Table 1 Studies on the the overweight/obesity risk associated with SSB consumption

Beck, 2013

Mexican American children recruited from enrollees of Kaiser Permanente Health Plan of Northern California

319

8-10 years

Youth/ Adolescent FFQ

Increment of a serving/day of soda (1 serving = 240ml)

Odds of obesity

Positive

OR = 1.29 [95%CI: 1.13, 1.47]*

Bremer, 2010A

Nationally representative sample of U.S. adolescents, NHANES, 1988-1994, 1999-2004

1988-1994:

3234

1999-2004:

6967

12-19 years

Single 24-hour dietary recall interview

Increment of a serving/day of SSB (1 serving =250g)

Change in BMI percentile for age-sex

Mixed

Null for one follow-up

Positive for one follow-up

1988-1994

β = 0.38 [SE: 0.45]

1999-2004

β = 0.93 [SE: 0.18]*

Bremer, 2010B

Nationally representative sample of U.S. adolescents, NHANES, 1999-2004

6967

12-19 years

Single 24-hour dietary recall interview

Increment of a serving/day of SSB (1 serving =250g)

Change in BMI percentile for age-sex

Mixed

Positive in two sub-groups

Null in one sub-group

Non-Hispanic White:

β = 1.08 [SE: 0.21]*

Mexican-American:

β = 0.59 [SE: 0.29]*

Non-Hispanic Black:

β = 0.37 [SE: 0.26]

Clifton, 2011

Australian children as part of Australian National Children’s Nutrition and Physical Activity Survey

4400

2-16 years

Single 24-hour dietary recall interview

Consumed any amount of SSB in last 24 hours

Proportion of overweight or obese children who consumed SSBs vs. proportion of non-overweight children

Proportion of obese children who consumed SSBs compared to proportion of non-overweight children

Mixed

Null for one comparison

Positive for one comparison

Overweight and Obese vs. Normal Weight

50% vs. 47%

No measure of variation reported

Obese vs. Normal Weight

59% vs. 47%*

No measure of variation reported

Coppinger, 2011

British schoolchildren in south-west London, UK

248

9-13 years

Three day diary (Friday-Sunday)

mL/day of SSB

Correlation with BMI or BMI z-score

Null

No significant correlation [r= 0.05 for soft drinks and BMI, r=0.10 for fruit beverages]

Danyliw, 2012

Representative survey of Canadian children and adolescents

10,038

2-18 years

Single 24-hour dietary recall interview

Soft drink beverage cluster vs. moderate beverage pattern (mean beverage consumption in each cluster differed by gender and age group)

Odds of overweight-obesity

Mixed

Positive in one sub-group

Null in other sub-groups

Males, 6-11 years old

OR= 2.3 [95%CI: 1.2, 4.1] *

Females, 6-11 years old

OR = 0.8 [95%CI: 0.4, 1.7]

Males, 12-18 years old

OR = 0.7 [95%CI: 0.4-1.2]

Females 12-18 years old

OR: 1.1 [0.6, 1.9]

Davis, 2012

Low-income Hispanic toddlers from Los Angeles WIC program, 2008 data

1483

2-4 years

Interview about early-life feeding practices and nutritional intake

No SSB vs. High SSB (≥2 SSBs/day) (1 serving = 12 ounces)

Odds of obesity

Positive

OR= 0.69 [95%CI: 0.47, 1.00]*

Davis, 2014

Low-income Hispanic toddlers from Los Angeles WIC program, 2011 data

2295

2-4 years

Interview about early-life feeding practices and nutritional intake

No SSB vs. High SSB (≥2 SSBs/day), (1 serving = 12 ounces)

Odds of obesity

Positive

AOR = 0.72 [95%CI: 0.5, 1.0]*

Denova-Gutiérrez, 2009

Adolescent children of workers at two institutes and one university in Mexico

1055

10-19 years

Semi-quantitative FFQ

Increment of a serving/day of sweetened beverage (1 serving = 240mL)

Change in BMI

Odds of obesity

Positive

β =0.33 95%CI: 0.2, 0.5]*

OR=1.55 [95%CI: 1.32, 1.80]*

Gibson, 2007

Children in the UK part of the UK National Dietary and Nutritional Survey of Young People

1294

7-18 years

Seven day weighed food records

Top tertile of caloric soft drink intake (>396kJ/day)) vs. bottom tertile (<163kj/day)

Odds of overweight

Weakly Positive

OR=1.39 [95%CI: 0.96, 2.0]

Grimes, 2013

Nationally representative sample of Australian children

4283

2-16 years

Two 24-hour dietary recalls

More than one serving/day vs. less than one serving/day (1 serving = 250g)

Odds of overweight-obese

Positive

OR=1.26 [95%CI: 1.03, 1.53]*

Gómez-Martinez, ²⁰⁰⁹

Representative sample of urban Spanish adolescents

1523

13-18 years

Single 24-hour dietary recall

Non-consumers vs. moderate consumption (<336g/day) vs. high consumption (>336g/day) of sweetened soft drinks

Mean BMI

Null

No significant differences in BMI across SSB consumption groups

Ha, 2016

Combination of 5 studies conducted on Korean children between 2002 and 2011

2599

9-14 years

Three day dietary records

More than one serving/day vs. no SSB (1 serving = 200mL)

Odds of obesity

Mixed

Negative in one sub-group

Null in one sub-group

Males

OR: 0.52 [95%CI: 0.26, 1.05]*

Females

OR: 1.36 [95%CI: 0.62, 2.97]

Jiménez-Aguilar, 2009

Representative sample of Mexican adolescents who participated in Mexican National Health and Nutrition Survey

10,689

10-19 years

Semi-quantitative FFQ

Increment of a serving/day of soda (1 serving = 240ml)

Change in BMI

Mixed

Positive in one sub-group

Null in one sub-group

Males

β =0.17 [95%CI: 0.02, 0.32]*

Females

β =-0.07 [95%CI: -0.23, 0.10]

Note: these results are for soda. See full paper for fruit drinks, sugar beverages and SSBs.

Kosova, 2013

Nationally representative sample of U.S. children from NHANES, 1994-2004

4880

3-11 years

Single 24-hour dietary recall interview

Increment of a serving/day of SSB (1serving = 250g)

Change in BMI percentile

Mixed

Null overall and in some sub-groups

Positive in one sub-group

Overall

β =0.71

[SE=0.38]

3-5 year olds

β =-0.46 [SE=0.68]

6-8 year olds

β =0.19

[SE=0.65]

9-11 year olds

β =1.42

[SE=0.46]*

Linardakis, 2008

Children in public kindergartens in a single county in Greece

856

4-7 years

Three day weighed dietary records

High consumers (>250g/day) vs. non/low consumers of sugar-added beverage

Odds of obesity

Positive

OR= 2.35*

No measure of variation reported

Papandreou, 2013

Greek children in Thessaloniki

607

7-15 years

Three 24-hour dietary recalls

High consumers (>360mL/day) vs. low (<180mL/day) of SSBs

Odds of obesity

Positive

OR = 2.57 [95%CI: 1.06, 3.38]*

Schröder, 2014

Representative sample of Spanish adolescents

1149

10-18 years

Single 24-hour dietary recall

Soft drink beverage cluster (mean= 553g) vs. whole milk cluster

One-unit increase in BMI z-score

Positive

Males

OR = 1.29 [95%CI: 1.01, 1.65]*

Note: No soft drink cluster was identified for females

Valente, 2010

Elementary school children in Portugal

1675

5-10 years

Semi-quantitative FFQ

>2 servings/day (330mL) vs. less than 1 serving/day

Odds of overweight

Null

Males

OR: 0.64

[95%CI: 0.33, 1.52]

Females

OR: 0.63 [95%CI: 0.33, 1.22]

Ambrosini, 2013

Adolescent offspring from Australian Pregnancy Cohort (Raine) Study

1433

14 years old, followed-up at 17 years old

FFQ, at baseline and follow-up

Movement into top tertile of SSB consumption (>1.3 servings/day) at follow-up vs. remaining in lower SSB tertile

Odds of overweight-obesity at follow-up

Mixed

Null in one sub-group

Positive in one sub-group

Males:

OR: 1.2 [95%CI: 0.6, 2.7]

Females

OR: 4.8 [95%CI: 2.1, 11.4] *

Chaidez, 2013

Convenience sample of Latino mother and toddler pairs

67 mothers

1-2 years, followed-up for 6 months

Four 24-hour dietary recall (2 at baseline, 2 at follow-up)

High SSB consumption (higher than median) vs. low SSB consumption (lower than median)

BMI z-score, weight for height z-score, and weight for age z-score at follow-up

Mixed

Positive for one measure.

Null for other measures.

Weight for height z-score

β =0.46*

BMI z-score

β =0.47

Weight for age z-score

β =0.13

No measure of variation reported

DeBoer, 2013

Nationally representative sample of toddlers in the U.S.

9600

9 months, 2, 4 and 5 years (followed-up at each age)

Computer-assisted interview with questions about beverage consumption, at each follow-u

≥1 serving/day vs. <1 serving/day of SSB (1 serving = 8 ounces)

BMI z-score at follow-up (between 2 and 4 years and between 4 and 5 years)

Mixed

Measure of association not reported.

Positive for change between 2 and 4 years, null for change between 4 and 5 years.

Dubois, 2007

Representative sample of children in Quebec, Canada

1944

2.5, 3.5, 4.5 years (followed-up at each age)

Single 24-hour dietary recall and FFQ at each follow-up

Regular consumers (4-6 servings/week between meals) between ages 2.5 and 4.5 years vs. non-consumers of SSBs

Odds of being overweight at follow-up

Positive

OR: 2.36 [OR: 1.10, 5.05]*

Field, 2014

Children of participants in the Nurses’ Health Study 2 in the U.S.

7559

9-16 years, followed-up for 7 years

Youth/ Adolescent FFQ, at baseline and follow-up

Increment of baseline and change in sports drink serving/day (serving =1 can)

BMI score at follow-up

Mixed

Results differed depending on type of SSB and whether predictor was baseline intake or change in intake. Results below are for sports drink intake.

Females

Baseline:

β =0.29 [95%CI: 0.03, 0.54]*

Change:

β =0.05 [95%CI: =-0.19, 0.29]

Males:

Baseline:

β =0.33 [95%CI: 0.09, 0.58]*

Change:

β =0.43 [95%CI: 0.19, 0.66]*

Fiorito, 2009

Non-Hispanic white girls in the U.S.

170

5 years, assessed biennially until 15 years

Three 24-hour dietary recalls at each follow-up

≥2 servings of SSB/day vs. < 1 serving of SSB/day at age 5, (1 serving = 8 ounces)

Percentage overweight in each SSB consumption group at each follow-up

Positive

5 years old

≥2: 38.5%

<1: 16.1%

7 years old

≥2: 46.2%

<1: 15.1 %

9 years old

≥2: 46.2%

<1: 24.2%

11 years old

≥2: 53.9%

<1: 21.7%

13 years old

≥2: 46.2%

<1: 22.2

15 years old

≥2: 32.0

<1: 18.5

*Significant main effect

Jensen, 2013A

Danish children entering school in Copenhagen participating in intervention study

366

6, 9, 13 years (followed-up at each age)

7 day dietary record at 6 and 9 years

Increment of a serving/day of SSBs at 6 or 9 years, (1 serving = 100g)

Change in BMI from 6 to 9 years, 6 to 13 years or 9 to 13 years

Null

Intake at age 6, change from 6 to 9 years

β =-0.005 [95%CI:

-0.059, 0.0489]

Intake at age 6, change from 6 to 13 years

β =-0.059 [95%CI:

-0.145, 0.027]

Intake at age 9, change from 9 to 13 years

β =0.008 [95%CI:

-0.098, 0.113]

Note: these results are for SSBs. See full paper for sweet drinks and soft drinks separately.

Jensen, 2013B

Comparison groups of two quasi-experimental intervention studies in Australia (BAEW, IYM)

1465

4-18 years, followed-up approximately 2 years later

Asked participants how much SSB consumed yesterday or last school day

Increment of a serving/day of sweet drink at baseline, (1 serving = 100mL)

BMI z-score at follow-up

Null

BAEW study:

Β=0.005 [95%CI:

-0.003, 0.012]

IYM study:

β =0.004 [95%CI:

-0.002, 0.01]

Kral, 2008

Cohort of white children in U.S. born at different risks for obesity (based on maternal pre-pregnancy BMI)

49

3-6 years, followed-up at ages 3, 4, 5 and 6 years

Three day weighed food record

Change in calories from SSB from ages 3-5

Change in BMI z-score over follow-up

Null

Measure of association not reported

Laska, 2012

Adolescents enrolled in two longitudinal cohort studies in the U.S. (IDEA, ECHO)

693

6^th to 11^th grade, followed-up 2 years later

Three telephone-administered 24-hour dietary recalls

Increment of a serving/day (1 serving = not reported)

BMI at follow-up

Mixed

Positive in one sub-group

Null in one sub-group

Males

β =0.25 [SE: 0.10]*

Females

β =-0.09 [SE: 0.16]

Note: Above association was no longer significant when correcting for multiple testing

Laurson, 2008

Cohort of children in three rural U.S. states

268

10 years, followed-up for 18 months

Questionnaire asking about SSB consumption

SSB consumption (1 serving = not reported)

Spearman correlation with BMI at baseline or follow-up or change in BMI

Null

Males

Baseline

r= 0.009

Follow-up

r= 0.033

Change

r=0.041

Females

Baseline

0.073

Follow-up

0.077

Change

-0.033

Lee, 2015

Non-Hispanic Caucasian and African-American girls in the U.S.

2021

9-10 years, followed-up for 1 year

Three day food records

Increment of one teaspoon of added sugar (liquid form)

Change in BMI z-score at follow-up

Positive

β = 0.002 [95%CI: 0.001, 0.003)*

Leermakers, 2015

Dutch children in population-based prospective cohort study

2371

13 months, followed-up at ages 2, 3, 4 and 6

Semi-quantitative FFQ, validation against 24-hour recalls

High intake (15 servings/week) vs. low intake (3 servings/week) of sugar-containing beverages at 13 months, (1 serving = 150ml)

Change in BMI z-score at different follow-up ages

Mixed

Null in some sub-groups

Positive in other sub-groups

Males

2 year olds

β =-0.01 [95%CI: -0.15, 0.12]

3 year olds

β = -0.01 [95%CI: -0.15, 0.12]

4 year olds

β =0.01 [95%CI:

-0.12, 0.09]

6 year olds

β =0.05 [95%CI:

-0.08, 0.18]

Females

2 year olds

β =0.15 [95%CI: 0.01, 0.30]*

3 year olds

β =0.14 [95%CI: 0.01, 0.27]*

4 year olds

β =0.13 [95%CI: 0.01, 0.25]*

6 year olds

β =0.11 [0.00, 0.23]*

Libuda, 2008

German adolescents participating in longitudinal study (DONALD)

244

9-18 years, followed-up for 5-years

Three day weighed dietary records

Baseline and change in regular soft drink consumption

BMI z-score at follow-up

Null

Males

Baseline soft drink consumption

β =0.046

Change in baseline soft drink consumption

β =0.009

Females

Baseline soft drink consumption

β =-0.291

Change in baseline soft drink consumption

β =0.055

Measures of variation not reported

Lim, 2009

Low-income African-American children

365

3-5 years, followed-up for 2 years

Block Kids FFQ

Increment of an ounce/day of SSB at baseline

Odds of incidence of overweight at 2-year follow-up

Positive

OR=1.04 [95%CI: 1.01, 1.07]*

Millar, 2014

Nationally representative cohort of Australian children

4164

4-10 years, followed-up for 6 years

Parental interview asked about SSB consumption in past 24 hours

Increment of a serving/day (serving = not reported)

Change in BMI z-score at follow-up

Positive

β =0.015 [95%CI: 0.004, 0.025]*

Pan, 2014

Children in Infant Feeding Practices Cohort Study in U.S.

1189

10-12 months, followed-up at 6 years

Survey including questions about SSB consumption

Ever consumed SSBs vs. never consumed during infancy

High intake of SSBs (≥3 times/week) vs. no intake of SSBs during infancy

Odds of obesity at 6 years

Positive

Ever Consumed vs. Never consumed:

OR: 1.71

[95%CI: 1.09, 2.68]*

High vs. No SSBs

OR: 2.00 [95%CI: 1.02, 3.90]*

Vanselow, 2009

U.S. Adolescents from various socioeconomic and ethnic background in Minneapolis/St Paul metropolitan area

2294

Adolescents, followed-up for 5 years

Youth/ Adolescent FFQ

Stratified by different number of soft drinks serving/week (0, 0.5-6, ≥6)

Change in BMI over 5-year follow-up

Null

0 servings

β =1.74 [SEM= 0.18]

0.5-6 servings

β =1.92 [SEM=0.10]

≥7 servings

1.80 [SEM=0.15]

No significant differences across groups

Note: these results are for soft drinks. See full paper for punch, low-calorie soft drinks, etc.

Weijs, 2011

Dutch children

120

4-13 months, followed-up 8 years later

Two day dietary record

Beverage sugar intake per one percent of energy intake

Odds of overweight

Positive

OR: 1.13 [95%CI: 1.03, 1.24]*

Zheng, 2014

Danish children part of European Youth Heart Study

283

9 years, followed-at ages 15 and 21

24-hour dietary recall, supplemented by qualitative food record from same day, conducted at baseline and first follow-up

≥1 serving (12 ounces) vs. none at 9 years or 15 years

Increase in SSB serving from 9 to 15 years vs. no change

Change in BMI from 9 to 21 years or from 15 to 21 years

Mixed

Change in BMI from 9 to 21 years, using 9 years SSB as predictor

1.42 [SE: 0.68]

Change in BMI from 15 to 21 years, using 15 years SSB as predictor

0.92 [SE: 0.54]*

Change in BMI from 15 to 21 years, using

change in SSB from 9 to 15 years as predictor

0.91 [SE: 0.57]

Author, Year

Setting

Sample Size

Sample Age

Intervention

Control

Primary Outcome

Direction of Association

Findings

de Ruyter, 2012

Normal weight Dutch children

641

4-11 years

250mL sugar-free, artificially sweetened beverage

Similar sugar-containing beverage (104 calories)

Difference in change of BMI z-score from baseline at 18-month follow-up

Positive

-0.13 [95%CI:

-0.21, -0.05]*

Ebbeling, 2012

Overweight and obese adolescents in U.S. who reported consuming at least 12oz of SSB/day

224

Grade 9 or 10

1-year intervention designed to decrease SSB consumption

No beverage (given supermarket gift cards as retention strategy)

Difference in change of BMI z-score from baseline to 1 year and from 1 year to 2 years (Change in experimental group minus change in control group)

Mixed

1-year follow-up

-0.57 [SE: 0.28]*

2-year follow-up

-0.3 [SE: 0.40]

James, 2007

Longitudinal follow-up of children involved in intervention in United Kingdom

434

7-11 years

Discouraged children from consuming SSBs and provided one hour of additional health education during each of four school terms

No beverage

Odds of overweight at 1 year and 3-years after baseline intervention (intervention ended at 1 year)

Mixed

1-year follow-up

OR=0.58 [95%CI: 0.37, 0.89] *

3-year follow-up

OR=0.79 [95%CI: 0.52, 1.21]