of alium, fruit and vegetables. This finding is supported by previous literature linking SSB onsumption with unhealthy dietary habits [10]. However, in ontrast with muh of the previous literature [8,9] there was no signifiant assoiation between high or regular onsumption of SSB and likelihood of being overweight or obese. This may be due to the ross-setional study design. As reported dietary patterns are likely to be influened by urrent weight status, drawing statistially valid and meaningful inferenes from ross setional data regarding the relationship between SSB onsumption and risk of weight gain is impossible. Growing evidene supports a mehanisti link between the high sugar, and speifially high frutose, ontent of SSBs and the risk of hroni disease and exess weight gain. The unique metaboli and hormonal

9 Page 9 of 10 effets of frutose are linked to a number of adverse short-term effets inluding enhaned de novo lipogenesis and triglyeride prodution resulting in dyslipidaemia, inreased systoli blood pressure, redued insulin and leptin sensitivity, impaired appetite ontrol and viseral adiposity [33-36]. These onditions may ontribute to an inreased risk of longer term hroni health onditions suh as type 2 diabetes, ardiovasular disease, ardio-renal disease, and obesity [37,38]. As ommon alori sweeteners suh as surose, high frutose orn syrup and fruit juie onentrate have relatively high onentrations of frutose they have similar metaboli effets [12]. Therefore, inreasing publi awareness of the dangers assoiated with high intake of alori sweeteners in the diets of hildren should be of great onern to publi health advoates. The urrent analysis is limited by its reliane on 2- day dietary reall, a ross setional study design and the potential for mislassifiation of beverage type or soure. As individual onsumption was based on 2 days of dietary reall it is possible that individual onsumption on the survey days was not representative of their typial onsumption patterns, in partiular, for hildren who did not onsume SSBs. However, as the survey days were randomly seleted for eah survey hild, population estimates of average onsumption of SSB should be unbiased. As data were self-reported, atual onsumption may vary from reorded intakes. Further, the involvement of the parent or aregiver in dietary data olletion in younger hildren (2-8 years) may have potentially influened the auray of results. However, these biases are likely to underestimate the onsumption of unhealthy foods suh as SSBs and therefore reported intakes are likely to be biased towards the null, leading to an underestimation of the reported assoiations. The ross-setional design means it is impossible to assoiate urrent onsumption with the development of long term health outomes suh as obesity or hroni disease. Due to the use of random digit dealing for partiipant reruitment no demographi information was available for nonrespondents. As a result, differenes between study partiipants and non-respondents ould not be determined and potential bias between these groups ould not be assessed. However, households partiipating in NCNPAS typially had slightly higher levels of eduation and inome in omparison to the general Australian population, and this ould potentially bias results and limit the generalizability of the researh. Conlusions These findings reveal opportunities to lose possible gaps in publi understanding of the role of both arbonated and non-arbonated SSBs in a healthy diet. Additionally, they highlight the need for evidene based, and age relevant publi health interventions whih target the primary soures and loation of SSB onsumption in order to effetively redue levels of intake by Australian hildren. Finally, the importane of periodi monitoring of hild health and nutrition status, to allow publi health strategies to remain effetive and relevant to the hanging needs of the population, is lear. Aknowledgements This study was funded by a program grant from the National Health and Medial Researh Counil (Grant Number ). The publi use file from the National Children s Nutrition and Physial Ativity Survey was provided by the Australian Soial Siene Data Arhive on behalf of the Commonwealth Sientifi Industrial and Researh Organisation and the University of Adelaide, who onduted the survey. The survey was funded by the Australian Government Department of Health and Ageing, the Australian Government Department of Agriulture, Fisheries and Forestry, and the Australian Food and Groery Counil. The views expressed in this paper are those of the authors. The organisations that funded and onduted the survey are responsible for the initial data only, and are not responsible for the analyses, interpretations or onlusions reahed by the authors. Authors ontributions KH, FM and DL oneived the original idea for the study. All authors ontributed to the development of the study methodology. KH undertook the data analysis and wrote the first draft of manusript. All authors ontributed to writing the paper, and approved the final manusript. Competing interests The authors delare that they have no ompeting interests. Reeived: 16 June 2011 Aepted: 22 Deember 2011 Published: 22 Deember 2011 Referenes 1. Booth ML, Chey T, Wake M, Norton K, Hesketh K, Dollman J, Robertson I: Change in the prevalene of overweight and obesity among young Australians, Am J Clin Nutr 2003, 77(1): Olds TS, Tomkinson GR, Ferrar KE, Maher CA: Trends in the prevalene of hildhood overweight and obesity in Australia between 1985 and Int J Obes 2010, 34(1): Slyper AH: The pediatri obesity epidemi: auses and ontroversies. J Clin Endorinol Metab 2004, 89(6): Somerset SM: Refined sugar intake in Australian hildren. Publ Health Nutr 2003, 6(8): Bleih SN, Wang YC, Wang Y, Gortmaker SL: Inreasing onsumption of sugar-sweetened beverages among US adults: to Am J Clin Nutr 2009, 89(1): Reedy J, Krebs-Smith SM: Dietary soures of energy, solid fats, and added sugars among hildren and adolesents in the United States. J Am Diet Asso 2010, 110(10): Duffey KJ, Popkin BM: Shifts in patterns and onsumption of beverages between 1965 and Obesity (Silver Spring) 2007, 15(11): Vartanian LR, Shwartz MB, Brownell KD: Effets of soft drink onsumption on nutrition and health: a systemati review and meta-analysis. Am J Publi Health 2007, 97(4): Malik VS, Shulze MB, Hu FB: Intake of sugar-sweetened beverages and weight gain: a systemati review. Am J Clin Nutr 2006, 84(2): Bes-Rastrollo M, Sanhez-Villegas A, Gomez-Graia E, Martinez JA, Pajares RM, Martinez-Gonzalez MA: Preditors of weight gain in a Mediterranean ohort: the Seguimiento Universidad de Navarra Study 1. Am J Clin Nutr 2006, 83(2): , quiz Ambrosini GL, Oddy WH, Robinson M, O Sullivan TA, Hands BP, de Klerk NH, Silburn SR, Zubrik SR, Kendall GE, Stanley FJ, et al: Adolesent dietary

10 Page 10 of 10 patterns are assoiated with lifestyle and family psyho-soial fators. Publ Health Nutr 2009, 12(10): Malik VS, Popkin BM, Bray GA, Despres JP, Hu FB: Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and ardiovasular disease risk. Cirulation 2010, 121(11): National Health and Medial Researh Counil: Dietary Guidelines for hildren and adolesents in Australia Canberra: Commonwealth of Australia AMA Position Statement on Obesity [ 3033] Federal Eletion Poliy Priorities. [ douments/phaaeletionpriorities.pdf]. 16. Marfell-Jones M, Olds T, Stewart A, Carter L: International Standards for Anthropometri Assessment Pothefstroom, South Afria; Univeristy of South Australia, CSIRO, I-view: User Guide National Children s Nutrition and Physial Ativity Survey Nutrient Referene Values for Australia and New Zealand. [ nhmr.gov.au]. 19. Cole TJ, Flegal KM, Niholls D, Jakson AA: Body mass index ut offs to define thinness in hildren and adolesents: international survey. Br Med J 2007, 335(7612): It s Your Health - Caffeine. [ 21. Wolter K: Introdution to Variane Estimation. New York Australian Bureau of Statistis (ABS): 2006 Census of Population and Housing Australia. Canberra Nielsen SJ, Popkin BM: Changes in beverage intake between 1977 and Am J Preventitive Med 2004, 27(3): Wang YC, Bleih SN, Gortmaker SL: Inreasing alori ontribution from sugar-sweetened beverages and 100% fruit juies among US hildren and adolesents, Pediatris 2008, 121(6):e Harnak L, Stang J, Story M: Soft drink onsumption among US hildren and adolesents: nutritional onsequenes. J Am Diet Asso 1999, 99(4): Marshall TA, Levy SM, Broffitt B, Warren JJ, Eihenberger-Gilmore JM, Burns TL, Stumbo PJ: Dental aries and beverage onsumption in young hildren. Pediatris 2003, 112(3 Pt 1):e Smith MM, Lifshitz F: Exess fruit juie onsumption as a ontributing fator in nonorgani failure to thrive. Pediatris 1994, 93(3): Committee on Nutrition: The Use and Misuse of Fruit Juie in Pediatris. Pediatris 2001, 107(5): Skinner JD, Carruth BR, Wendy B, Ziegler PJ: Children s food preferenes: a longitudinal analysis. J Am Diet Asso 2002, 102(11): Fiorito LM, Marini M, Mithell DC, Smiiklas-Wright H, Birh LL: Girls early sweetened arbonated beverage intake predits different patterns of beverage and nutrient intake aross hildhood and adolesene. JAm Diet Asso 2010, 110(4): Ranjit N, Evans MH, Byrd-Williams C, Evans AE, Hoelsher DM: Dietary and ativity orrelates of sugar-sweetened beverage onsumption among adolesents. Pediatris 2010, 126(4):e Department of Health and Family Servies: The Australian guide to healthy eating. Canberra DHFS; Basiano H, Federio L, Adeli K: Frutose, insulin resistane, and metaboli dyslipidemia. Nutr Metab 2005, 2(1): Jalal DI, Smits G, Johnson RJ, Chonhol M: Inreased frutose assoiates with elevated blood pressure. J Am So Nephrol 2010, 21(9): Stanhope KL, Shwarz JM, Keim NL, Griffen SC, Bremer AA, Graham JL, Hather B, Cox CL, Dyahenko A, Zhang W, et al: Consuming frutosesweetened, not gluose-sweetened, beverages inreases viseral adiposity and lipids and dereases insulin sensitivity in overweight/ obese humans. J Clin Invest 2009, 119(5): Teff KL, Elliott SS, Tshop M, Kieffer TJ, Rader D, Heiman M, Townsend RR, Keim NL, D Alessio D, Havel PJ: Dietary frutose redues irulating insulin and leptin, attenuates postprandial suppression of ghrelin, and inreases triglyerides in women. J Clin Endorinol Metab 2004, 89(6): Johnson RJ, Segal MS, Sautin Y, Nakagawa T, Feig DI, Kang DH, Gersh MS, Benner S, Sanhez-Lozada LG: Potential role of sugar (frutose) in the epidemi of hypertension, obesity and the metaboli syndrome, diabetes, kidney disease, and ardiovasular disease. Am J Clin Nutr 2007, 86(4): Miller A, Adeli K: Dietary frutose and the metaboli syndrome. Curr Opin Gastroenterology 2008, 24(2): Pre-publiation history The pre-publiation history for this paper an be aessed here: /prepub doi: / Cite this artile as: Hafekost et al.: Sugar sweetened beverage onsumption by Australian hildren: Impliations for publi health strategy. BMC Publi Health :950. Submit your next manusript to BioMed Central and take full advantage of: Convenient online submission Thorough peer review No spae onstraints or olor figure harges Immediate publiation on aeptane Inlusion in PubMed, CAS, Sopus and Google Sholar Researh whih is freely available for redistribution Submit your manusript at