Trends in Beverage Consumption Among Children and Adults,

Transcription

1 Trends in Beverage Consumption Among Children and Adults, Sara N. Bleich, Kelsey A. Vercammen, Jonathan Wyatt Koma, and Zhonghe Li Objective: This study aimed to provide the most recent national estimates for beverage consumption among children and adults in the United States. Methods: Dietary data were collected from 18,600 children aged 2 to 19 years and from 27,652 adults aged 20 years in the 2003 to 2014 National Health and Nutrition Examination Survey. Total beverage and sugar-sweetened beverage (SSB) consumption was measured by 24-hour dietary recall. Results: From 2003 to 2014, per capita consumption of all beverages declined significantly among children ( calories; P < 0.001) and adults ( calories; P < 0.001). In the survey, 60.7% of children and 50.0% of adults drank SSBs on a given day, which is significantly lower than , when 79.7% of children and 61.5% of adults reported drinking SSBs. From 2003 to 2014, per capita consumption of SSBs declined from calories to calories (P < 0.001) for children and from calories to calories (P < 0.001) for adults. The absolute levels for the percentage of SSB drinkers and per capita consumption of SSBs were highest among black, Mexican American, and non-mexican Hispanic children, adolescents, and young adults for all years of the study. Conclusions: Overall, beverage and SSB consumption declined for children and adults from 2003 to The levels of consumption are highest among black, Mexican American, and non-mexican Hispanic participants. (2018) 26, doi: /oby Introduction Sugar-sweetened beverages (SSBs) are a leading source of added sugar to the diet for adults and children in the United States (1). The consumption of SSBs is strongly linked to obesity (2-5), though some studies have found null results (6-8). On a given day, about half of adults and two-thirds of children consume at least one SSB (1). In recent years, SSB consumption has begun to decline (9); however, rates of SSB consumption have remained persistently higher for racial and ethnic minorities (1), who are also at higher risk for obesity (10). Reducing SSB consumption is the focus of many public health efforts in the United States. The Dietary Guidelines for Americans recommend reducing the consumption of added sugar to less than 10% of calories per day and, specifically, to choose beverages with no added sugars (11). Several jurisdictions have recently passed beverage taxes that increase the price of SSBs relative to noncalorie beverages such as water. There have also been a variety of unsuccessful legislative attempts, such as SSB warning labels, portion size caps, and restrictions on the purchase of SSBs with benefits from the Supplemental Nutrition Assistance Program. Given the considerable research, policy, and advocacy activities focused on SSB consumption, surveillance on trends in SSB consumption remains important. Also critical is understanding how beverage intake has changed overall in order to identify key future areas to target, as SSBs represent only one source of beverage calories. Though there have been several studies on patterns and trends in SSB consumption (1,12-16), few studies have examined national changes in beverage consumption overall and by sociodemographic groups (17-20). No study has presented the most recent national data on SSB consumption from the National Health and Nutrition Examination Survey (NHANES). Available evidence looking at the consumption of all beverages is several years old (19,20). The purpose of this study is to provide the most recent national estimates of the percentage of drinkers and amount consumed by beverage type as well as to analyze trends in beverage consumption between 2003 and Methods Nationally representative data from NHANES for 2003 to 2014 were used. A complete description of data collection procedures and analytic guidelines are available elsewhere ( Department of Health Policy and Management, Harvard T.H. Chan School of Public Health, Harvard College, Boston, Massachusetts, USA. Correspondence: Sara N. Bleich (sbleich@hsph.harvard.edu) Disclosure: The authors declared no conflict of interest. Additional Supporting Information may be found in the online version of this article. Received: 22 June 2017; Accepted: 29 September 2017; Published online 14 November doi: /oby VOLUME 26 NUMBER 2 FEBRUARY

2 TABLE 1 Unweighted sample sizes for children (aged 2-19) and adults (aged 201) by gender, age group, and race/ethnicity, NHANES Age groups, y All groups Non-Hispanic white Non-Hispanic black Mexican American Non-Mexican Hispanic All children ,600 5,735 5,704 5,461 1, ,384 1,426 1,197 1, ,808 1,785 1,760 1, ,408 2,524 2,747 2, Males ,443 2,955 2,894 2, , , ,307 1,316 1,416 1, Females ,157 2,780 2,810 2, , , ,101 1,208 1,331 1, All adults ,652 14,061 6,310 4,910 2, ,407 4,352 2,183 2, ,792 4,252 2,152 1, ,453 5,457 1,975 1, Males ,417 6,917 3,036 2,391 1, ,461 2,083 1, ,281 2,101 1, ,675 2,733 9, Females ,235 7,144 3,274 2,519 1, ,946 2,269 1,139 1, ,511 2,151 1, ,778 2, The NHANES is a population-based survey designed to collect information on the health and nutrition of the US noninstitutionalized population. Participants were selected based on a multistage, clustered, probability sampling strategy. Since 2003, data have been collected annually. The most recent data presented in this analysis are from the survey, and trends were analyzed between and with six time periods. The response rates for individuals selected for the survey interview were 79% in , 80% in , 78% in , 79% in , 73% in , and 71% in The study sample consisted of children aged 2 to 19 years old and adults 20 years old with completed 24-hour dietary recalls from each of the time periods. For the 24-hour dietary recall information, responses for children ages 2 to 5 years were provided by caretakers, responses for children ages 6 to 11 years were assisted by an adult, and responses for adolescents were completed independently. Survey respondents were excluded if their dietary recall was incomplete or unreliable (as determined by the NHANES staff). Because of the small sample size of the other race and ethnicity category, and because oversampling of non-hispanic Asians did not begin until the NHANES , we only reported the results for non- Hispanic white participants, non-hispanic black participants (hereafter referred to as NH white participants and NH black participants), Mexican American participants, and other Hispanic participants (hereafter referred to as non-mexican Hispanic participants) in the main analyses. The results for the other race group are available upon request. Participants self-reported their race and whether they were of Hispanic origin. Survey respondents reported all food and beverages consumed in a prior 24-hour period (midnight to midnight) and reported the type and quantity for each item. After the dietary interview, all reported food and beverage items were systemically coded with the use of the US Department of Agriculture Food and Nutrient Database that corresponded with the survey year (21). Caloric content and other nutrients derived from each consumed food or beverage item were calculated based on the quantity of food and beverages reported and the corresponding nutrient contents by the National Center for Health Statistics. We only used the first dietary recall from each survey for this analysis, as nonresponse is higher for the second dietary interview across all waves, and including it would make our generalizability to VOLUME 26 NUMBER 2 FEBRUARY

3 Trends in Beverage Consumption Bleich et al. Figure 1 Percentage of U.S. children (aged 2 to 19) and adults (aged 20+) consuming SSBs from 2003 to NHANES is not a continuous sample, and data are instead collected at discrete points in time. the US population nonrobust. We also relied on the first dietary recall because the interview settings differ between the two; day 1 recall was collected via a face-to-face interview by trained dietary interviewers in a mobile examination center. Day 2 recall was collected via telephone 3 to 10 days after day 1. We identified seven mutually exclusive beverage categories from 662 beverage items, including (1) SSBs (soda, fruit drinks and punches, sports drinks, low-calorie drinks, and other sweetened beverages), (2) 100% juice, (3) diet beverages, (4) milk (including flavored milk, almond milk, soy milk, and other milk alternatives), (5) unsweetened coffee or tea (hereafter referred to as coffee or tea), (6) alcohol, and (7) water (first available in the NHANES ). Of note, coffee and tea with added sugar were coded as either other SSB or low-calorie SSB. To create the analytic data set for this analysis, we treated each of the six NHANES waves as six samples of the US population (assuming that the population is constant across time) and used the day 1 dietary response weight, which adjusted for the complex survey weight and dietary nonresponse weight. Estimates of beverage consumption (measured by the percentage of drinkers and calories) are presented with 95% confidence intervals. In the text, we report P values for the linear trend from 2003 to Differences in beverage consumption levels were tested separately in children and adults. All analyses were conducted by using Stata software, version 14.2 (StataCorp LLC, College Station, Texas). The Stata survey procedures were used to model the outcome variables, which take into account sampling clustering, strata, and sampling weight. We used the Stata survey logistic procedure to estimate the yearly proportions of drinkers in each category and Stata survey regression to estimate national yearly total calories. We conducted age-stratified regression analyses for beverages overall, SSBs, and SSB subtypes. Multivariate regressions were used to adjust for potential changes in population composition over time, including age, race or ethnicity, and gender. For adults, we additionally adjusted for education level. All results were weighted to be representative of the general population (except for Table 1, which provides the unweighted sample sizes), the complex survey variance was estimated by using Taylor Series linearization (SAS Institute Inc., Cary, North Carolina), and all tables and figures report predicted means based on the adjusted models. To account for the possibility of false positives because of multiple comparisons, we used the Bonferroni correction (22), setting the significance value at P < 0.01 for the five mutually exclusive beverage categories for which we had complete data for all years. Results There were 18,600 children (9,443 males and 9,157 females) and 27,652 adults (13,417 males and 14,235 females) with complete dietary recall in the NHANES from 2003 to Table 1 reports sample sizes by age, gender, and race or ethnic group. The prevalence of SSB consumption declined significantly from 2003 to 2014 among children (79.7%-60.7%; P < 0.001) and adults (61.5%- 50.0%; P < 0.001) (Figure 1). Table 2 reports the prevalence of SSB consumption stratified by age and race or ethnic group. Among children, the prevalence of SSB consumption decreased significantly across all age categories (P < for all three). Within these agespecific categories, there was a significant decrease in SSB consumption prevalence for each gender (Supporting Information Table S1) except females ages 12 to 19 years old. With respect to race and ethnicity, SSB consumption prevalence declined significantly across all age categories of NH white participants, among NH black participants ages 12 to 19 years old (89.6%-78.3%; P < 0.001), and among Mexican American participants ages 2 to 5 years old (80.4%-53.6%; P ) and 12 to 19 years old (86.3%-71.6%; P < 0.001). No significant declines in the prevalence of SSB consumption were observed among non-mexican Hispanic children. Among adults, the prevalence of SSB consumption declined significantly among 20- to 39-year-olds (72.5%-60.8%; P < 0.001) and 40- to 59-year-olds (61.0%-46.4%; P < 0.001). Within age-specific categories of adults, significant declines in SSB consumption were observed for NH white participants 434 VOLUME 26 NUMBER 2 FEBRUARY

4 TABLE 2 Age-specific trends in the percentage of US children (aged 2-19) and adults (aged 201) consuming sugary beverages from 2003 to 2014 by race/ethnicity Survey year Ages 2-5 Ages 6-11 Ages Ages Ages Ages 601 All ( ) 81.8 ( ) 83.3 ( ) 72.5 ( ) 61.0 ( ) 46.4 ( ) ( ) 71.2 ( ) 78.5 ( ) 67.2 ( ) 55.0 ( ) 44.5 ( ) ( ) 76.9 ( ) 77.9 ( ) 68.1 ( ) 59.4 ( ) 45.7 ( ) ( ) 68.9 ( ) 76.2 ( ) 65.5 ( ) 54.3 ( ) 45.6 ( ) ( ) 75.0 ( ) 75.0 ( ) 63.7 ( ) 56.5 ( ) 45.2 ( ) ( ) 63.5 ( ) 65.4 ( ) 60.8 ( ) 46.4 ( ) 39.6 ( ) P value for linear trend < < < < < NH white ( ) 82.5 ( ) 81.8 ( ) 69.5 ( ) 58.8 ( ) 43.1 ( ) ( ) 66.2 ( ) 75.1 ( ) 65.1 ( ) 51.6 ( ) 41.5 ( ) ( ) 76.4 ( ) 77.8 ( ) 65.3 ( ) 57.9 ( ) 42.4 ( ) ( ) 65.3 ( ) 74.8 ( ) 62.3 ( ) 50.0 ( ) 42.4 ( ) ( ) 76.9 ( ) 77.5 ( ) 60.0 ( ) 54.3 ( ) 41.7 ( ) ( ) 59.8 ( ) 62.8 ( ) 56.2 ( ) 42.9 ( ) 36.7 ( ) P value for linear trend < NH black ( ) 78.5 ( ) 89.6 ( ) 86.0 ( ) 74.5 ( ) 60.4 ( ) ( ) 69.8 ( ) 87.0 ( ) 76.1 ( ) 71.5 ( ) 59.1 ( ) ( ) 62.7 ( ) 81.1 ( ) 77.2 ( ) 75.1 ( ) 58.2 ( ) ( ) 69.3 ( ) 82.4 ( ) 73.9 ( ) 70.2 ( ) 65.1 ( ) ( ) 62.2 ( ) 78.1 ( ) 76.0 ( ) 69.0 ( ) 64.7 ( ) ( ) 65.5 ( ) 78.3 ( ) 77.6 ( ) 58.3 ( ) 51.6 ( ) P value for linear trend < Mexican American ( ) 86.5 ( ) 86.3 ( ) 82.6 ( ) 65.8 ( ) 56.5 ( ) ( ) 80.4 ( ) 80.8 ( ) 81.0 ( ) 71.7 ( ) 47.1 ( ) ( ) 78.6 ( ) 79.8 ( ) 82.3 ( ) 62.3 ( ) 52.7 ( ) ( ) 85.7 ( ) 80.4 ( ) 72.9 ( ) 70.5 ( ) 54.1 ( ) ( ) 79.3 ( ) 72.4 ( ) 71.4 ( ) 63.2 ( ) 58.2 ( ) ( ) 69.7 ( ) 71.6 ( ) 77.3 ( ) 58.5 ( ) 49.5 ( ) P value for linear trend < Non-Mexican Hispanic ( ) 97.4 ( ) 88.9 ( ) 69.2 ( ) 59.5 ( ) 56.5 ( ) ( ) 76.4 ( ) 81.0 ( ) 71.5 ( ) 44.6 ( ) 46.5 ( ) ( ) 68.6 ( ) 77.9 ( ) 71.8 ( ) 65.8 ( ) 47.5 ( ) ( ) 71.0 ( ) 77.7 ( ) 77.7 ( ) 59.7 ( ) 53.1 ( ) ( ) 75.3 ( ) 73.4 ( ) 73.7 ( ) 52.9 ( ) 49.1 ( ) ( ) 76.9 ( ) 73.8 ( ) 63.4 ( ) 56.6 ( ) 46.7 ( ) P value for linear trend Data reported as mean and 95% confidence interval. ages 20 to 39 years old (69.5%-56.2%; P < 0.001) and 40 to 59 years old (58.8%-42.9%; P ) and NH black participants ages 40 to 59 years old (74.5%-58.3%; P ). No significant declines in the prevalence of SSB consumption were observed for Mexican American or non-mexican Hispanic adults. Supporting Information Table S2 reports the prevalence of SSB consumption for non-hispanic Asian Americans by using the 4 years available ( ). Table 3 summarizes the prevalence of beverage consumption among children and adults by type of beverage. Among children, the prevalence of 100% juice consumption declined significantly among 12- to 19-year-olds (26.2%-18.6%; P ), water increased significantly across all age groups (P < for all three), and no significant changes were seen for coffee and tea, milk, or diet beverage consumption for any age group. Among adults, the prevalence of VOLUME 26 NUMBER 2 FEBRUARY

5 Trends in Beverage Consumption Bleich et al. TABLE 3 Percentage of US children (aged 2-19) and adults (aged 201) consuming beverages by type of beverage from 2003 to 2014 Survey year Ages 2-5 Ages 6-11 Ages Ages Ages Ages % juice ( ) 35.8 ( ) 26.2 ( ) 19.8 ( ) 22.0 ( ) 32.5 ( ) ( ) 30.2 ( ) 23.4 ( ) 20.1 ( ) 20.6 ( ) 35.7 ( ) ( ) 38.1 ( ) 20.8 ( ) 21.6 ( ) 19.3 ( ) 28.7 ( ) ( ) 31.7 ( ) 21.9 ( ) 22.7 ( ) 21.5 ( ) 31.6 ( ) ( ) 33.2 ( ) 22.9 ( ) 19.3 ( ) 19.1 ( ) 27.3 ( ) ( ) 31.8 ( ) 18.6 ( ) 18.1 ( ) 15.8 ( ) 24.7 ( ) P value for linear trend < Diet ( ) 5.4 ( ) 6.0 ( ) 16.3 ( ) 21.8 ( ) 18.9 ( ) ( ) 8.7 ( ) 6.8 ( ) 19.3 ( ) 26.6 ( ) 20.9 ( ) ( ) 6.5 ( ) 8.3 ( ) 18.2 ( ) 22.8 ( ) 19.4 ( ) ( ) 6.6 ( ) 8.5 ( ) 17.9 ( ) 23.2 ( ) 20.3 ( ) ( ) 4.6 ( ) 7.9 ( ) 15.0 ( ) 19.3 ( ) 20.1 ( ) ( ) 6.4 ( ) 5.9 ( ) 11.6 ( ) 19.2 ( ) 17.5 ( ) P value for linear trend Milk ( ) 76.4 ( ) 56.3 ( ) 44.2 ( ) 42.6 ( ) 54.9 ( ) ( ) 76.8 ( ) 52.8 ( ) 45.5 ( ) 44.3 ( ) 56.0 ( ) ( ) 72.7 ( ) 57.7 ( ) 40.9 ( ) 45.8 ( ) 53.5 ( ) ( ) 77.0 ( ) 57.2 ( ) 45.2 ( ) 47.1 ( ) 53.4 ( ) ( ) 78.6 ( ) 55.4 ( ) 37.1 ( ) 40.2 ( ) 51.7 ( ) ( ) 72.6 ( ) 51.7 ( ) 35.2 ( ) 38.3 ( ) 44.8 ( ) P value for linear trend < < Coffee/tea ( ) 3.4 ( ) 9.2 ( ) 39.8 ( ) 65.8 ( ) 80.3 ( ) ( ) 4.9 ( ) 8.9 ( ) 40.7 ( ) 67.0 ( ) 79.6 ( ) ( ) 5.5 ( ) 15.5 ( ) 41.3 ( ) 64.3 ( ) 77.8 ( ) ( ) 3.0 ( ) 10.8 ( ) 45.4 ( ) 66.7 ( ) 76.6 ( ) ( ) 4.5 ( ) 10.5 ( ) 43.8 ( ) 65.5 ( ) 78.3 ( ) ( ) 6.2 ( ) 15.2 ( ) 45.2 ( ) 65.6 ( ) 72.5 ( ) P value for linear trend < Water ( ) 79.6 ( ) 73.8 ( ) 77.4 ( ) 76.8 ( ) 81.3 ( ) ( ) 72.3 ( ) 73.3 ( ) 77.1 ( ) 74.2 ( ) 79.0 ( ) ( ) 74.5 ( ) 75.4 ( ) 78.9 ( ) 76.5 ( ) 81.6 ( ) ( ) 77.9 ( ) 80.1 ( ) 84.5 ( ) 81.3 ( ) 82.9 ( ) ( ) 81.9 ( ) 79.6 ( ) 81.9 ( ) 81.2 ( ) 81.9 ( ) P value for linear trend < < < < < < Alcohol ( ) 26.4 ( ) 17.8 ( ) ( ) 26.9 ( ) 22.5 ( ) ( ) 22.7 ( ) 20.7 ( ) ( ) 28.3 ( ) 24.7 ( ) ( ) 28.6 ( ) 23.8 ( ) ( ) 25.7 ( ) 24.3 ( ) P value for linear trend Data reported as mean and 95% confidence interval. 436 VOLUME 26 NUMBER 2 FEBRUARY

6 TABLE 4 Percentage of US children (aged 2-19) and adults (aged 201) consuming SSBs by type of SSB from 2003 to 2014 Survey year Ages 2-5 Ages 6-11 Ages Ages Ages Ages 601 Soda ( ) 55.4 ( ) 61.4 ( ) 50.7 ( ) 40.2 ( ) 21.6 ( ) ( ) 39.5 ( ) 51.4 ( ) 41.3 ( ) 30.9 ( ) 21.5 ( ) ( ) 40.3 ( ) 41.1 ( ) 39.0 ( ) 31.0 ( ) 20.0 ( ) ( ) 32.4 ( ) 38.3 ( ) 36.7 ( ) 26.5 ( ) 18.7 ( ) ( ) 35.0 ( ) 38.4 ( ) 34.9 ( ) 28.9 ( ) 20.4 ( ) ( ) 24.2 ( ) 33.3 ( ) 35.2 ( ) 25.5 ( ) 19.0 ( ) P value for linear trend < < < < < Fruit drinks ( ) 35.3 ( ) 28.1 ( ) 18.4 ( ) 14.5 ( ) 16.2 ( ) ( ) 36.7 ( ) 25.5 ( ) 16.7 ( ) 11.5 ( ) 12.2 ( ) ( ) 31.3 ( ) 19.1 ( ) 13.7 ( ) 12.0 ( ) 8.8 ( ) ( ) 26.3 ( ) 22.3 ( ) 13.4 ( ) 12.3 ( ) 9.9 ( ) ( ) 30.3 ( ) 18.5 ( ) 14.5 ( ) 8.8 ( ) 8.5 ( ) ( ) 27.2 ( ) 20.6 ( ) 10.9 ( ) 9.0 ( ) 7.6 ( ) P value for linear trend < < < < Sports drinks ( ) 7.2 ( ) 7.2 ( ) 3.6 ( ) 1.2 ( ) 0.7 ( ) ( ) 7.5 ( ) 10.9 ( ) 6.0 ( ) 3.0 ( ) 1.8 ( ) ( ) 7.8 ( ) 10.5 ( ) 6.2 ( ) 4.5 ( ) 1.7 ( ) ( ) 6.6 ( ) 12.4 ( ) 7.6 ( ) 3.5 ( ) 1.8 ( ) ( ) 6.5 ( ) 9.9 ( ) 6.9 ( ) 4.3 ( ) 3.2 ( ) ( ) 8.5 ( ) 8.9 ( ) 6.3 ( ) 3.0 ( ) 2.2 ( ) P value for linear trend Low-calorie ( ) 2.4 (0.0-5) 2.2 ( ) 2.3 ( ) 4.3 ( ) 5.6 ( ) ( ) 6.8 ( ) 3.6 ( ) 4.7 ( ) 7.4 ( ) 8.9 ( ) ( ) 6.4 ( ) 8.1 ( ) 6 ( ) 8.6 ( ) 11.1 ( ) ( ) 6.1 ( ) 8.2 ( ) 6.1 ( ) 8.9 ( ) 9.0 ( ) ( ) 8.7 ( ) 7.2 ( ) 5.3 ( ) 7.1 ( ) 7.1 ( ) ( ) 6.4 ( ) 4.8 ( ) 5.2 ( ) 6.2 ( ) 7.7 ( ) P value for linear trend Other ( ) 15.3 ( ) 18.6 ( ) 19.4 ( ) 12.7 ( ) 10.3 ( ) ( ) 9.4 ( ) 20.7 ( ) 22 ( ) 14.4 ( ) 7.7 ( ) ( ) 23.0 ( ) 27.3 ( ) 21.8 ( ) 18.9 ( ) 10.9 ( ) ( ) 24.4 ( ) 24.8 ( ) 21.0 ( ) 16.3 ( ) 13.4 ( ) ( ) 24 ( ) 27.6 ( ) 20.5 ( ) 20 ( ) 12.4 ( ) ( ) 18.5 ( ) 17.4 ( ) 15.6 ( ) 12.3 ( ) 7.4 ( ) P value for linear trend Data reported as mean and 95% confidence interval. 100% juice consumption decreased significantly among those 60 years or older (32.5%-24.7%; P < 0.001), milk decreased significantly among 20- to 30-year-olds (44.2%-35.2%; P < 0.001) and those 60 years or older (54.9%-44.8%; P < 0.001), coffee and tea decreased significantly among those 60 years or older (80.3%- 72.5%; P < 0.001), and diet beverages decreased significantly among those 20 to 39 years old (16.3%-11.6%; P ) and 40 to 59 years old (21.8%-19.2%; P ). The prevalence of water consumption increased across all age groups (P < for all three), and no significant changes were seen for alcohol. Table 4 summarizes the prevalence of SSB consumption among children and adults by type of SSB. Across all age groups of children, the prevalence of soda and fruit drink consumption decreased significantly. The prevalence of low-calorie SSB consumption increased significantly among 12- to 19-year-olds (2.2%-4.8%; P ), and other SSB consumption increased among 6- to 11-year-olds (15.3%- 18.5%; P ). No significant changes were seen for sports drink consumption in children. Across all adult age categories, the prevalence of fruit drink consumption decreased significantly, and soda consumption decreased among 20- to 39-year-olds (50.7%-35.2%; VOLUME 26 NUMBER 2 FEBRUARY

7 Trends in Beverage Consumption Bleich et al. TABLE 5 Per capita consumption of beverage calories among US children (aged 2-19) and adults (aged 201) from 2003 to 2014 Survey year Ages 2-5 Ages 6-11 Ages Ages Ages Ages 601 SSB ( ) ( ) ( ) ( ) ( ) 75.3 ( ) ( ) ( ) ( ) ( ) ( ) 75.2 ( ) ( ) ( ) ( ) ( ) ( ) 68.1 ( ) ( ) ( ) ( ) ( ) ( ) 81.4 ( ) ( ) ( ) ( ) ( ) ( ) 87.7 ( ) ( ) ( ) ( ) ( ) ( ) 68.1 ( ) P value for linear trend < < < < % juice ( ) 46.9 ( ) 50.1 ( ) 33.7 ( ) 35.7 ( ) 34.6 ( ) ( ) 37.7 ( ) 43.8 ( ) 33.4 ( ) 33.7 ( ) 36.2 ( ) ( ) 43.6 ( ) 36.3 ( ) 38.9 ( ) 28.2 ( ) 30.7 ( ) ( ) 37.8 ( ) 39.7 ( ) 42.6 ( ) 29.9 ( ) 34.8 ( ) ( ) 40.6 ( ) 38.7 ( ) 33.7 ( ) 26.2 ( ) 31.0 ( ) ( ) 34.8 ( ) 29.4 ( ) 24.2 ( ) 21.0 ( ) 24.6 ( ) P value for linear trend Milk ( ) ( ) ( ) ( ) 76.3 ( ) 72.5 ( ) ( ) ( ) ( ) 88.7 ( ) 78.6 ( ) 77.1 ( ) ( ) ( ) ( ) 70.1 ( ) 73.2 ( ) 66.2 ( ) ( ) ( ) ( ) 72.8 ( ) 78.7 ( ) 68.7 ( ) ( ) ( ) ( ) 60.6 ( ) 66.2 ( ) 72.9 ( ) ( ) ( ) ( ) 67.5 ( ) 55.7 ( ) 61.6 ( ) P value for linear trend < < < Alcohol ( ) ( ) 56.8 ( ) ( ) ( ) 60.8 ( ) ( ) ( ) 59.8 ( ) ( ) ( ) 75.8 ( ) ( ) ( ) 70.0 ( ) ( ) ( ) 69.0 ( ) P value for linear trend Data reported as mean and 95% confidence interval. Per capita consumption of diet beverages, coffee or tea, and water is not reported given their negligible contributions to calories. P < 0.001) and 40- to 59-year-olds (40.2%-25.5%; P < 0.001). The consumption of sports drinks, low-calorie SSBs, and other SSBs did not change significantly for adults. The per capita consumption of beverage calories is presented in Table 5. Among children, calories from SSBs and milk decreased significantly across all ages, while calories from 100% juice decreased among 2- to 5-year-olds ( kcal; P ) and 12- to 19-year-olds ( kcal; P ). Among adults, calories from SSBs decreased significantly among 20- to 39-year-olds ( kcal; P < 0.001), while calories from 100% juice declined among 40- to 59-year-olds ( kcal; P ). Calories from milk decreased significantly among 20- to 39-yearolds ( kcal; P < 0.001) and among 40- to 59-year-olds ( kcal; P ), while calories from alcohol did not change significantly across any age category. The per capita consumption of SSB calories is presented in Figure 2 (children) and Figure 3 (adults). Supporting Information Table S3 reports the point estimates. Among children, calories from soda and fruit drinks decreased significantly for all ages (P < for all). Calories from other SSB consumption increased among 6- to 11-yearolds ( kcal; P ) and 12- to 19-year-olds ( kcal; P ). There was no significant change in calories from sports drinks or low-calorie SSBs for any age group. Among adults, calories from fruit drinks significantly decreased across all age groups (P < for all), while soda declined significantly among 20- to 39- year-olds ( kcal; P < 0.001) and 40- to 59-year-olds ( kcal; P < 0.001). There was no significant change in calories from sports drinks, low-calorie SSBs, or other SSBs. Supporting Information Table S4 presents the per capita consumption of soda calories stratified by age, gender, and race or ethnic 438 VOLUME 26 NUMBER 2 FEBRUARY

8 Figure 2 Per capita consumption of SSB calories among US children (aged 2 to 19) from 2003 to 2014, by SSB category. group. There was a significant decrease in soda calories for all age categories of NH white and Mexican American children as well as for NH black participants ages 12 to 19 years old. No significant declines in soda calories were observed for non-mexican Hispanic children. Within age-specific categories of adults, there was a significant decrease in soda calories for males aged 20 to 39 years and 49 to 59 years and females aged 20 to 39 years. Significant declines in calories from soda were observed for NH white participants aged 20 to 39 years and 40 to 59 years, NH black participants aged 40 to 59 years, and Mexican American participants aged 40 to 59 years. No significant declines in soda calories were observed for non-mexican Hispanic adults. Discussion From 2003 to 2014, the per capita consumption of all beverages declined significantly among children (from kcal to kcal; P < 0.001) and adults (from kcal to kcal; P < 0.001). This was primarily driven by a decline in the percentage of SSB drinkers and lower per capita consumption of SSBs. In the survey, 60.7% of children and 50.0% of adults drank SSBs on a given day, which is significantly lower than , when 79.7% of children and 61.5% of adults drank SSBs. From 2003 to 2014, the per capita consumption of SSBs declined by 92.3 calories (from kcal to kcal; P < 0.001) for children and Figure 3 Per capita consumption of SSB calories among US adults (aged 201) from 2003 to 2014, by SSB category. VOLUME 26 NUMBER 2 FEBRUARY

9 Trends in Beverage Consumption Bleich et al calories (from kcal to kcal; P < 0.001) for adults. This overall decline in both beverage and SSB consumption is consistent with previous literature, suggesting a recent turning point (23) toward lower energy intake in the US diet, potentially attributable to widespread discussion and media coverage of the role of certain foods (e.g., SSBs) in promoting obesity, changes to food allowances within the Special Supplemental Nutrition Program for Women, Infants and Children, improvements to school feeding programs, and product reformulations by food manufacturers and retailers. From an energy balance perspective, it is encouraging that children and adults are consuming fewer calories from beverages, as liquids have lower satiety and are less well compensated than calories from solid foods (24,25). The overall declines in beverage calories may help to explain the leveling of obesity in the general population and reductions among young children ages 2 to 5 years. Notably, in almost all years of the study, the percentage of drinkers and per capita consumption was higher for milk than for SSBs among younger children (aged 2-11 years). Over the period, milk was the largest source of per capita beverage calories among younger children (aged 2-11 years), and SSBs (mostly soda) were the largest source of beverage calories for adolescents and adults. For the years of the study in which water consumption was collected (2005 and onward), the percentage of drinkers was high and increased significantly over the period for all age groups. The prevalence of 100% juice consumption remained constant over time across all age categories of children. The continual monitoring of juice consumption trends remains important given the strengthening evidence base linking 100% juice consumption to weight gain, particularly among young children (26,27). The differences in SSB consumption trends by sociodemographic characteristics were also notable. While NH white adults experienced declines in SSB consumption across almost all age groups, there were no significant changes for any age category of Mexican American and non-mexican Hispanic participants or for NH black participants aged 20 to 39 years or 60 1 years. Furthermore, the absolute levels for the percentage of SSB drinkers and per capita consumption of soda remained highest among black, Mexican American, and non-mexican Hispanic adolescents (12-19 years old) for all years of the study, which is consistent with results from other papers (14,19,28) and important because these groups are also at higher risk for obesity and diabetes (10,29). Also like prior studies (14,19), we observed a considerable gradient by age, with young children and the elderly consuming the lowest quantity of SSBs and adolescents and young adults consuming the most. Although our results suggested that SSB consumption is declining overall, they also highlighted the need for reducing disparities in SSB consumption by race and ethnicity. Current public health efforts may be helpful for narrowing this gap. Key among them are beverage taxes, which are gaining momentum in the United States. In 2014, Berkeley, California passed the nation s first beverage tax. Last year, six jurisdictions successfully followed suit (Philadelphia, Pennsylvania; Boulder, Colorado; San Francisco, California; Oakland, California; Albany, California; and Cook County, Illinois), affecting roughly 8 million residents. Most recently, Seattle, Washington passed a beverage tax in June Several other jurisdictions are actively debating a beverage tax. In Berkeley, California and Mexico where evidence is available, the data suggests that the taxes have a larger impact among low-income households or neighborhoods (30,31). Targeted policies that aim to have a differentially large impact on reducing SSB consumption among groups at higher risk will also be an important strategy going forward. Another way to encourage greater consumption of healthier beverages (such as water) could be through procurement policies (32), which place restrictions on the types of beverages that can be made available for purchase in places such as schools, worksites, or government institutions. These healthy beverage procurement policies may have the added benefit of catalyzing the beverage industry to reformulate beverages to meet a heathier profile (such as flavored water rather than soda). These data provide the most recent national estimates for beverage consumption among children and adults in the United States (overall and by demographic characteristics) and analyze trends in beverage consumption from 2003 to Despite these strengths, this study had several limitations. Our reliance on single 24-hour dietary recalls may have introduced inaccuracy and bias to our analyses because of underreporting. However, evidence suggested better recall accuracy with packaged beverage items such as SSBs. It also did not consider the within-person variability in SSB consumption. Therefore, the results should be interpreted as representing intake on a given day rather than as usual intake over a longer period of time. The NHANES data are cross-sectional, which only allowed us to address associations rather than causality. These analyses did not control for some health behavior variables that may be associated with SSB consumption, such as smoking or physical activity. Conclusion Overall, beverage consumption declined for children and adults from 2003 to 2014, driven primarily by a decrease in the percentage of SSB drinkers and lower per capita consumption of SSBs. However, adolescents and young adults still consume more than the recommended amount of SSBs set by the Dietary Guidelines for Americans, and levels of SSB consumption are persistently highest among black, Mexican American, and non-mexican Hispanic individuals, who are also at higher risk for obesity. A notable positive trend is the relatively higher consumption of milk (compared to SSBs) among younger children and the increasing percentage of water drinkers among children and adults.o VC 2017 The Society References 1. Rosinger A, Herrick K, Gahche J, Park S. Sugar-sweetened beverage consumption among U.S. youth, NCHS data brief, no Hyattsville, MD: National Center for Health Statistics; Hu FB. Resolved: there is sufficient scientific evidence that decreasing sugarsweetened beverage consumption will reduce the prevalence of obesity and obesityrelated diseases. Obes Rev 2013;14: Malik VS, Popkin BM, Bray GA, Despres JP, Hu FB. Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation 2010; 121: Schulze MB, Manson JE, Ludwig DS, et al. Sugar-sweetened beverages, weight gain, and incidence of type 2 diabetes in young and middle-aged women. JAMA 2004;292: Vartanian LR, Schwartz MB, Brownell KD. Effects of soft drink consumption on nutrition and health: a systematic review and meta-analysis. Am J Public Health 2007;97: Gomez-Martınez S, Martın A, Romeo Marın J, et al. Is soft drink consumption associated with body composition? A cross-sectional study in Spanish adolescents. Nutr Hosp 2009;24: VOLUME 26 NUMBER 2 FEBRUARY

10 7. Jensen BW, Nielsen BM, Husby I, et al. Association between sweet drink intake and adiposity in Danish children participating in a long-term intervention study. Pediatr Obes 2013;8: Valente H, Teixeira V, Padrao P, et al. Sugar-sweetened beverage intake and overweight in children from a Mediterranean country. Public Health Nutr 2011;14: Kit BK, Fakhouri TH, Park S, Nielsen SJ, Ogden CL. Trends in sugar-sweetened beverage consumption among youth and adults in the United States: Am J Clin Nutr 2013;98: Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, JAMA 2014;311: United States Department of Agriculture. Scientific Report of the 2015 Dietary Guidelines Advisory Committee. Washington, DC: United States Department of Agriculture and Department of Health and Human Services; Fagherazzi G, Vilier A, Saes Sartorelli D, Lajous M, Balkau B, Clavel-Chapelon F. Consumption of artificially and sugar-sweetened beverages and incident type 2 diabetes in the Etude Epidemiologique aupres des femmes de la Mutuelle Generale de l Education Nationale-European Prospective Investigation into Cancer and Nutrition cohort. Am J Clin Nutr 2013;97: Garnett BR, Rosenberg KD, Morris DS. Consumption of soda and other sugarsweetened beverages by 2-year-olds: findings from a population-based survey. Public Health Nutr 2013;16: Han E, Powell LM. Consumption patterns of sugar-sweetened beverages in the United States. J Acad Nutr Diet 2013;113: Simon PA, Lightstone AS, Baldwin S, Kuo T, Shih M, Fielding JE. Declines in sugar-sweetened beverage consumption among children in Los Angeles county, 2007 and Prev Chronic Dis 2013;10:E131-E Shi L, van Meijgaard J. Substantial decline in sugar-sweetened beverage consumption among California s children and adolescents. Int J Gen Med 2010;3: Nielson SJ, Popkin BM. Changes in beverage intake between 1977 and Am J Prev Med 2004;27: Storey ML, Forshee RA, Anderson PA. Beverage consumption in the US population. J Am Diet Assoc 2006;106: Bleich SN, Wang YC, Wang Y, Gortmaker SL. Increasing consumption of sugarsweetened beverages among US adults: to Am J Clin Nutr 2009;89: Wang YC, Bleich SN, Gortmaker SL. Increasing caloric contribution from sugarsweetened beverages and 100% fruit juices among US children and adolescents, Pediatrics 2008;121:e1604-e Food and Nutrient Database for Dietary Studies Beltsville, MD: US Department of Agriculture, Agricultural Research Service; usda.gov/northeast-area/beltsville-md/beltsville-human-nutrition-research-center/ food-surveys-research-group/docs/fndds-download-databases/. Accessed April 10, Armstrong RA. When to use the Bonferroni correction. Ophthalmic Physiol Opt 2014;34: Ng SW, Slining MM, Popkin BM. Turning point for US diets? Recessionary effects or behavioral shifts in foods purchased and consumed. Am J Clin Nutr 2014;99: Mattes RD. Dietary compensation by humans for supplemental energy provided as ethanol or carbohydrate in fluids. Physiol Behav 1996;59: De Castro JM. The effects of the spontaneous ingestion of particular foods or beverages on the meal pattern and overall nutrient intake of humans. Physiol Behav 1993;53: Dennison BA, Rockwell HL, Baker SL. Excess fruit juice consumption by preschool-aged children is associated with short stature and obesity. Pediatrics 1997;99: Shefferly A, Scharf RJ, DeBoer MD. Longitudinal evaluation of 100% fruit juice consumption on BMI status in 2 5-year-old children. Pediatr Obes 2016;11: Park S, Blanck HM, Sherry B, Brener N, O Toole T. Factors associated with sugarsweetened beverage intake among United States high school students. J Nutr 2012; 142: Zhang Q, Wang Y, Huang ES. Changes in racial/ethnic disparities in the prevalence of Type 2 diabetes by obesity level among US adults. Ethn Health 2009;14: Cochero MA, Rivera-Dommarco J, Popkin BM, Ng SW. In Mexico, evidence of sustained consumer response two years after implementing a sugar-sweetened beverage tax. Health Aff (Millwood) 2017;36: Falbe J, Thompson HR, Becker CM, Rojas N, McCulloch CE, Madsen KA. Impact of the Berkeley excise tax on sugar-sweetened beverage consumption. Am J Public Health 2016;106: Niebylski ML, Lu T, Campbell NR, et al. Healthy food procurement policies and their impact. Int J Environ Res Public Health 2014;11: VOLUME 26 NUMBER 2 FEBRUARY