Table 5

 Summary of prospective cohort studies examining association between free sugars exposures and adiposity in adults

First author (year), country, study namePopulation and recruitmentFollow-up (years)Exposure and assessment methodOutcome and assessment methodCovariates and stratificationFindings*+/0/−†
Bes-Rastrollo (2010),59 Spain, SUN Project10 162 Spanish university graduates; convenience sampling4.4SSB intake separated into thirds; FFQWeight change (kg) (continuous), weight gain ≥3 kg (OR), incident obesity (HR); self reportedAdjusted for age, alcohol intake, baseline BMI, dietary intake, physical activity, sex, sitting, smoking, total energy intake, TV viewingConsumption of SSB was associated with greater weight gain in the group with the highest intake relative to the group with the lowest. However, there was no association with risk of developing obesity+/0
Colditz (1990),76 USA, Nurses’ Health Study31 940 married registered female nurses aged 30-55 years; consecutive sampling4Sucrose (g/day); FFQWeight gain (kg) (continuous); self reportedAdjusted for age, baseline BMI, total energy intakeSucrose intake not significantly associated with weight gain at follow-up0
Dhingra (2007),60 USA, Framingham Offspring Study4028 middle aged adults, mean age 51-56 years, whose parents were in the Framingham Heart Study; random sampling used in original Framingham Heart Study cohort4SSSD (0, <1, ≥1, ≥2 servings/day); questionnaireIncident obesity (BMI ≥30) and incident high WC (men ≥102 cm, women ≥88 cm) (OR); measuredAdjusted for age, baseline BMI/WC, dietary intake, glycaemic index, physical activity, sex, smoking, total energy intakeCompared with no SSSD intake, SSSD consumption was associated with incident obesity (1 serving/day: OR 1.21 (95% CI 0.90 to 1.62); ≥1: 1.31 (1.02 to 1.68); ≥2: 1.50 (1.06 to 2.11)). SSSD consumption was significantly associated with developing a high WC across the intake categories (1: 1.25 (1.02 to 1.54); ≥1: 1.40 (1.08 to 1.83); ≥2: 1.30 (1.09 to 1.56))+
Drapeau (2004),61 Canada, Québec Family Study248 adults aged 18-65 years, living within 80 km radius of Québec; convenience sampling5.9Self perceived change (increase, maintenance, or decrease) in the intake of sugar, sweet foods, and SSSD; questionnaireChange in weight (kg), BF (%), sum of 6 skinfolds (cm), WC (cm) (all continuous); measuredAdjusted for age, body weight indicators at baseline, change in dietary intake, change in physical activityIn unadjusted ANCOVA analyses, participants reporting an increase in intake of sugar/sweet foods had a significantly higher increase in WC and sum of 6 skinfolds than those reporting a decrease (P<0.05), while no differences were seen for weight change or BF (%). No differences were seen for any body weight indicators after an increase in SSSD consumption. In adjusted regression analyses, an increase in consumption of sugar/sweet foods was significantly associated with an increased in WC (0.16 cm) and sum of 6 skinfolds (1.62 cm; both P=0.03)+/0
French (1994),62 USA, Healthy Worker Project1639 working men (mean age 39.1 years, SD 9.8) and 1913 women (mean age 37.3 years, SD 0.7) participating in an intervention study of smoking cessation and obesity prevention; convenience sampling of worksites, random sampling of workers2Sweet foods and SSSD (servings/week); FFQWeight change (in pounds) (continuous); measuredAdjusted for dieting behaviour at baseline and follow-up, baseline intake of sweets/SSSD, baseline weight, education, intervention group, marital status, occupation, smoking, worksite; stratified by sexConsumption of sweet foods was associated with weight gain over time in both men and women. Each serving/week of sweet foods at baseline was associated with an weight increase of 0.28 pounds (0.13 kg, SE 0.04 kg) and 0.19 pounds (0.09 kg, 0.04 kg) in women and men, respectively, over the 2 years (both P≤0.02). No significant association was seen between SSSD intake and weight gain+
Halkjaer (2004),63 Denmark, MONICA2275 Danish adults aged 30, 40, 50, or 60 years; random sampling6Intake of sweet foods divided into fifths; FFQChange in WC (cm) (continuous); measuredAdjusted for age, alcohol intake, baseline BMI and current BMI, baseline hip circumference, baseline WC, change in dietary intake, education, physical activity, smoking; stratified by sexNo relation was seen between sweet foods consumption and WC0
Halkjaer (2006 and 2009),64 65 Denmark, Danish Diet, Cancer and Health Study20 126 men and 22 570 women aged 50-64 years, living in greater Copenhagen or Aarhus areas; consecutive sampling5.3Jams, sugars and syrups (250 kJ/day), SSSD (250 kJ/day), and energy intake from foods with added sugar (MJ/day); FFQChange in WC (cm/5 years) (continuous); measured at baseline, self reported at follow-upAdjusted for age, alcohol intake, baseline BMI and WC, energy intake from other sources, physical activity, smokingIntake of jams, syrups, sugars, or SSSD was not associated with change in WC in women or men. In women only, energy intake from foods with added sugar at baseline was significantly associated with a 5 year increase in WC (per MJ/day: 0.39 cm increase (95% CI 0.18 to 0.60); P<0.001)+ (women only)
Hendriksen (2011),66 Holland, MORGEN-EPIC11 111 adults aged 20-64 years, living in Amsterdam, Maastricht, or Doetinchem; random sampling4.9/9.9Sweets (418 kJ/day) and cakes (418 kJ/day) intake; FFQ Annual weight change (g/year) (continuous); self reportedAdjusted for age, baseline weight and height, dietary intake, duration of follow-up, education, physical activity, sex, smoking, SSSD intake, total energy intakeNo association between weight change and consumption of sweets or cakes was seen0
Kvaavik (2004),67 Norway, Oslo Youth Study371 Norwegian adults aged 23-27 years who participated in a school based intervention study 10 years previously; convenience sampling8Long term intake of SSSD (low consumption (<3 times/week at first and second follow-up), high consumption (≥3 times/week at first and second follow-up), or inconsistent consumption); questionnairePrevalent overweight (BMI ≥25) or obesity (BMI ≥30) at second follow-up (OR); self reportedAdjusted for baseline (adolescent) BMI; stratified by sexNo associations were seen between long term consumption of SSSD and overweight or obese status0
Mozaffarian (2011),68 USA, Nurses’ Health Study I and II, Health Professionals Follow-up Study50 422 married registered female nurses aged 30-55 years, 47 898 registered female nurses aged 24-44 years, and 120 877 male health professionals aged 40-75 years; consecutive sampling4Increase in consumption (servings/day) of SSB, FJ, sweets, or desserts; FFQWeight gain (in pounds) over 4 year periods (continuous); self reportedAdjusted for age, baseline BMI, change in alcohol intake, change in dietary intake, change in physical activity, change in smoking, change in TV viewing, sleep durationIn pooled analysis, each serving/day increase in SSB intake was associated with a 1 pound increase in weight (0.45 kg (95% CI 0.36 to 0.53); P<0.001). Each serving/day increase in FJ was associated with a 0.31 pound increase in weight (0.14 kg (0.06 to 0.21); P<0.001). Each serving/day increase in sweets or desserts was associated with a 0.41 pound increase in weight (0.19 kg, 0.07 to 0.30; P<0.001)+
Nooyens (2005),69 Holland, Doetinchem Cohort Study288 men aged 50-60 years, attending a municipal health clinic; consecutive sampling5Change in intake of SSSD (glasses/day); FFQChange in body weight (kg/year) and change in WC (cm/year) (both continuous); measuredAdjusted for age, alcohol intake, dietary intake, occupation, physical activity, retirement status, smoking, total energy intakeIn unadjusted analyses, an increase in SSSD intake from baseline to follow-up was positively associated with change in body weight and WC (per glass/day increase, β 0.2 and 0.16, respectively; both P≤0.04). After adjustment, associations were rendered non-significant (both P≥0.05)+/0
Odegaard (2010),70 Singapore, Singapore Chinese Health Study43 580 Hokkien and Cantonese speaking Singaporeans aged 45-74 years, residing in housing estates built by the government; consecutive sampling5.7Intake of SSSD and intake of fruit drinks or FJ (servings per week or per month); FFQWeight change (kg) (continuous); self reportedAdjusted for age, alcohol intake, BMI, dialect, dietary intake, education, person years, sex, smoking, year of interviewParticipants in the highest category of SSSD consumption (>2 servings/week) had a significant increase in weight (0.53 kg) compared with those who did not consume SSSD or reported only monthly consumption (P<0.001). There was no association between intake of fruit drinks/FJ and change in mean weight between baseline and follow-up. A test for interaction between SSSD intake and weight gain over time was significant (P=0.007)+ (SSSD), 0 (FJ)
Palmer (2008),71 USA, Black Women’s Health Study43 960 African American women aged 21-69 years; convenience sampling6Change in intake of SSSD and change in intake of fruit drinks (servings/day); FFQWeight gain (kg) (continuous); self reportedAdjusted for baseline age and BMI, dietary intake and change in dietary intake, education, family history of diabetes, physical activity and change in physical activity, smoking and change in smokingCompared with women who reduced intake of SSSD (≥1 to ≤1 servings/day), those who increased intake (≤1 to ≥1) gained significantly more weight over the 6 year follow-up (4.1 kg (SD 0.22) v 6.8 kg (0.28); P<0.01). No significant associations were noted between weight gain and change in fruit drink intake+ (SSSD)
0 (FJ)
Parker (1997),72 USA, Pawtucket Heart Health Program465 adults aged 18-64 years, participating in a community based programme for cardiovascular disease prevention; random sampling4Intake of sugar (sucrose; g/day) and sweets (servings/week); FFQWeight change (kg) (continuous); measuredAdjusted for age, BMI, physical activity, smoking, total energy intakeIntake of sucrose and sweets at baseline was not associated with change in weight at follow-up0
Schulz (2002),73 Germany, EPIC-Potsdam Cohort17 396 adults aged 19-70 years, enrolled in EPIC in Potsdam; convenience sampling2.2100 g/day increments in intake of sweets, cakes, and biscuits; SSSD; and desserts. FFQWeight change (OR; kg/year; large gain (≥2), small gain (1-2), stable (gain or loss of 1), small loss (1-2), large loss (≥2); self reported Adjusted for age, baseline weight and height, change in dietary intake, education, life and health contentment, drugs use, prevalent diabetes, prevalent stroke, weight cycling history; stratified by sexPer 100 g/day higher increment in intake of sweets at baseline, men were more likely to have a large gain or small loss in weight than remain stable (OR 1.48 and 1.43, respectively; both P<0.05). In men, SSSD intake at baseline was positively associated with large weight gain (OR 1.03), small weight loss (1.02) and large weight loss (1.03; all P<0.05). Per 100 g/day higher increment in intake of sweets or cakes/biscuits at baseline, women were less likely to have a large loss in weight than remain stable (0.67 and 0.88, respectively; both P<0.05). In women, SSSD intake at baseline was positively associated with large weight loss (1.02, P<0.05)+ and −
Schulze (2004),74 USA, Nurses’ Health Study II51 603 registered female nurses aged 24-44 years; consecutive sampling4Change in consumption of SSSD, fruit drinks, FJ (from ≤1 time/week to ≥1 time/day, ≥1 time/day to ≤1 time/week, consistently ≤1 time/week, consistently ≥1 time/day); FFQWeight change (kg) and BMI change (both continuous); self reportedAdjusted for baseline values of age, alcohol intake, BMI, dietary intake, oral contraceptive use, physical activity, postmenopausal hormone use, and smoking, and changes in all variables over timeWomen who increased consumption of SSSD gained more weight and reported a higher BMI at follow-up than those who decreased consumption or maintained a high or low intake (all P<0.001). Similarly, women who increased consumption of fruit drink/FJ gained more weight than those who decreased consumption (P<0.001)+
(only in women who increased intake)

ANCOVA=analysis of covariance; BF=body fat; FFQ=food frequency questionnaire; FJ=100% fruit juice; HR=hazard ratio; OR=odds ratio; SE=standard error; SD=standard deviation; SSSD=sugar sweetened soft drinks; SSB=sugar sweetened beverages (including cordials, energy drinks, fruit drinks, iced tea, soft drinks); TV=television; WC=waist circumference.

*Most adjusted results are reported unless otherwise stated.

†Higher sugar intake positively associated with weight gain (+), not associated with weight gain (0), and negatively associated with weight gain (−).