R

FAPM

ADsrsi((oalEhDadsty(wPbsmdmflbfAs

MSiCl

RU3u

A

3

RESEARCH

esearch and Professional Briefs

ive-Year Longitudinal and Secular Shifts indolescent Beverage Intake: Findings fromroject EAT (Eating Among Teens)-II

ELISSA C. NELSON, PhD, RD; DIANNE NEUMARK-SZTAINER, PhD, RD; PETER J. HANNAN, MStat; MARY STORY, PhD, RD

twsJ

Rcastc

iscHpwfaasdmwc

oic(baotldsals

MSP

BSTRACTetailed research examining concurrent longitudinal and

ecular changes in adolescent beverage intake is not cur-ently available, particularly since the year 2000. Thistudy’s objective was to evaluate these trends in beveragentake in a large, diverse adolescent cohort. Project EATEating Among Teens)-II is a 5-year longitudinal studyn�2,516) including two cohorts, which allows for thebservation of longitudinal changes from early to mid-dolescence (junior high to high school) and from mid- toate adolescence (high school to post high school). ProjectAT-II also examined secular trends in adolescentealth behavior from 1999-2004 in mid-adolescence.aily beverage servings were assessed using the Youthnd Adolescent Food Frequency Questionnaire. Longitu-inal findings indicate that intake of soda and sugar-weetened beverages (including soda, sweetened icedeas, and fruit drinks) increased significantly amongounger males, and alcohol increased across all groupsP�0.01). Consumption of certain beverages decreasedith age: fruit juice (among all males and older females,�0.02), milk (older adolescents, P�0.01), other milkeverages (all females and older males, P�0.01), dietoda (younger adolescents, P�0.01), and coffee/tea (allales and younger females, P�0.01). Significant secular

ecreases were observed in fruit juice and coffee/tea forales and females (P�0.05). Overall, these findings re-ect recent secular and longitudinal shifts in adolescenteverage consumption during the critical transition periodrom early to mid-adolescence and mid- to late adolescence.lthough additional research is needed to better under-tand nuances in adolescent consumption patterns, regis-

. C. Nelson is an assistant professor, D. Neumark-ztainer and M. Story are professors, and P. J. Hannan

s a senior research fellow, Division of Epidemiology &ommunity Health, University of Minnesota, Minneapo-

is.Address correspondence to: Melissa C. Nelson, PhD,D, Division of Epidemiology and Community Health,niversity of Minnesota, 1300 S 2nd St, WBOB Suite00, Minneapolis, MN 55454-1015. E-mail: nelson@epi.mn.eduManuscript accepted: July 2, 2008.Copyright © 2009 by the American Dietetic

ssociation.0002-8223/09/10902-0012$36.00/0

ldoi: 10.1016/j.jada.2008.10.043

08 Journal of the AMERICAN DIETETIC ASSOCIATION

ered dietitians and other health care practitioners workingith adolescents should address the importance of limiting

ugar-sweetened beverages with low nutrient density.Am Diet Assoc. 2009;109:308-312.

educing sugar-sweetened beverage consumption is ahighly targeted goal for obesity prevention andtreatment (1-3). The Dietary Guidelines for Ameri-

ans 2005 now include recommendations to limit bever-ges with added sugars and caloric sweeteners (4). De-pite this, adolescents and young adults have becomeargeted market segments for soft drink and fast-foodompanies in recent years (5,6).Previous studies have illustrated secular increases (ie,

ncreases over time) in sugar-sweetened beverage con-umption since the 1970s (7-9), with accompanying in-reased interest in youth-centered marketing (5,6,10-12).owever, little research has evaluated secular changes,articularly among adolescents, since the early 2000shen sugar-sweetened beverages started to become the

ocus of extensive public health scrutiny. Furthermore,lthough some research has illustrated longitudinal (ie,ge-related) trends in beverage intake, such as increasingoda and decreasing milk consumption with age (13,14),etailed research about beverage intakes during early,id-, and late adolescence is limited, particularly thathich examines concurrent secular and longitudinal

hanges.Project EAT (Eating Among Teens), a longitudinal ad-

lescent cohort study, provided the opportunity to exam-ne beverage trends over time, assessing trends occurringoncurrently due to age-related developmental changesie, following up with individuals who were assessed inoth early and mid-adolescence, or both mid- and latedolescence) and time (ie, assessing two different cohortsf mid-adolescents of similar age, one cohort in 1999 andhe other in 2004). Based on published research high-ighting the continued increasing prevalence of obesityuring the past 30 years (15), we hypothesized that re-ults would be similar to trends reported between 1970nd 2000 (7-9,13,14). Thus, we expected both secular andongitudinal increases in sugar-sweetened beverage con-umption over the 5-year study period.

ETHODStudy Designroject EAT examined diet and weight status among a

arge, ethnically diverse adolescent population. In EAT-I,

© 2009 by the American Dietetic Association

4nmcvt(o(2pyoays

Bgrbetfoswfsapi

pAsPt0amsrsaqaspPs

capwow

AMyh

attpiwsnmeatI

tptnQMot(dtoespHom

RLAtaafdAt0c(amm

SSerjsmcou

,746 junior and senior high school students in 31 Min-esota schools completed in-class surveys and anthropo-etric measures (1998-1999); all students in selected

lassrooms were eligible to participate. EAT-II resur-eyed participants by mail 5 years later (2003-2004) ashe younger cohort progressed from early adolescencejunior high) to mid-adolescence (high school), and thelder cohort progressed from mid- to late adolescencehigh school to post high school). Of the original sample,2.6% were lost to follow-up. Among remaining partici-ants, 2,516 completed surveys: 806 adolescents in theounger cohort (440 females/366 males) and 1,710 in thelder cohort (946 females/764 males). At Time 1, meange (�standard deviation) was 12.8�0.8 years for theounger and 15.8�0.8 years for the older cohort. Detailedtudy procedures are described elsewhere (16,17).The University of Minnesota’s Institutional Reviewoard approved study protocols. All EAT-I participantsave written assent before participation. Based on theequirements of the participating schools’ researchoards, parental consent was obtained by requiring par-nts to either: (a) return signed consent forms agreeing toheir child’s participation, or (b) return signed forms re-using their child’s participation (16). In EAT-II, parentsf adolescents younger than 18 years were first sent con-ent forms, and surveys were not sent to adolescentshose parents mailed back signed forms indicating re-

usal of their child’s participation. Adolescents were alsoent assent forms with the survey and were asked to signnd return the form if they were not interested in studyarticipation. Completion of the Project EAT-II surveymplied written assent.

Reported beverage intake was assessed once at each timeoint (in 1998-1999 and 2003-2004) using the Youth anddolescent Food Frequency Questionnaire, which has rea-onable reliability and validity in assessing youth diets.revious research indicated that mean correlations be-ween questionnaire responses and dietary recalls was.54 (18,19). Servings of the following beverages weressessed: fruit juice (eg, orange, apple, other), milk, otherilk beverages (eg, flavored milk, milkshakes), regular

oda, diet soda, sugar-sweetened beverages (includingegular soda, punch drinks, iced teas, but excludingweetened or flavored milk), alcohol (beer, wine, liquor),nd coffee/tea. The Youth and Adolescent Food Fre-uency Questionnaire allowed participants to write indditional foods on the questionnaire; 4% of this studyample had write-in values. A limited number of thesearticipants reported consuming Gatorade (PepsiCo, Inc,urchase, NY); these values were included as sugar-weetened beverages.Serving size was specified as: a can for soda and beer, a

up for tea/coffee, a drink/shot for liquor, and a glass forll other beverages. Four to seven response options wererovided for each beverage, specifying servings per day,eek, or month; for example, options for soda were: neverr �1/month, 1 to 3 cans/month, 1 can/week, 2 to 6 cans/eek, 1 can/day, or �2 cans/day.

nalysisixed model regressions (20) including a main effect for

ear (1999/2004), cohort (younger/older), and year by co-

ort interaction along with a random effect for individu- e

ls (accounting for longitudinal correlation) were used toest differences in means (ie, servings/week of beverageypes) across time, both within and across cohorts. Givenrevious research that suggests differences in beveragentake and trends by sex (14), sex-stratified analysesere used to examine the data. All mixed-model regres-

ions were adjusted for baseline race/ethnicity, socioeco-omic status (SES), and total energy intake. Age adjust-ent was also included in the mixed-model regression, so

stimates and tests for secular changes in high schooldolescents (1999-2004) would compare adolescents athe same mean age (16.5 years). SAS software (v9.1, SASnstitute Inc, Cary, NC, 2002) was used for analyses.

Longitudinal and secular trends were estimated usinghose individuals who had no missing data at both timeoints for the outcome being examined (ie, beverageype). From the full sample (n�2,516), 98 participants didot complete the Youth and Adolescent Food Frequencyuestionnaire and were excluded from these analyses.issing data for specific items ranged from 4.3% to 6.4%

f the longitudinal samples. The 2,418 participants inhis analysis represent 51% of the original cohortn�4,746) and 66% of those for whom valid mailing ad-resses were available (n�3,672). Because attrition inhe study population during the 5-year period did notccur at random, data were weighted to adjust for differ-ntial response rates in EAT-II using a response propen-ity method (21). The weighted ethnic/racial and SESroportions were: 52% white, 16% African American, 5%ispanic, 19% Asian, 4% Native American, and 4% mixed

r other race. SES was: low (16%), middle-low (19%),iddle (27%), middle-high (24%), and high (14%).

ESULTS AND DISCUSSIONongitudinal Trendscross this 5-year adolescent period, there were longi-

udinal shifts in beverage consumption (Table). Sodand sugar-sweetened beverages increased significantlymong younger males, (eg, increasing from 33% to 45%,rom 0.38 to 0.55 servings/day, and 0.94 to 1.25 servings/ay, respectively). Alcohol increased across all groups.mong younger females, sugar-sweetened beverage in-

ake was 0.86 servings/day while in middle school and.97 servings/day in high school (P�0.06). In contrast,onsumption decreased significantly for: fruit juiceamong all males and older females only), milk (olderdolescents), other milk beverages (all females and olderales), diet soda (younger adolescents), and coffee/tea (allales and younger females).

ecular Trendsecular changes in mid-adolescent (ie, high school) bev-rage intake were observed from 1999 to 2004 (Table,ight panels). Significant decreases were observed in fruituice and coffee/tea for high school males and females (eg,hifting from 1.26 to 1.06 fruit juice servings/day amongid-adolescent females in 1999 vs 2004). Significant de-

reases for diet soda and other milk beverages were alsobserved among high school females. No significant sec-lar changes were observed in milk, soda, or sugar-sweet-

ned beverages.

February 2009 ● Journal of the AMERICAN DIETETIC ASSOCIATION 309

ci1atcigdsesshsWefeAss

soas(

ifehisilea

fiomlo

ced tea

3

In summary, the aim of this study was to evaluateoncurrent longitudinal and secular trends in beveragentake among early, mid-, and late adolescents between999 and 2004. Although patterns in the observed trendsre complex in nature, three major findings emerge fromhis study and should be highlighted. These results in-lude the following: (a) sugar-sweetened beverage intakencreased longitudinally (ie, due to age) as youth pro-ressed through early adolescence, seeming to stabilizeuring the later transition from high school to post highchool; (b) adolescent intake of more nutrient-dense bev-rages (eg, fruit juice, milk) decreased with age; and (c)ecular trends (ie, trends related to time, rather than age)uggested a decrease in fruit juice consumption amongigh school youth from 1999 to 2004, whereas soda andugar-sweetened beverages showed no change over time.hereas previous research has suggested possible differ-

nces in beverage intake and trends by sex (14), findingsrom this study do not indicate clear differences in bev-rage patterns over time between males and females.lthough study findings support the hypothesis that ob-erved longitudinal changes in beverage intake would be

Table. Five-year longitudinal trends in early, mid-, and late adolescbeverage consumption (Project Eating Among Teens [EAT]-II, 1999-2

Younger Cohort

Earlyadolescence,1999 meanintake(servings/d)

Mid-adolescence,2004 meanintake(servings/d)

Pvalue

FemalesFruit juice 1.14 1.06 0.2Milk 0.96 0.87 0.1Other milk beverages 0.31 0.18 <0.0Soda 0.36 0.39 0.4Diet soda 0.15 0.07 <0.0Sugar-sweetened beveragesf 0.86 0.97 0.0Alcohol 0.003 0.04 <0.0Coffee/tea 0.28 0.19 <0.0MalesFruit juice 1.20 1.05 0.0Milk 1.22 1.21 0.9Other milk beverages 0.35 0.31 0.3Soda 0.38 0.55 <0.0Diet soda 0.13 0.04 <0.0Sugar-sweetened beveragesf 0.94 1.25 <0.0Alcohol 0.03 0.07 <0.0Coffee/tea 0.34 0.15 <0.0

aAll models are adjusted for propensity weights, age, ethnicity/race, socioeconomic statuto missing data. Younger cohort refers to the cohort transitioning from early adolescencthose transitioning from mid-adolescence (high school) in EAT-I to late adolescence (posthe older cohort is comprised of 946 females and 764 males.bP value for longitudinal trend in younger cohort (testing differences between early andcP value for longitudinal trend in older cohort (testing differences between mid- and latdP value for secular trend in mid-adolescence (testing differences between 1999 and 2eValues in boldface type denote statistically significant differences (P�0.05).fSugar-sweetened beverages included soda (regular only, not diet), punch drinks, and i

imilar to those previously reported (13,14), they do not f

10 February 2009 Volume 109 Number 2

upport the a priori hypothesis around secular trends. Inther words, secular changes in sugar-sweetened bever-ge consumption over the 5-year study period were notimilar to those increases reported over the past 30 years7-9).

These findings indicate that sugar-sweetened beveragentake longitudinally increased 33% in males and 13% inemales during the middle to high school years. Sweet-ned beverage intake remained at this higher level afterigh school, showing no additional increase with age. It is

nteresting to note that nondiet soda generally repre-ented less than half of total sugar-sweetened beveragentake during this early adolescent period, thus high-ighting the importance of other nonsoda sugar-sweet-ned beverages (eg, sweetened teas, fruit drinks) at thisge.Consistent with previous research (13,14,22-24), these

ndings suggest age-related decreases, as much as 33%,f more nutrient-dense beverages, such as fruit juice andilk, among males and females. Current national guide-

ines (25) recommend that teenagers consume 3 cups/dayf low-fat milk (or equivalent products) and 2 cups/day of

everage consumption, and 5-year secular trends in mid-adolescenta

Older Cohort Secular Trend

id-dolescence,999 meantakeervings/d)

Lateadolescence,2004 meanintake(servings/d)

Pvaluec

Mid-adolescence:2004-1999 meanchange(servings/d)

Pvalued

.26e 0.97 <0.01 �0.20 <0.01

.99 0.72 <0.01 �0.13 0.11

.24 0.17 <0.01 �0.06 0.05

.46 0.42 0.15 �0.06 0.16

.15 0.16 0.66 �0.07 0.03

.99 0.93 0.14 �0.02 0.72

.03 0.14 <0.01 0.004 0.74

.29 0.29 0.84 �0.10 <0.01

.21 0.98 <0.01 �0.16 0.03

.28 0.86 <0.01 �0.07 0.44

.36 0.18 <0.01 �0.05 0.24

.61 0.56 0.16 �0.05 0.31

.10 0.10 0.98 �0.05 0.06

.23 1.20 0.59 0.02 0.76

.07 0.23 <0.01 �0.01 0.66

.27 0.22 <0.01 �0.12 <0.01

total caloric intake (based on self-report from Time 1) (17). Model sizes vary slightly dueor high) in EAT-I to mid-adolescence (high school) in EAT-II, and older cohort refers toschool) in EAT-II. The younger cohort is comprised of 440 females and 366 males, and

dolescence).scence)., among mid-adolescent females fruit juice decreased by 0.2 servings/day, P�0.002).

s.

ent b004)

b

Ma1in(s

0 17 01 00 01 06 01 01 0

2 12 12 01 01 01 11 01 0

s, ande (junit high

mid-ae adole004, eg

ruit (which can include 100% fruit juice). Although fruit

jcbmhac

idfaltnastficfltiebbpat

a2gtaHiaottpmpcTnaqFt

CTlsefcrcmsda

ccTaas1eletilhaatf

TtStdv

R

uice and milk do contain calories and thus should not beonsumed in excess, they confer important nutritionalenefits. Decreases in the consumption of these beveragesay have important implications for adolescents’ futureealth by potentially increasing the risk of diseases suchs such as osteoporosis, cardiovascular disease, and can-er (4).

In addition to the longitudinal changes in beveragentake illustrated by these findings, significant secularecreases were also observed in fruit juice consumptionor the two cohorts of similarly aged high school malesnd females between the years 1999-2004. Despite highevels of soda and sugar-sweetened beverage consump-ion (ie, approximately 1 serving/day), these analyses didot detect significant secular increases in these bever-ges. This is in contrast to previous research that hashown large, secular increases in soda consumption be-ween the late 1970s and 2001 (7,8). It is possible that ourndings reflect a preliminary leveling-off of soft drinkonsumption in the United States, potentially also re-ecting early market reports that soda sales have beguno plateau (26). However, it is also possible that with thencreasing variety in sugar-sweetened beverages (eg, en-rgy drinks, sport beverages) and serving sizes, it hasecome increasingly difficult for adolescents to reporteverage intake, leading to misreporting and/or underre-orting. It is a challenge for researchers to develop validnd reliable food frequency assessment tools that reflecthe most current market offerings.

In recent years, sugar-sweetenedbeverage consumption has begun toreceive extensive public attention,especially as a potential target for

obesity prevention.

This population-based study is among the first to ex-mine trends in beverage consumption patterns since001 and is unique in its examination of concurrent lon-itudinal and secular changes, providing important dis-inctions between the transitions of early, mid-, and latedolescence, which are often not examined separately.owever, this study has several limitations. This study

ncluded two measurement points over a 5-year period;lthough this design allows for the evaluation of changever time, additional measurement points would enhancehis work. In addition, whereas validated food frequencyools are a well-recognized dietary assessment method foropulation studies, it can be difficult to accurately esti-ate serving sizes and intake frequency given the options

rovided. Serving size estimation may be particularlyhallenging given recent increases in portion sizes (27).he Youth and Adolescent Food Frequency Question-aire did not allow for the collection of data on all bever-ges consumed, particularly important given the fre-uent appearance of new beverages in the marketplace.inally, these data do not provide information about wa-

er intake, which may be replacing other beverages.

ONCLUSIONShese results illustrate shifts that have occurred in ado-

escent beverage consumption. In recent years, sugar-weetened beverage consumption has begun to receivextensive public attention, especially as a potential targetor obesity prevention. Given environmental and policyhanges about sugar-sweetened beverage consumptionesulting from this recent attention, such as reduced ac-ess to soft drinks in schools, future research is needed toonitor trends in beverage intake over time. It is pos-

ible that such public health initiatives to limit softrink consumption may impact population-wide bever-ge trends.The transition through adolescence is marked by in-

reasing independent food purchasing decisions and ac-essibility to a wide array of foods outside of the home.his is a critical age at which food choices need to beddressed. The findings from this study indicate thatdolescents are consuming approximately one sugar-weetened beverage each day, likely representing 100 to50 calories/serving. This may translate into 700 to 1,000xcess calories per week. Because these beverages offerittle nutritional value, sweetened beverages may be anasy and important target for caloric reduction. Regis-ered dietitians and other health care practitioners work-ng with adolescents should address the importance ofimiting intake of sugar-sweetened beverages, whichave a low nutrient density. In addition, practitionersnd clinicians may need to work closely with adolescentsnd young adults to help them to incorporate more nu-rient-dense beverages, such as 100% fruit juice and low-at milk, in a healthy, energy-balanced diet.

his study was supported by grant R40 MC 00319 fromhe Maternal and Child Health Bureau (Title V, Socialecurity Act), Health Resources and Services Adminis-ration, Department of Health and Human Services. Ad-itional salary support was provided by the Obesity Pre-ention Center at the University of Minnesota.

eferences1. Harnack L, Stang J, Story M. Soft drink consumption among US

children and adolescents: Nutritional consequences. J Am Diet Assoc.1999;99:436-441.

2. Popkin BM, Armstrong LE, Bray GM, Caballero B, Frei B, WillettWC. A new proposed guidance system for beverage consumption inthe United States. Am J Clin Nutr. 2006;83:529-542.

3. Bray GA, Nielsen SJ, Popkin BM. Consumption of high-fructose cornsyrup in beverages may play a role in the epidemic of obesity. Am JClin Nutr. 2004;79:537-543.

4. US Department of Health and Human Services and US Department ofAgriculture. Dietary Guidelines for Americans, 2005. 6th Edition.Washington, DC: US Government Printing Office; 2005.

5. Powell LM, Szczypka G, Chaloupka FJ. Adolescent exposure to foodadvertising on television. Am J Prev Med. 2007;33(suppl):S251-S256.

6. Nelson MC, Story M, Larson NI, Nuemark-Sztainer D, Lytle LA.Emerging adulthood and college-aged youth: An overlooked age forweight-related behavior change. Obesity. 2008;16:2205-2211.

7. French SA, Lin BH, Guthrie JF. National trends in soft drink con-sumption among children and adolescents age 6 to 17 years: Preva-lence, amounts, and sources, 1977/1978 to 1994/1998. J Am DietAssoc. 2003;103:1326-1331.

8. Nielsen SJ, Popkin BM. Changes in beverage intake between 1977and 2001. Am J Prev Med. 2004;27:205-210.

9. Rajeshwari R, Yang SJ, Nicklas TA, Berenson GS. Secular trends in

children’s sweetened-beverage consumption (1973 to 1994): The Bo-galusa Heart Study. J Am Diet Assoc. 2005;105:208-214.

February 2009 ● Journal of the AMERICAN DIETETIC ASSOCIATION 311

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

3

0. Institute of Medicine. Food marketing to children and youth: Threator opportunity? Washington, DC: National Academy of Sciences;2005.

1. Gamble M, Cotugna N. A quarter century of TV food advertisingtargeted at children. Am J Health Behav. 1999;23:261-267.

2. Hawkes C. Marketing activities of global soft drink and fast foodcompanies in emerging markets: A review. In: Globalization, Diets,and Non-Communicable Diseases. Geneva, Switzerland: WorldHealth Organization; 2002.

3. Rampersaud GC, Bailey LB, Kauwell GP. National survey beverageconsumption data for children and adolescents indicate the need toencourage a shift toward more nutritive beverages. J Am Diet Assoc.2003;103:97-100.

4. Cavadini C, Siega-Riz AM, Popkin BM. US adolescent food intaketrends from 1965 to 1996. Arch Dis Child. 2000;83:18-24.

5. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, FlegalKM. Prevalence of overweight and obesity in the United States, 1999-2004. J Am Med Assoc. 2006;295:1549-1555.

6. Neumark-Sztainer D, Wall M, Guo J, Story M, Haines J, Eisenberg M.Obesity, disordered eating, and eating disorders in a longitudinalstudy of adolescents: How do dieters fare 5 years later? J Am DietAssoc. 2006;106:559-568.

7. Neumark-Sztainer D, Story M, Hannan PJ, Croll J. Overweight sta-tus and eating patterns among adolescents: Where do youths stand incomparison with the healthy people 2010 objectives? Am J PublicHealth. 2002;92:844-851.

8. Rockett HR, Breitenbach M, Frazier AL, Witschi J, Wolf AM, Field

AE, Colditz GA. Validation of a youth/adolescent food frequency ques-tionnaire. Prev Med. 1997;26:808-816.

2

12 February 2009 Volume 109 Number 2

9. Rockett HR, Wolf AM, Colditz GA. Development and reproducibilityof a food frequency questionnaire to assess diets of older children andadolescents. J Am Diet Assoc. 1995;95:336-340.

0. Diggle P, Heagerty P, Liang K, Zeger S. Analysis of longitudinal data.2nd ed. New York, NY: Oxford University Press; 2002.

1. Little R. Survey nonresponse adjustments for estimates of means. IntStat Rev. 1986;54:137-139.

2. Demory-Luce D, Morales M, Nicklas T, Baranowski T, Zakeri I, Be-renson G. Changes in food group consumption patterns from child-hood to young adulthood: The Bogalusa Heart Study. J Am Diet Assoc.2004;104:1684-1691.

3. Sichert-Hellert W, Kersting M, Manz F. Fifteen year trends in waterintake in German children and adolescents: results of the DONALDStudy. Dortmund Nutritional and Anthropometric Longitudinally De-signed Study. Acta Paediatr. 2001;90:732-737.

4. Fulgoni V 3rd, Nicholls J, Reed A, Buckley R, Kafer K, Huth P,Dirienzo D, Miller GD. Dairy consumption and related nutrient in-take in African-American adults and children in the United States:Continuing Survey of Food Intakes by Individuals 1994-1996, 1998,and the National Health And Nutrition Examination Survey 1999-2000. J Am Diet Assoc. 2007;107:256-264.

5. MyPyramid.gov: Steps to a Healthier You. United States Departmentof Agriculture. MyPyramid.gov Web site. http://www.mypyramid.gov.Accessed February 14, 2008.

6. QSR Web. Beverage trends impact QSR [Quick Service Restaurant]sales. QSR Web site. http://www.qsrweb.com/article.php?id�8033.Accessed May 1, 2008.

7. Nielsen SJ, Popkin BM. Patterns and trends in food portion sizes,1977-1998. JAMA. 289:450-453.