NHANES is a series of complex, stratified, multistage probability surveys of the US civilian, noninstitutionalized population conducted by the National Center for Health Statistics, CDC. Participants in the NHANES surveys receive a detailed home interview followed by a physical examination and a dietary interview at a mobile examination center. The surveys were approved by the National Center for Health Statistics’s Ethics Review Board. Informed consent was obtained for persons aged ≥18 y. For those younger than 18 y, written parental consent was obtained and child assent was obtained for those 7–17 y. Data from NHANES 1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, and 2009–2010 were included in this analysis. Unweighted total examination response rates during the survey years ranged from 75% to 80% (21).

The type and quantity of all foods and beverages consumed in one 24-h period, specifically the 24-h period before the dietary-recall interview (midnight to midnight), were collected by trained interviewers with the use of a computer-assisted dietary interview system that included a multiple-pass format with standardized probes (22, 23). Since 2003–2004, two 24-h dietary recalls are publicly available in NHANES, but only one 24-h dietary recall was included in this analysis to maximize comparability between surveys. Furthermore, one 24-h recall is sufficient to estimate population means because the effects of random errors associated with dietary recall, including day-to-day variability, can be generally assumed to cancel out if days of the week are evenly represented (24). Estimates for frequency of consumption are also reported; however, because these estimates are calculated at the individual level using one 24-h recall, they are referred to as intake “on a given day.” Proxies, most commonly a parent, reported dietary intake for children aged ≤5 y and assisted with the dietary interview for children aged 6–11 y; dietary intake was self-reported for participants aged ≥12 y. The dietary intakes of each reported participant were assessed for reliability; dietary recalls deemed as unreliable, on the basis of criteria described elsewhere (25), are denoted as such on the public data files. Nutrient intakes, including calories, from all reported foods and beverages were calculated by using the USDA food-composition databases. In 1999–2000, the USDA 1994–1998 Survey Nutrient Database was the food-composition database used; in subsequent surveys, the USDA Food and Nutrient Database for Dietary Studies was used (26). In addition to changes in the food-composition database, between 1999 and 2002, there were other design changes incorporated to improve data quality, including changes to the automated data collection system and coding system (26). In 2002, the USDA and the National Center for Health Statistics integrated their dietary data collection efforts, and the dietary recall data are now collected in a joint effort.

The 2010 Dietary Guidelines for Americans define SSBs as those beverages that are sweetened with sugars that can add calories (13), including but not limited to sucrose, high-fructose corn syrup, and glucose. For the purposes of this report, the definition of SSBs includes soda, fruit drinks (including sweetened bottled waters and fruit juices and nectars with added sugars), sports and energy drinks, sweetened coffees and teas, and other SSBs (including horchata and sugar cane beverages), consistent with previous reports (18, 19). In this report, SSBs do not include diet drinks [defined as approximately <40 kcal/240 mL (8 oz) of the beverage (27)]; 100% fruit juice; beverages sweetened by the participant, including coffee and teas; alcohol; or flavored milks. These beverage categories were also excluded from previous analyses of SSB consumption (18, 19). Sports and energy drinks were combined into a single category, because, in several instances, the estimate for energy drinks was not statistically reliable. All reported beverages consistent with SSBs were identified by one author (BKK), and the classification strategy was reviewed by each of the other authors. Differences in classification were resolved by discussion. The focus of this analysis is on SSB consumption overall, but trends by SSB type are also assessed.

During the dietary interview, respondents were asked to identify both the location and eating occasion of each reported item. For this analysis, location of consumption was dichotomized as home and away from home, and eating occasion was categorized as meal (breakfast, lunch, or dinner) and snack, consistent with previous reports (28). Data for location of consumption was missing in <1% of SSB records; because a large portion of all food calories are consumed at home [~67% in 2009–2010 (29)], beverages missing a location of consumption were assumed to be consumed at home. In <1% of SSB records, the assigned eating occasion was not consistent with either meal or snack, most commonly recorded as “other,” and these beverages were excluded from the eating occasion analyses. The response options for location and eating occasion were coded slightly differently in 1999–2000 compared with subsequent surveys, but comparable categories were constructed.

Approximately one-third of participants reported consuming an SSB more than once during their 24-h recall. To examine whether the frequency of SSB consumption has changed, the number of SSBs reported for an individual was categorized as 0, 1, or ≥2.

Age was categorized as 2–5, 6–11, 12–19, 20–39, 40–59, or ≥60 y, consistent with NHANES sample-selection methods (30). Race-ethnicity was categorized as non-Hispanic white, non-Hispanic black, Mexican American, and other, based on self-reported race and ethnicity. Participants with a race-ethnicity categorized as other are included in overall estimates but are not separately reported.

All persons aged ≥2 y who participated in the examination component of NHANES were eligible for the dietary interview. In each of the surveys, between 4% and 6% of the participants were excluded because their recall did not meet standards of reliability (25). Pregnant or lactating women were not excluded. The final analytic sample, after the exclusion of persons with unreliable dietary data, consisted of 8074, 9033, 8273, 8549, 8529, and 9042 persons aged ≥2 y in 1999–2002, 2001–2002, 2003–2004, 2005–2006, 2007–2008, and 2009–2010. Detailed sample sizes are provided elsewhere (see Supplemental Tables 1 and 2 under “Supplemental data” in the online issue).

Statistical analyses were conducted by using SAS version 9.2 (SAS Institute Inc) and SUDAAN version 10.0 (RTI). Energy intake in this analysis is expressed as kcal; the conversion factor for converting kilocalories to kilojoules is 4.184. Mean kcal/d from SSBs were estimated and percentage of energy intake from SSB were calculated by using the PROC RATIO procedure in SUDAAN (31, 32). Percentage of energy intake from SSBs adjusts for differences in total energy intake, which facilitates comparisons across sex and age groups. Additional analyses examined mean SSB consumption among those who reported consuming a SSB. The percentages of persons reporting 0, 1, or ≥2 SSBs are reported. Dietary sample weights, which account for differential probabilities of selection, nonresponse, noncoverage, and sample design, were used to obtain estimates representative of the noninstitutionalized US population. The publicly available dietary sample weights include a poststratification step to balance recalls across days of the week. SEs were estimated by using Taylor series linearization. Statistical hypotheses were tested by using Student’s t statistic and an α level of 0.05. The hypothesis of no linear trend across the survey periods was tested by using orthogonal contrast matrices. No adjustments were made for multiple comparisons. Estimates for adults were age-adjusted to the 2000 Census population estimates by using the direct method and the following age groups: 20–39, 40–59, ≥60 y.

In 2009–2010, youth consumed a mean (±SE) of 155 ± 7 kcal/d from SSB, and adults consumed an age-adjusted mean (±SE) of 151 ± 5 kcal/d from SSBs—a decrease from 1999 to 2000 of 68 ± 14 and 45 ± 11 kcal/d, respectively (P-trend <0.001 for each) (Figure 1, A and B).

Over the 12-y study period, there were relatively large declines in soda consumption. Among youth and adults, soda consumption was 67 ± 11 and 58 ± 10 kcal/d lower in 2009–2010 than in 1999–2000 (P-trend <0.001 for both). Among youth and adults, fruit drink consumption was 16 ± 6.4 and 8 ± 3 kcal/d lower in 2009–2010 than in 1999–2000 (P-trend <0.001 for both). Consumption of sports and energy drinks increased over the 12-y period, but in 2009–2010 it continued to contribute a relatively small amount of energy intake from SSBs (12 ± 1.8 and 10 ± 0.9 kcal/d for youth and adults, respectively; P-trend <0.001 for both). Energy intake from sweetened coffee and teas increased over the study period (P-trend <0.05 and <0.001 for youth and adults, respectively). Beverages classified as “other” sweetened beverages were included in the overall estimates but are not displayed in Figure 1 (A and B) because they made only a small contribution to SSB intake, ranging from 0.7 ± 0.3 to 1.7 ± 0.4 kcal/d.

Among youth, energy intake from SSB decreased between 1999–2000 and 2009–2010 for each age group (P-trend ≤ 0.01 for each) (Table 1). The decreases in SSB were 37 ± 12, 71 ± 13, and 84 ± 23 kcal/d for youth aged 2–5, 6–11, and 12–19 y, respectively. Within age-specific groups, there was a decrease in energy intake from SSB over the 12 y for each sex and race-ethnicity, except among non-Hispanic black children aged 2–5 y. The percentage of energy intake from SSBs decreased among youth overall and in each age group across the study period (P-trend < 0.05 for each). Among non-Hispanic black youth aged 2–5 and 6–11 y and Mexican American youth aged 6–11 y, the change in percentage of energy intake from SSBs was not significant (P-trend ≥ 0.05).

Among adults, energy intake from SSBs decreased over the 12-y study period for those aged 20–39 and ≥60 y (P < 0.01 for both), but not for adults aged 40–59 y (Table 2). Among adults aged 20–39 y, energy intake from SSBs decreased among men and women and in each race-ethnicity group; however, among adults aged 40–59 y, only non-Hispanic blacks had a significant linear decrease over the 12-y study period. Among men aged 40–59 y, there was an increase in energy intake over the study period from SSB followed by a decrease (P-quadratic trend < 0.01). Among adults aged ≥60 y, men and non-Hispanic whites had a significant linear decrease in energy intake from SSBs.

Among adults, as observed for youth, the percentage of energy intake from SSBs decreased across the study period overall (P-trend < 0.001). Among adults aged 20–39 y, there was an overall decrease, and a decrease for both sexes and each examined race-ethnicity group, in the percentage of energy intake from SSBs. Among adults aged 40–59 y, the changes in percentage of energy intake from SSB were not significant except for a decrease in percentage of energy intake from SSB among non-Hispanic blacks (P < 0.001). Among adults aged ≥60 y, there was an overall decrease, and a decrease for men and each examined race-ethnicity group, in the percentage of energy intake from SSBs. The decrease in percentage of energy intake from SSBs among adults ≥60 y was generally smaller than that among adults aged 20–39 y.

Declines SSB consumption both at home and away from home contributed to the overall declines in energy intake from SSBs (Table 3). Among youth, energy intake from SSBs consumed at home and away from home were 35 ± 8 and 33 ± 8 kcal/d lower in 2009–2010 than in 1999–2000 (P-trend < 0.001 for both); among adults, age-adjusted energy intakes from SSBs consumed at home and away from home were 16 ± 9 and 29 ± 7 kcal/d lower in 2009–2010 than in 1999–2000 (P-trend < 0.01 for both). For both youth and adults, energy intake from SSBs consumed at lunch, at dinner, and as a snack decreased across the study period (P-trend < 0.01 for each), but no significant change in energy intake consumed at breakfast was observed.

In analyses restricted to those who reported consuming an SSB in the previous 24 h, mean energy intakes from SSBs were 288 ± 13, 268 ± 8, 278 ± 9, 271 ± 11, 234 ± 8, and 241 ± 10 kcal/d in 1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, and 2009–2010 for youth, respectively; mean age-adjusted energy intakes from SSBs were 325 ± 9, 321 ± 9, 306 ± 8, 296 ± 8, 276 ± 14, and 283 ± 5 kcal/d for adults during these time periods, respectively. For both youth and adults, the decrease in mean energy intake from SSBs (kcal/d) among individuals reporting SSB consumption was significant (P-trend, < 0.0.01 for both) (data not tabulated).

Among youth, in 2009–2010, more than one-third reported no SSB consumption (35.8% ± 1.3%), whereas 33.1% ± 1.3% and 31.2% ± 0.8% reported consumption of 1 SSB and ≥2 SSBs on a given day (Figure 2A). The percentage of youth reporting no SSB consumption increased (P-trend < 0.001) and of those reporting 1 SSB and ≥2 SSBs decreased (P-trend < 0.05 and < 0.001, respectively) between 1999–2000 and 2009–2010. An increase in the percentage of adults reporting no SSB consumption and a decrease in adults reporting consumption of ≥2 SSBs on a given day was observed (P-trend < 0.01 for both) (Figure 2B). Among adults, no significant change was observed over the 12-y study period in consumption of 1 SSB on a given day. For both youth and adults, a linear decrease in the consumption of one or more SSBs was observed (youth: 77.4% in 1999–2000 compared with 64.2% in 2009–2010; adults: 57.6% in 1999–2000 compared with 50.6% in 2009–2010; P < 0.001 for both, data not tabulated).