April 2007, Vol 97, No. 4 | American Journal of Public Health Vartanian et al. | Peer Reviewed | Framing Health Matters | 667
FRAMING HEALTH MATTERS
In a meta-analysis of 88 studies, we examined the association between soft
drink consumption and nutrition and health outcomes. We found clear associa-
tions of soft drink intake with increased energy intake and body weight. Soft drink
intake also was associated with lower intakes of milk, calcium, and other nutri-
ents and with an increased risk of several medical problems (e.g., diabetes).
Study design significantly influenced results: larger effect sizes were observed
in studies with stronger methods (longitudinal and experimental vs cross-sectional
studies). Several other factors also moderated effect sizes (e.g., gender, age, bev-
erage type). Finally, studies funded by the food industry reported significantly
smaller effects than did non–industry-funded studies. Recommendations to re-
duce population soft drink consumption are strongly supported by the available
science. (Am J Public Health. 2007;97:667–675. doi:10.2105/AJPH.2005.083782)
Effects of Soft Drink Consumption on Nutrition and
Health: A Systematic Review and Meta-Analysis
| Lenny R. Vartanian, PhD, Marlene B. Schwartz, PhD, and Kelly D. Brownell, PhD
they displace other foods and beverages and,
hence, nutrients; whether they contribute to
diseases such as obesity and diabetes; and
whether soft drink marketing practices repre-
sent commercial exploitation of children.3–5
The industry trade association in the
United States (the American Beverage Associ-
ation, formerly the National Soft Drink Asso-
ciation) counters nutrition concerns with sev-
eral key points: (1) the science linking soft
drink consumption to negative health out-
comes is flawed or insufficient, (2) soft drinks
are a good source of hydration, (3) soft drink
sales in schools help education by providing
needed funding, (4) physical activity is more
important than food intake, and (5) it is unfair
to “pick on” soft drinks because there are
many causes of obesity and there are no
“good” or “bad” foods. Similar positions have
been taken by other trade associations such
as the British Soft Drinks Association and the
Australian Beverages Council.
Legislative and legal discussions focusing
on soft drink sales often take place on political
and philosophical grounds with scant atten-
tion to existing science. Our objectives were to
review the available science, examine studies
that involved the use of a variety of methods,
and address whether soft drink consumption
is associated with increased energy intake, in-
creased body weight, displacement of nutri-
ents, and increased risk of chronic diseases.
We focused on research investigating the
effects of sugar-sweetened beverages; diet
and artificially sweetened beverages are
noted only in certain cases for comparison
purposes. We conducted a computer search
through MEDLINE and PsycINFO using the
key terms “soft drink,” “soda,” and “sweetened
beverage.” We identified articles that assessed
the association of soft drink consumption with
4 primary outcomes (energy intake, body
weight, milk intake, and calcium intake) and
2 secondary outcomes (nutrition and health).
We identified additional articles by searching
each article’s reference section and the Web
of Science database. Finally, we contacted the
authors of each included article with a re-
quest for unpublished or in-press work, and
we asked each author to forward our request
to other researchers who might have relevant
work. Our searches yielded a total of 88
articles that were included in the present
There is a great deal of variability in re-
search methods in this literature. Studies
vary in their design (i.e., cross-sectional, longi-
tudinal, or experimental studies), sample
characteristics (e.g., male vs female, adults vs
children), and operational definitions of inde-
pendent and dependent variables. Because
such heterogeneity of research methods is
likely to produce heterogeneity of effect sizes
across studies (an effect size represents the
magnitude of the relationship between 2 var-
iables), we took 2 steps to assess the impact
of research method on outcome.
Initially, for each primary outcome (energy
intake, body weight, milk intake, and calcium
intake), we assessed the degree of heteroge-
neity of effect sizes by testing the significance
of the Q statistic, which is the sum of the
squared deviations of each effect size from
the overall weighted mean effect size. We did
not assess the degree of heterogeneity for
Soft drink consumption has become a highly
visible and controversial public health and
public policy issue. Soft drinks are viewed by
many as a major contributor to obesity and
related health problems and have conse-
quently been targeted as a means to help cur-
tail the rising prevalence of obesity, particu-
larly among children. Soft drinks have been
banned from schools in Britain and France,
and in the United States, school systems as
large as those in Los Angeles, Philadelphia,
and Miami have banned or severely limited
soft drink sales. Many US states have consid-
ered statewide bans or limits on soft drink
sales in schools, with California passing
such legislation in 2005. A key question is
whether actions taken to decrease soft drink
consumption are warranted given the avail-
able science and whether decreasing popula-
tion consumption of soft drinks would benefit
The issue is not new. In 1942 the Ameri-
can Medical Association mentioned soft
drinks specifically in a strong recommenda-
tion to limit intake of added sugar.1At that
time, annual US production of carbonated
soft drinks was 90 8-oz (240-mL) servings
per person; by 2000 this number had risen
to more than 600 servings.2In the interven-
ing years, controversy arose over several fun-
damental concerns: whether these beverages
lead to energy overconsumption; whether
American Journal of Public Health | April 2007, Vol 97, No. 4 668 | Framing Health Matters | Peer Reviewed | Vartanian et al.
FRAMING HEALTH MATTERS
secondary outcomes (nutrition and health)
because there were relatively few studies in
these domains. Our analysis of primary out-
comes revealed a significant degree of heter-
ogeneity of effect sizes in each case, and thus
we separated the studies according to re-
search design. This procedure reduces the
likelihood of aggregating effect-size estimates
across heterogeneous studies. Moreover,
some research designs are viewed as more
powerful than others. Cross-sectional studies
represent the weakest design, because such
studies cannot determine causality. Longitu-
dinal designs are considered stronger, but ex-
perimental designs are the strongest test of
causal relationships. Thus, separating studies
according to type of design allowed us to ex-
amine effect magnitudes as a function of
strength of research design.
We further explored variability in effect
sizes by examining a number of potential
moderator variables, including (1) population
studied (children and adolescents vs adults),
(2) gender of participants (only male, only fe-
male, or male and female combined), (3) type
of beverage (sugar-sweetened carbonated soft
drinks vs a mix of sugar-sweetened and diet
beverages), (4) whether the reported results
were adjusted for covariates (e.g., age, gender,
ethnicity, activity level), (5) assessment method
(self-reports vs observations or measure-
ments), and (6) presence or absence of food
industry funding. A study was coded as “in-
dustry funded” if the authors acknowledged
support from food companies, beverage com-
panies, or trade associations. Articles that did
not report a funding source or cited support
from other sources (e.g., pharmaceutical in-
dustry, university, foundation, or government
grants) were coded as “non–industry funded.”
We calculated average effect sizes (r values)
using Comprehensive Meta-Analysis version x2
(Biostat, Englewood, NJ). In most cases, we en-
tered data in the form in which they appeared
in each individual study, including group
means and standard deviations, correlation co-
efficients, t values, P values, and odds ratios and
confidence intervals. In certain cases, it was
necessary to manually calculate effect sizes. For
example, when means for more than 2 groups
were presented (e.g., low, moderate, and high
soft drink consumption), we used the formulas
for 1-way contrasts described by Rosenthal et
al.6In other cases, odds ratios were reported
with uneven confidence intervals (as a result of
rounding), and effect sizes were calculated di-
rectly from the odds ratio according to the
method described by Chinn.7
When data from different subgroups were
presented separately (e.g., data for male and
female participants were presented indepen-
dently), we calculated effect sizes separately
for each subgroup. In the case of studies that
reported multiple measures of a particular
construct (e.g., both body weight and body
mass index [BMI]), we computed the average
effect size of the reported measures. When
there was extraordinary variability in sample
sizes across studies, we employed the conser-
vative approach of limiting the sample size of
the largest study in a particular domain (e.g.,
cross-sectional studies of energy intake) to the
maximum sample size of the other studies in
that domain. This approach ensured that the
calculated average effect size would not be
dominated by a single study. We considered
an effect size of 0.10 or less as small, an ef-
fect size of 0.25 as medium, and an effect
size of 0.40 or above as large.8
To assess the presence of publication bias,
we computed a “fail-safe N” for each of the
main outcomes; this value is an estimate of
the number of unretrieved or unpublished
studies with null results that would be re-
quired to render the observed effect non-
significant. Rosenthal9suggested that a fail-
safe N greater than 5k + 10 (with k being the
number of studies included in the analysis)
indicates a robust effect; in the present analy-
ses, each fail-safe N far exceeded Rosenthal’s
recommendation, suggesting a low probability
of publication bias.
Soft Drink Consumption and Energy
The overall effect size (r) across all studies
for the relation between soft drink consump-
tion and energy intake was 0.16 (P<.001,
Q46=715.46, fail-safe N=9726). Because
there was a significant degree of heterogene-
ity among the effect sizes, we separated stud-
ies according to type of research design. Ef-
fect sizes for soft drink consumption and
energy intake are shown in Table 1.
Of the 12 cross-sectional studies examining
the relation between soft drink consumption
and energy intake, 10 reported a significant
positive association,10–191 reported mixed re-
sults,20and 1 reported no statistically signifi-
cant effect.21Two studies showed that the in-
crease in energy intake associated with soft
drink consumption was greater than what
could be explained by consumption of the
beverages alone,1 1,1 7suggesting that such bev-
erages might stimulate appetite or suppress
satiety, perhaps because of a high glycemic
index (foods with a high glycemic index pro-
duce a rapid rise in blood sugar).22The aver-
age effect size of the association between soft
drink consumption and energy intake across
all cross-sectional studies was 0.13 (P<.001;
The 5 longitudinal studies that we identi-
fied all reported positive associations between
soft drink consumption and overall energy
intake.1 7,23–26The average effect size for
these studies was 0.24 (P<.001; Q6=109.11,
Four long-term experimental studies in
which participants consumed soft drinks for
between 3 and 10 weeks showed that individ-
uals failed to compensate for the extra energy
consumed in the form of sugar-sweetened
beverages in that they did not reduce the rest
of their food energy intake, resulting in a
greater total daily energy intake.27–30One
study revealed that participants consumed
17% more energy than in their typical diet
even after the energy from the soft drinks
they consumed had been taken into ac-
count,27suggesting again that soft drinks may
influence other aspects of dietary intake. The
average effect size was 0.30 (P<.001;
Q4=2.37, P=.667). Because the Q statistic
was not statistically significant, we did not in-
vestigate moderators for long-term experi-
Findings from short-term experimental
studies (i.e., those examining energy intake
over the course of a subsequent meal or a
single day) were mixed. Of 12 studies, 5 re-
ported that individuals who consumed soft
drinks consequently took in a greater
amount of total energy (food energy plus
beverage energy) than did those who con-
sumed water.31–35One study also revealed
higher-than-expected energy intakes among
April 2007, Vol 97, No. 4 | American Journal of Public HealthVartanian et al. | Peer Reviewed | Framing Health Matters | 675
FRAMING HEALTH MATTERS
in Wyoming, Montana, and Idaho. Int J Obes Relat Metab
53. Nicklas TA, Yang S-J, Baranowski T, Zakeri I,
Berenson G. Eating patterns and obesity in children: the
Bogalusa Heart Study. Am J Prev Med. 2003;25:9–16.
54. Bes-Rastrollo M, Sanchez-Villegas A, Gomez-
Gracia E, Martinez JA, Pajares RM, Martinez-Gonzalez
MA. Predictors of weight gain in a Mediterranean co-
hort: the Seguimiento Universidad de Navarra Study 1.
Am J Clin Nutr. 2006;83:362–370.
55. Novotny R, Daida YG, Acharya S, Grove JS, Vogt
TM. Dairy intake is associated with lower body fat
and soda intake with greater weight in adolescent girls.
J Nutr. 2004;134:1905–1909.
56. Ludwig DS, Peterson KE, Gortmaker SL. Relation
between consumption of sugar-sweetened drinks and
childhood obesity: a prospective, observational analysis.
57. Drapeau V, Despres JP, Bouchard C, et al. Modifi-
cations in food-group consumption are related to long-
term body-weight changes. Am J Clin Nutr. 2004;80:
58. Phillips SM, Bandini LG, Naumova EN, et al.
Energy-dense snack food intake in adolescence: longi-
tudinal relationship to weight and fatness. Obes Res.
59. Blum JW, Jacobsen DJ, Donnelly JE. Beverage
consumption patterns in elementary school aged chil-
dren across a two-year period. J Am Coll Nutr. 2005;
60. Ebbeling CB, Feldman HA, Osganian SK, Chomitz
VR, Ellenbogen SJ, Ludwig DS. Effects of decreasing
sugar-sweetened beverage consumption on body
weight in adolescents: a randomized, controlled pilot
study. Pediatrics. 2006;117:673–680.
61. James J, Thomas P, Cavan D, Kerr D. Preventing
childhood obesity by reducing consumption of carbon-
ated drinks: cluster randomised controlled trial. BMJ.
62. Grandjean AC, Reimers KJ, Bannick KE, Haven MC.
The effect of caffeinated, non-caffeinated, caloric and
non-caloric beverages on hydration. J Am Coll Nutr.
63. Storey ML, Forshee RA, Anderson PA. Associations
of adequate intake of calcium with diet, beverage con-
sumption, and demographic characteristics among chil-
dren and adolescents. J Am Coll Nutr. 2004;23:18–33.
64. Frary CD, Johnson RK, Wang MQ. Children and
adolescents’ choices of foods and beverages high in
added sugars are associated with intakes of key nutri-
ents and food groups. J Adolesc Health. 2004;34:56–63.
65. Ballew C, Kuester S, Gillespie C. Beverage choices
affect adequacy of children’s nutrient intakes. Arch Pe-
diatr Adolesc Med. 2000;154:1148–1152.
66. Skinner JD, Carruth BR, Moran J III, Houck K,
Coletta F. Fruit juice intake is not related to children’s
growth. Pediatrics. 1999;103:58–64.
67. Fisher JO, Mitchell DC, Smiciklas-Wright H,
Birch LL. Maternal milk consumption predicts the
tradeoff between milk and soft drinks in young girls’
diets. J Nutr. 2000;131:246–250.
68. Ma D, Jones G. Soft drink and milk consumption,
physical activity, bone mass, and upper limb fractures
in children: a population-based case-control study.
Calcif Tissue Int. 2004;75:286–291.
69. Barr SI. Associations of social and demographic
variables with calcium intakes of high school students.
J Am Diet Assoc. 1994;94:260–266, 269.
70. Guthrie JF. Dietary patterns and personal charac-
teristics of women consuming recommended amounts
of calcium. Fam Econ Nutr Rev. 1996;9:33–49.
71. Skinner JD, Bounds W, Carruth BR, Ziegler P.
Longitudinal calcium intake is negatively related to
children’s body fat indexes. J Am Diet Assoc. 2003;
72. Salamoun MM, Kizirian AS, Tannous RI, et al.
Low calcium and vitamin D intake in healthy children
and adolescents and their correlates. Eur J Clin Nutr.
73. Fisher JO, Mitchell DC, Smiciklas-Wright H,
Mannino ML, Birch LL. Meeting calcium recommenda-
tions during middle childhood reflects mother-daughter
beverage choices and predicts bone mineral status. Am
J Clin Nutr. 2004;79:698–706.
74. Marshall TA, Eichenberger Gilmore JM, Broffitt B,
Stumbo PJ, Levy SM. Diet quality in young children is
influenced by beverage consumption. J Am Coll Nutr.
75. Whiting SJ, Healey A, Psiuk S, Mirwald R, Kowalski
K, Bailey DA. Relationship between carbonated and
other low nutrient dense beverages and bone mineral
content of adolescents. Nutr Res. 2001;21:1107–1115.
76. Curhan GC, Willett WC, Rimm EB, Spiegelman D,
Stampfer MJ. Prospective study of beverage use and
the risk of kidney stones. Am J Epidemiol. 1996;143:
77. Curhan GC, Willett WC, Speizer FE, Stampfer MJ.
Beverage use and risk for kidney stones in women.
Ann Intern Med. 1998;128:534–540.
78. Yoo S, Nicklas T, Baranowski T, et al. Comparison
of dietary intakes associated with metabolic syndrome
risk factors in young adults: the Bogalusa Heart Study.
Am J Clin Nutr. 2004;80:841–848.
79. Guerrero-Romero F, Rodriguez-Moran M, Reyes E.
Consumption of soft drinks with phosphoric acid as a
risk factor for the development of hypocalcemia in
postmenopausal women. J Clin Epidemiol. 1999;52:
80. Mazariegos-Ramos E, Guerrero-Romero F,
Rodriguez-Moran M, Lazcano-Burciaga G, Paniagua R,
Amato D. Consumption of soft drinks with phosphoric
acid as a risk factor for the development of hypocal-
cemia in children: a case-control study. J Pediatr. 1995;
81. Wyshak G. Teenaged girls, carbonated beverage
consumption, and bone fractures. Arch Pediatr Adolesc
82. Wyshak G, Frisch RE. Carbonated beverages, di-
etary calcium, the dietary calcium/phosphorus ratio,
and bone fractures in girls and boys. J Adolesc Health.
83. Wyshak G, Frisch RE, Albright TE, Albright NL,
Schiff I, Witschi J. Nonalcoholic carbonated beverage
consumption and bone fractures among women former
college athletes. J Orthop Res. 1989;7:91–99.
84. Heaney RP, Rafferty K. Carbonated beverages
and urinary calcium excretion. Am J Clin Nutr. 2001;
85. Smith S, Swain J, Brown EM, et al. A preliminary
report of the short-term effect of carbonated beverage
consumption on calcium metabolism in normal women.
Arch Intern Med. 1989;149:2517–2519.
86. Forshee RA, Storey ML. Evaluation of the associa-
tion of demographics and beverage consumption with
dental caries. Food Chem Toxicol. 2004;42:1805–1816.
87. Jensdottir T, Arnadottir IB, Thorsdottir I, et al.
Relationship between dental erosion, soft drink con-
sumption, and gastroesophageal reflux among Ice-
landers. Clin Oral Investig. 2004;8:91–96.
88. Marshall TA, Levy SM, Broffitt B, et al. Dental
caries and beverage consumption in young children.
89. Steinberg AD, Zimmerman SO, Bramer ML. The
Lincoln Dental Caries Study: II. The effect of acidu-
lated carbonated beverages on the incidence of dental
caries. J Am Dent Assoc. 1972;85:81–89.
90. Heller KE, Burt BA, Eklund SA. Sugared soda
consumption and dental caries in the United States.
J Dent Res. 2001;80:1949–1953.
91. Ismail AI, Burt BA, Eklund SA. The cariogenicity
of soft drinks in the United States. J Am Dent Assoc.
92. Hirvonen T, Pietinen P, Virtanen M, Albanes D,
Virtamo J. Nutrient intake and use of beverages and
the risk of kidney stones among male smokers. Am J
93. Krieger JN, Kronmal RA, Coxon V, Wortley P,
Thompson L, Sherrard DJ. Dietary and behavioral risk
factors for urolithiasis: potential implications for pre-
vention. Am J Kidney Dis. 1996;28:195–201.
94. Shuster J, Finlayson B, Scheaffer RL, Sierakowski R,
Zoltek J, Dzegede S. Primary liquid intake and urinary
stone disease. J Chronic Dis. 1985;38:907–914.
95. Shuster J, Jenkins A, Logan C, et al. Soft drink con-
sumption and urinary stone recurrence: a randomized
prevention trial. J Clin Epidemiol. 1992;45:911–916.
96. Soucie JM, Coates RJ, McClellan W, Austin H,
Thun M. Relation between geographic variability in
kidney stones prevalence and risk factors for stones.
Am J Epidemiol. 1996;143:487–495.
97. Savoca MR, Evans CD, Wilson ME, Harshfield GA,
Ludwig DA. The association of caffeinated beverages
with blood pressure in adolescents. Arch Pediatr Ado-
lesc Med. 2004;158:473–477.
98. Winkelmayer WC, Stampfer MJ, Willett WC,
Curhan GC. Habitual caffeine intake and the risk of hy-
pertension in women. JAMA. 2005;294:2330–2335.
99. Putnam JJ, Allshouse JE. Food Consumption, Prices
and Expenditures, 1970–1997. Washington, DC: Eco-
nomic Research Service, US Dept of Agriculture; 1999.
100. Flegal KM, Carroll MD, Ogden CL, Johnson CL.
Prevalence and trends in obesity among US adults,
1999–2000. JAMA. 2002;288:1723–1727.
101. Popkin BM, Nielsen SJ. The sweetening of the
world’s diet. Obes Res. 2003;11:1325–1332.
102. Bray GA, Nielsen SJ, Popkin BM. Consumption of
high-fructose corn syrup in beverages may play a role
in the epidemic of obesity. Am J Clin Nutr. 2004;79:
103. Bekelman JE, Li Y, Gross CP. Scope and impact
of financial conflicts of interest in biomedical research:
a systematic review. JAMA. 2003;289:454–465.
104. Levine J, Gussow JD, Hastings D, Eccher A. Au-
thors’ financial relationships with the food and beverage
industry and their published positions on the fat substi-
tute olestra. Am J Public Health. 2003;93:664–669.