Substitution of sugar-sweetened beverages for other beverages and the risk of developing coronary heart disease: Results from the Harvard Pooling Project of Diet and Coronary Disease

Journal Pre-proof Substitution of sugar-sweetened beverages for other beverages and the risk of developing coronary heart disease: Results from the Harvard Pooling Project of Diet and Coronary Disease

Amélie Keller, Eilis J. O'Reilly, Vasanti Malik, Julie E. Buring, Ingelise Andersen, Lyn Steffen, Kim Robien, Satu Männistö, Eric B. Rimm, Walter Willett, Berit Heitmann PII:

S0091-7435(19)30453-0

DOI:

https://doi.org/10.1016/j.ypmed.2019.105970

Reference:

YPMED 105970

To appear in:

Preventive Medicine

Received date:

12 July 2019

Revised date:

9 October 2019

Accepted date:

22 December 2019

Please cite this article as: A. Keller, E.J. O'Reilly, V. Malik, et al., Substitution of sugarsweetened beverages for other beverages and the risk of developing coronary heart disease: Results from the Harvard Pooling Project of Diet and Coronary Disease, Preventive Medicine(2018), https://doi.org/10.1016/j.ypmed.2019.105970

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2018 Published by Elsevier.

Journal Pre-proof

1

Substitution of sugar-sweetened beverages for other beverages and the risk of developing coronary heart disease: results from the Harvard Pooling Project of Diet and Coronary Disease Amélie Keller1 (MPH), Eilis J O’Reilly2,4 (ScD), Vasanti Malik2,3(ScD), Julie E. Buring5(Dr.), Ingelise Andersen6(PhD), Lyn Steffen7(PhD), Kim Robien8(PhD), Satu Männistö9(PhD), Eric B. Rimm 2,3,10(ScD), Walter

of

Willett2,3,10(M.D., Dr. P.H), Berit Heitmann1, 11,12(PhD) 1

Research Unit for Dietary Studies at The Parker Institute, Bispebjerg and Frederiksberg Hospital, Part of Copenhagen University Hospital, The Capital Region, Frederiksberg, Denmark 3

ro

2

Department of Nutrition, Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA

4

Schoolof Public Health, University College Cork, Ireland

Division of Preventive Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA

-p

5 6

Section of Social Medicine, Department of Public Health, University of Copenhagen, Denmark

7

Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, USA

re

8

Department of Exercise and Nutrition Sciences, Milken Institute School of Public Health, George Washington University, USA Institute for Health and Welfare Chronic Disease Prevention Unit, Finland

10

Channing Division of Network Medicine, Brigham and Women's Hospital, Boston, MA, USA

11

lP

9

The Boden Institute of Obesity, Nutrition, Exercise & Eating Disorders, The University of Sydney, Sydney, New South Wales Australia Section for General Practice, Department of Public Health, University of Copenhagen, Denmark

na

12

Jo ur

Corresponding author: Amélie Keller Post-doc Research Unit for Dietary Studies at the Parker Institute, Frederiksberg og Bispebjerg Hospital Nordre Fasanvej 57, Vej 8, Indgang 11 2000 Frederiksberg Denmark Email: [email protected] Funding: This project was supported by the Danish Heart Foundation, Region Hovedstaden, Etly og Jørgen Stjerngrens Fond, the Swiss Foundation for Nutrition Research, Oticon fonden, Niehls Bohr fondet. The funders had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. Conflict of interest: None of the authors had any conflicts of interest. Running title: Sugar-sweetened beverages and heart diseases Word count: 2529

Journal Pre-proof

Abbreviations: ARIC - The Atherosclerosis Risk in Communities Study ASB - Artificially-Sweetened Beverages ATBC - The Alpha-Tocophenol and Beta-Carotene Cancer Prevention Study BMI – Body Mass Index CHD – Coronary Heart Disease CI - Confidence Intervals

of

CVD – Cardiovascular Disease

ro

FFQ – Food Frequency Questionnaire

-p

HPFS - The Health Professionals Follow-up Study HPP - Harvard Pooling Project

re

HR – Hazard Ratio

lP

IWHS - The Iowa Women’s Health Study

NHS - The Nurses’ Health Study TEI - Total Energy Intake

Jo ur

T2DM – Type 2 diabetes mellitus

na

MI – Myocardial Infarction

SSB – Sugar-Sweetened Beverages WHS - The Women’s Health Study

2

Journal Pre-proof

3

Abstract Sugar-sweetened beverage (SSB) intake is associated with metabolic disorders. The reduction of SSB intake has been promoted to prevent death and disability from chronic diseases. We investigated the association between SSB intake and the risk of coronary events and death, and assessed if substitution of coffee, tea, milk, fruit juice and artificially-sweetened beverages (ASB) for SSBs was associated with a reduced risk of coronary events and death. This was a follow-up study in which data from six studies were pooled and

of

standard observational analyses were performed. Diet intake was assessed at baseline by food-frequency

ro

questionnaires. Hazard ratios (HRs) with 95% confidence intervals for the incidence of coronary events and

-p

deaths were calculated by Cox proportional hazards regression. The effect of substituting another beverage for SSBs was calculated by taking the difference in the individual effect estimates. During the median 8.2-

re

year follow-up, 4,248 coronary events and 1,630 coronary deaths were documented among 284,345

lP

individuals. 355 ml daily increase of SSB intake was associated with an increased risk of coronary events (HR: 1.08; 95%CI: 1.02, 1.14) and possibly coronary death (HR: 1.05; 95%CI: 0.96, 1.16). Substitution

na

analyses suggested that replacing SSBs with coffee (HR: 0.93; 95%CI: 0.87, 1.00) or ASB (HR: 0.89; 95%CI:

Jo ur

0.83, 0.97), might be associated with a lower risk of developing coronary events. We found that SSB intake was associated with an increased risk of coronary events and possibly coronary death. Our findings also suggest that replacing SSB’s with ASBs or coffee may lower the risk of developing CHD.

Key words Sugar-sweetened beverages; coronary heart diseases, substitution, Harvard Pooling Project

Journal Pre-proof

4

Introduction One-half of US adults consumes ≥1 sugar-sweetened beverage (SSB) daily, representing 6.5% of total energy intake (TEI) (1). Direct associations between SSB consumption and weight gain, obesity (2) and type2 diabetes mellitus (T2DM) have been shown (3,4); and each additional daily serving of SSBs has been associated with a 22% increased risk of myocardial infarction (MI) and 13% increased risk of stroke (5). A recent meta-analysis of published studies found that SSB consumption is positively associated with

of

coronary heart disease (CHD), stroke and heart failure (6). Furthermore, high intake of SSBs has also been

ro

associated with increased blood pressure and hyperlipidemia (7,8). As one-third of deaths worldwide are

-p

attributable to cardiovascular disease (CVD), mainly CHD (9), evaluation of potential strategies for reducing the burden of CHD is needed.

re

Although individual studies have shown an inverse association between coffee, tea and possibly milk

lP

intakes (10–12) and the risk of CVD and limiting SSB consumption is recommended by public health agencies, there is insufficient evidence of what constitutes appropriate replacement beverages to make

na

recommendations related to lowering the risk of developing CHD. Previous substitution studies have shown

Jo ur

that substitution of water, milk, artificially-sweetened beverage (ASB), plain tea or coffee for SSBs was associated with a lower risk of weight gain, obesity and T2DM (4,13–16). In regards to cardiovascular health, results from intervention studies evaluating the effect of replacing SSBs by other beverages on CVDs showed discrepant results (17–19); while results from observational studies suggested that substituting SSBs with fruit juice and water or coffee was associated with a lower risk of metabolic syndrome (20) and stroke (21), respectively. Therefore, we aimed to (1) supplement the growing evidence regarding the association between SSB intake and the risk of incident CHD using standardized modeling of the exposure and confounding variables to remove potential sources of noncomparability and heterogeneity that occur in the published literature, and (2) to assess whether substitution of coffee, tea, milk, fruit juice and ASB for SSBs was associated with a lower risk of CHD events and deaths.

Journal Pre-proof

5

Subjects and Methods Six of the eleven studies from the Harvard Pooling Project (HPP) of Diet and Coronary Disease were included. The HPP inclusion criteria were prospective studies with ≥150 incident coronary cases, assessment of usual dietary intake, and a validation of the diet assessment method(22). Only participants aged ≥35 years, without a history of CVD, diabetes, or cancer or extreme energy intake and with information on specific beverage consumption were included. The following studies met these criteria: The

of

Atherosclerosis Risk in Communities Study (ARIC)(23); The Alpha-Tocophenol and Beta-Carotene Cancer

ro

Prevention Study (ATBC)(24); The Health Professionals Follow-up Study (HPFS)(25); The Iowa Women’s

-p

Health Study (IWHS)(26); The Women’s Health Study (WHS)(27); The Nurses’ Health Study (1980-1986 (NHS 80); 1986-1996 (NHS 86))(28). Two of the six included studies were randomized primary prevention studies

re

and the others were prospective cohort studies. In the ATBC study, which was a clinical trial, analyses were

lP

also adjusted for the arm to which participants were randomized (placebo or intervention). WHS analyses

Exposure measures

na

were not adjusted for trial arm as the study was ongoing at inclusion into the HPP.

Jo ur

Diet intake was assessed at baseline by food-frequency questionnaires (FFQ). SSBs included carbonated/non-carbonated and caffeinated/non-caffeinated sodas, sport drinks and fruit drinks with any type of added sugar. Caffeinated coffee included all types of plain (unsweetened) coffee with caffeine and total coffee also included decaffeinated coffee. Tea included all types of plain (unsweetened) tea. Milk included only non-sweetened cow milk, either whole-fat, low-fat or total milk (whole-fat and low-fat combined). Fruit juice included only 100% fruit juice. ASB included any diet drinks sweetened with artificial sweeteners. The volume or frequency of a serving was used to assess the quantity of each beverage

consumed daily. Frequency data were converted to volume consumed/day on the basis of the frequency and study-specific serving size for each item. We calculated the consumption of beverage types by summing the related individual beverages listed in each study.

Journal Pre-proof

6

Outcome measures The outcome measures were fatal CHD and nonfatal MI. Standardized criteria, questionnaires supplemented by medical records, autopsy reports or death certificates reviewed by physicians(29,30), were used to ascertain CHD events and death in each study. In this study, CHD events refer to any first incident CHD events, first event can be fatal CHD, and CHD death refers to total incident CHD death. The IWHS used self-reported CHD events; therefore only CHD death was used. Due to insufficient number of

of

CHD deaths among women in ARIC and WHS, only total coronary events were included in this analysis.

ro

Statistical analyses

-p

Hazard ratios (HRs) with 95% confidence intervals (CIs) for the incidence of coronary events and deaths were calculated by Cox proportional hazards regression. The time metric was survival from entry into the

re

study and stratification on age in years at baseline and calendar year at baseline questionnaire was

lP

performed. Person-years of follow-up were calculated from baseline until the date of event, death, or end of follow-up. Follow-up was censored at 10 years to minimize misclassification of exposure.

na

The study-logs (sex-specific in ARIC) of HRs were weighted by the inverse of their variances, and a pooled

Jo ur

estimate of the HRs was computed using the fixed-effects model, analyses using the random-effects model were also run as sensitivity analyses (data not shown). Between-study heterogeneity was calculated using the I-squared (I2) test. Effect modification by sex was also investigated by the I2 test for between-group heterogeneity. SSB consumption was primarily modelled continuously. Categorical analyses (<1 serving/d; 1-2 serving/d; >2 servings/d) were also performed (Supplementary data: Tables S2-3). The effect of substituting another beverage for SSBs was calculated by including them in the same model and taking the difference in the individual effect estimates (ß-coefficients). The 95% CI for the substitution effects were calculated using variances and covariance of the effect estimates for each beverage (the square root of the sum of their variances − (2 × covariance) (31,32). Women and men have different risks of CVD(33), therefore, sexspecific associations were also examined. Potential confounders and mediators were considered and

Journal Pre-proof

7

stepwise added into the analyses. Nutrient intakes, including quintiles of cereal-fiber; trans-fatty acid; polyunsaturated fatty acid: saturated fatty acid ratio, were included in the model as potential confounders. TEI and Body Mass Index (BMI) were considered both confounders and mediators; therefore, they were added to the model individually. Baseline hypertension and high cholesterol were also added to the model as potential confounders. The crude model included only SSB and the stratification variables (smoking, physical activity, education,

of

alcohol intake); model 1 further considered confounding by nutrient intakes; model 2 further included total energy; model 3 further included BMI; and model 4 further included baseline hypertension and high

ro

cholesterol.

-p

Individuals developing CHD in the first years of follow-up may have changed their dietary consumption due

re

to pre-existing symptoms. Therefore, sensitivity analyses excluding individuals who developed CHD within

lP

the first two years of follow-up were performed. Ethics

Jo ur

all participants was obtained.

na

All included studies were previously approved by a national or institutional review board and consent from

Journal Pre-proof

8

Results Characteristics of participants from included studies are presented in Table 1. During the median 8.2-year follow-up, 4,248 CHD events and 1,630 CHD deaths occurred among 284,345 individuals (Supplementary data: Figure S1). The daily median intake of SSBs was 48ml with 37% reporting no drinking of SSBs in the previous year. In the following, all results refer to Model 3. Combined HRs for 355 ml (12oz) daily increase of SSB intake and coronary events are shown in Table 2 and

of

Figure 1, and CHD deaths are shown in Supplementary data: Table S1 and Figure S2. There was a positive

ro

association between daily increase in SSB intake and overall risk of coronary events in men (HR for 355

-p

ml/day:1.10; 95%CI:1.02-1.17) and in men and women combined (HR for 355 ml/day:1.08; 95%CI:1.021.14) but the results were not significant among women (HR for 355 ml/day:1.06; 95%CI:0.97-1.15). There

re

was non-significant positive association between SSB intake and risk of coronary deaths in men and women

lP

combined (HR for 355 ml/day :1.05; 95%CI:0.96-1.16 I2=0%, p-heterogeneity=0.50). There was no indication of effect modification by sex (p>0.05). Results from the random effects model were similar (data not

na

shown). No associations were seen between categories of SSB intake and the risk of coronary deaths, but

Jo ur

there was an increased risk of coronary events among individuals consuming 1-2 servings SSBs per day compared to less than 1 serving (HR:1.15; 95%CI:1.00-1.33) (Supplementary data: Tables S2-S3). Combined HRs for substitution of SSB’s with coffee, tea, milk, fruit juice, and ASB (355ml/d) and coronary events are shown in Table 3, and coronary deaths are shown in Supplementary data: Table S4. Substituting caffeinated coffee for SSBs was inversely associated with the overall risk of coronary events (HR:0.92; 95%CI:0.86-0.99; I2=0%, p-heterogeneity=0.42). There was an inverse association between substitution of total coffee for SSBs and the risk of coronary events among women (HR:0.91; 95%CI:0.840.98; I2=25%, p-heterogeneity=0.25) and women and men combined (HR:0.93; 95%CI:0.87-1.00; I2=10%, pheterogeneity=0.34), but not among men (HR:1.00; 95%CI:0.87-1.15; I2=0%, p-heterogeneity=0.44) (Figure 2.

Journal Pre-proof

9

There was a suggestion of an inverse association between substitution of tea for SSBs with the risk of coronary events (HR:0.94; 95%CI: 0.87-1.01; I2=0%, p-heterogeneity=0.72). Substituting ASBs for SSBs was inversely associated with coronary events (HR:0.89; 95%CI:0.83-0.97; I2=28%, p-heterogeneity=0.23) (Figure 3). There was a suggestion of an overall inverse association between substitution of low-fat milk (HR:0.96; 95%CI:0.90-1.03; I2=0%, p-heterogeneity=0.76) and total milk (HR:0.96; 95%CI:0.90-1.03; I2=0%, p-

of

heterogeneity=0.50) for SSBs and the risk of coronary events. Substituting whole-fat milk for SSBs tended to decrease the risk of coronary events among men (HR:0.89; 95%CI:0.80-1.00; I2=58%, p-heterogeneity=0.08)

ro

but not among women (HR:1.13; 95%CI:0.96-1.34; I2=0%, P=0.84; p-heterogeneity by sex=0.02).

-p

Overall, substituting SSBs with other beverages was not associated with a different risk of coronary death.

re

There was no indication of effect modifications by sex (p>0.05), except for substitution of whole-fat milk for

lP

SSBs and coronary events (p-heterogeneity=0.02).

Results from the random-effects model were similar to the fixed-effects model (data not shown).

Jo ur

similar results (data not shown).

na

Sensitivity analyses excluding individuals developing CHD within the first two years of follow-up showed

Journal Pre-proof

10

Discussion This study suggests that daily increase of SSB intake is associated with an increased risk of coronary events. These results are in accordance with recent evidence suggesting a direct association between SSB intake and CHD (6,8). In addition, this study reports several novel findings. Substitution analyses suggested that replacing SSBs with ASBs or coffee lowers the risk of developing CHD events. Although not statistically significant, substituting tea, total milk, or low-fat milk for SSBs might also have inverse associations with

of

coronary events risk development

ro

Recent evidence suggests that tea and coffee consumption is associated with a reduced CVD risk (11,12).

-p

One previous study found that substituting SSBs with coffee was associated with a lower risk of stroke and a trend towards a reduced risk with substitution with tea (21). To our knowledge, no previous studies

re

estimated the effect of substituting tea or coffee for SSBs and the subsequent CHD risk. Tea constituents

lP

might ameliorate endothelial function, lipid profile and have anti-inflammatory, antioxidant or antithrombotic properties(12). Regular coffee contains caffeine which may elevate blood pressure;

na

however it has been shown that caffeine consumption in the form of coffee was less likely to be associated

Jo ur

with hypertension than caffeine consumed as SSBs. Thus, non-caffeine components in coffee may have antagonist physiological effects to those of caffeine (34). This study is the first to show a lower incidence of CHD events associated with replacing SSBs by ASBs. These findings might be counter intuitive as some recent evidence suggests a positive association between ASB intake and adverse health outcomes (35). However, challenges related to assessing the association between ASB intake and diseases development from cross-sectional studies or cohorts with short duration, such as reverse causation and residual confounding, may account for some studies findings(35). In line with our results, substitution of ASBs for SSBs was associated with reduced risk of stroke among men and women; and weight losses of 2% to 2.5% in two recent studies (17,21). Thus, compared with SSB’s, ASBs may be a healthier option for the prevention of coronary events.

Journal Pre-proof

11

A pooled analysis of 10 prospective studies suggested a beneficial effect of milk consumption in reducing the risk of coronary heart disease and ischemic stroke (36); however, one previous study did not find any association between substitution of low fat milk for SSBs and the risk of stroke (21). Hence, the present study is the first to show a possible beneficial effect of substituting SSBs by milk, especially low-fat milk. Milk intake has been shown to increase high density lipoprotein cholesterol level which has antiatherogenic properties and may also have antihypertensive properties due to its mineral content (calcium,

of

potassium, phosphorus, magnesium) (10,37).

ro

Data about water intake were incomplete or missing in most of the included studies; therefore, substitution of water for SSBs was not assessed. The paucity of data on water intake can be explained by the lack of

-p

focus on collecting such information in past cohort studies(38).

re

Strengths and limitations

lP

To our knowledge, no previous study has examined the long-term effects of replacing SSBs with other beverages on the risk of coronary events and deaths. Our results have good generalizability among Western

na

middle-aged populations, but may not be extrapolated to other populations. By using the Harvard Pooling

Jo ur

Project of Diet and Coronary Disease we had a large sample size and great statistical power for total CHD, but power for fatal CHD was much less. Publication bias did not affect our pooled analyses as individual study inclusion criteria were independent from publication status. Selection bias was also limited as dietary assessment was conducted, at baseline, prior to disease development(39). The availability of dietary intake at baseline only prevented us from investigating change in beverage intake across time or to reduce random measurement error through statistical analysis. Individuals who changed drinking patterns may have been misclassified with respect to exposure; therefore, follow-up was censored at 10 years. Furthermore, sensitivity analyses excluding individuals who developed CHD within the first two-years of follow-up were performed; and similar results were found.

Journal Pre-proof

12

Average SSB intake was relatively low; therefore it is likely that the effect of substituting SSBs by other beverages would be particularly important among populations with a higher SSB consumption such as young adults, non-Hispanic black and Mexican-American adults, individuals with low socioeconomic statuses and recent cohorts of same age individuals(40). BMI might be considered to be more of a mediator than a confounder in the association between SSB intake and CHD development; therefore, by controlling for BMI the effect of SSB’s may have been

of

underestimated. Indeed, BMI might influence the frequency and quantity of SSB intake and

ro

overweight/obese individuals may have changed their beverage intake pattern in an attempt to lose weight prior to baseline diet assessment.

-p

We adjusted for a number of known potential confounders; however, residual confounding can never be

lP

re

excluded and additional confounding from other CHD risk factors could potentially explain our findings.

na

Conclusion

This study shows that SSB intake is associated with an increased risk of coronary events and possibly

Jo ur

coronary death. Our findings also suggest that replacing SSBs with ASBs or coffee may lower the risk of developing CHD events.

Journal Pre-proof

13

Acknowledgments We would like to thank Dr. Ascherio for his contribution in obtaining the data. The Authors thank the ARIC, NHS, HPFS, WHS, IWHS and ATBC study participants. The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health

of

and Human Services (contract numbers HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700004I and HHSN268201700005I). The authors thank the staff and

ro

participants of the ARIC study for their important contributions. Funding was also supported by

-p

R01HL59367.

re

Funding

lP

This project was supported by the Danish Heart Foundation, Region Hovedstaden, Etly og Jørgen Stjerngrens Fond, the Swiss Foundation for Nutrition Research, Oticon fonden, Niehls Bohr fondet. The

na

funders had no role in the study design; in the collection, analysis, and interpretation of data; in the writing

Jo ur

of the report; and in the decision to submit the article for publication.

Authors contribution

Berit L. Heitmann, Ingelise Andersen and Amélie Keller designed the study and formulated the research question. Eilis O’Reilly and Amélie Keller performed the statistical analyses. Vasanti Malik and Eilis O´Reilly provided feedback on the study design. Amélie Keller wrote the manuscript. Eric Rimm and Walter Willet provided interpretation of data and critical revision of the manuscript for important intellectual content. The remaining investigators (Julie E. Buring, Lyn Steffen, Kim Robien and Satu Männistö) represent the offsite investigators of the individual cohort studies. All of these investigators provided feedback on the manuscript. Conflict of interest: None.

Journal Pre-proof

14

Transparency declaration The lead author, Amélie Cléo Keller, affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that

Jo ur

na

lP

re

-p

ro

of

any discrepancies from the study as originally planned (and, if relevant, registered) have been explained.

Journal Pre-proof

15

References

6.

7. 8.

9. 10.

11.

12.

13.

of

ro

-p

5.

re

4.

lP

3.

na

2.

Rosinger A, Herrick K, Gahche J, Park S. Key findings Data from the National Health and Nutrition Examination Survey. 2011 [cited 2017 Feb 15]; Available from: https://www.cdc.gov/nchs/data/databriefs/db270_table.pdf#1. Hu FB. Resolved: there is sufficient scientific evidence that decreasing sugar-sweetened beverage consumption will reduce the prevalence of obesity and obesity-related diseases. Obes Rev [Internet]. NIH Public Access; 2013 [cited 2018 Apr 9];14:606–19. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23763695 Malik VS, Popkin BM, Bray GA, Despres J-P, Willett WC, Hu FB. Sugar-Sweetened Beverages and Risk of Metabolic Syndrome and Type 2 Diabetes: A meta-analysis. Diabetes Care [Internet]. 2010 [cited 2017 Dec 1];33:2477–83. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20693348 Imamura F, O’Connor L, Ye Z, Mursu J, Hayashino Y, Bhupathiraju SN, Forouhi NG. Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction. BMJ [Internet]. 2015 [cited 2017 Feb 15];h3576. Available from: http://www.bmj.com/lookup/doi/10.1136/bmj.h3576 Narain A, Kwok CS, Mamas MA. Soft drinks and sweetened beverages and the risk of cardiovascular disease and mortality: a systematic review and meta-analysis. Int J Clin Pract [Internet]. 2016 [cited 2017 Jul 10];70:791–805. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27456347 Bechthold A, Boeing H, Schwedhelm C, Hoffmann G, Knüppel S, Iqbal K, De Henauw S, Michels N, Devleesschauwer B, Schlesinger S, et al. Food groups and risk of coronary heart disease, stroke and heart failure: A systematic review and dose-response meta-analysis of prospective studies. Crit Rev Food Sci Nutr [Internet]. 2017 [cited 2019 Mar 14];1–20. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29039970 Keller A, Heitmann BL, Olsen N. Sugar-sweetened beverages, vascular risk factors and events: a systematic literature review. Public Health Nutr. 2015;18:1145–54. Xi B, Huang Y, Reilly KH, Li S, Zheng R, Barrio-Lopez MT, Martinez-Gonzalez MA, Zhou D. Sugarsweetened beverages and risk of hypertension and CVD: a dose–response meta-analysis. Br J Nutr [Internet]. 2015 [cited 2018 Apr 9];113:709–17. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25735740 WHO | Cardiovascular diseases (CVDs) [Internet]. WHO. World Health Organization; 2017 [cited 2018 Apr 9]. Available from: http://www.who.int/mediacentre/factsheets/fs317/en/ Soedamah-Muthu SS, Ding EL, Al-Delaimy WK, Hu FB, Engberink MF, Willett WC, Geleijnse JM. Milk and dairy consumption and incidence of cardiovascular diseases and all-cause mortality: doseresponse meta-analysis of prospective cohort studies. Am J Clin Nutr [Internet]. 2011 [cited 2018 Apr 9];93:158–71. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21068345 Ding M, Bhupathiraju SN, Satija A, van Dam RM, Hu FB. Long-term coffee consumption and risk of cardiovascular disease: a systematic review and a dose-response meta-analysis of prospective cohort studies. Circulation [Internet]. NIH Public Access; 2014 [cited 2017 May 19];129:643–59. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24201300 Stangl V, Lorenz M, Stangl K. The role of tea and tea flavonoids in cardiovascular health. Mol Nutr Food Res [Internet]. 2006 [cited 2017 May 29];50:218–28. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16404706 Zheng M, Allman-Farinelli M, Heitmann BL, Rangan A. Substitution of Sugar-Sweetened Beverages with Other Beverage Alternatives: A Review of Long-Term Health Outcomes. J Acad Nutr Diet

Jo ur

1.

Journal Pre-proof

15.

16.

20.

21.

22.

23.

24.

25.

re

lP

na

19.

Jo ur

18.

-p

ro

17.

[Internet]. 2015 [cited 2018 Apr 9];115:767–79. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25746935 Fresán U, Gea A, Bes-Rastrollo M, Ruiz-Canela M, Martínez-Gonzalez M. Substitution Models of Water for Other Beverages, and the Incidence of Obesity and Weight Gain in the SUN Cohort. Nutrients [Internet]. 2016 [cited 2017 Jun 26];8:688. Available from: http://www.ncbi.nlm.nih.gov/pubmed/27809239 Pan A, Malik VS, Hao T, Willett WC, Mozaffarian D, Hu FB. Changes in water and beverage intake and long-term weight changes: results from three prospective cohort studies. Int J Obes [Internet]. 2013 [cited 2017 Jun 26];37:1378–85. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23318721 de Koning L, Malik VS, Rimm EB, Willett WC, Hu FB. Sugar-sweetened and artificially sweetened beverage consumption and risk of type 2 diabetes in men. Am J Clin Nutr [Internet]. American Society for Nutrition; 2011 [cited 2017 Jun 23];93:1321–7. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21430119 Tate DF, Turner-McGrievy G, Lyons E, Stevens J, Erickson K, Polzien K, Diamond M, Wang X, Popkin B. Replacing caloric beverages with water or diet beverages for weight loss in adults: main results of the Choose Healthy Options Consciously Everyday (CHOICE) randomized clinical trial. Am J Clin Nutr [Internet]. 2012 [cited 2018 Apr 9];95:555–63. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22301929 Hernández-Cordero S, Cuevas-Nasu L, Morán-Ruán MC, Méndez-Gómez Humarán I, Ávila-Arcos MA, Rivera-Dommarco JA. Overweight and obesity in Mexican children and adolescents during the last 25 years. Nutr Diabetes [Internet]. Nature Publishing Group; 2017 [cited 2018 Jan 12];7:e247. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28287630 Hernandez-Cordero S, Barquera S, Rodriguez-Ramirez S, Villanueva-Borbolla MA, Gonzalez de Cossio T, Dommarco JR, Popkin B. Substituting Water for Sugar-Sweetened Beverages Reduces Circulating Triglycerides and the Prevalence of Metabolic Syndrome in Obese but Not in Overweight Mexican Women in a Randomized Controlled Trial. J Nutr [Internet]. 2014 [cited 2017 Jun 26];144:1742–52. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25332472 Mattei J, Malik V, Hu FB, Campos H. Substituting Homemade Fruit Juice for Sugar-Sweetened Beverages Is Associated with Lower Odds of Metabolic Syndrome among Hispanic Adults. J Nutr [Internet]. 2012 [cited 2018 Apr 9];142:1081–7. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22551801 Bernstein AM, de Koning L, Flint AJ, Rexrode KM, Willett WC. Soda consumption and the risk of stroke in men and women. Am J Clin Nutr [Internet]. American Society for Nutrition; 2012 [cited 2018 Apr 9];95:1190–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22492378 Jakobsen MU, O’Reilly EJ, Heitmann BL, Pereira MA, Bälter K, Fraser GE, Goldbourt U, Hallmans G, Knekt P, Liu S, et al. Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies. Am J Clin Nutr [Internet]. 2009 [cited 2018 Apr 9];89:1425–32. Available from: https://academic.oup.com/ajcn/article/89/5/1425/4596851 The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol [Internet]. 1989 [cited 2017 Feb 17];129:687–702. Available from: http://www.ncbi.nlm.nih.gov/pubmed/2646917 Virtamo J, Taylor PR, Kontto J, Männistö S, Utriainen M, Weinstein SJ, Huttunen J, Albanes D. Effects of α-tocopherol and β-carotene supplementation on cancer incidence and mortality: 18-Year postintervention follow-up of the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study. Int J Cancer [Internet]. 2014 [cited 2017 Nov 28];135:178–85. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24338499 Hu FB, Rimm E, Smith-Warner SA, Feskanich D, Stampfer MJ, Ascherio A, Sampson L, Willett WC.

of

14.

16

Journal Pre-proof

29.

30.

34.

35.

36.

37.

38.

na

33.

Jo ur

32.

lP

re

31.

of

28.

ro

27.

Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am J Clin Nutr [Internet]. American Society for Nutrition; 1999 [cited 2017 Feb 17];69:243–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/9989687 Munger RG, Folsom AR, Kushi LH, Kaye SA, Sellers TA. Dietary assessment of older Iowa women with a food frequency questionnaire: nutrient intake, reproducibility, and comparison with 24-hour dietary recall interviews. Am J Epidemiol [Internet]. 1992 [cited 2017 Nov 28];136:192–200. Available from: http://www.ncbi.nlm.nih.gov/pubmed/1415141 Rexrode KM, Lee I-M, Cook NR, Hennekens CH, Buring JE. Baseline Characteristics of Participants in the Women’s Health Study. J Womens Health Gend Based Med [Internet]. 2000 [cited 2018 Apr 9];9:19–27. Available from: http://www.ncbi.nlm.nih.gov/pubmed/10718501 Belanger CF, Hennekens CH, Rosner B, Speizer FE. The Nurses’ Health Study. Am J Nurs *Internet+. 1978 [cited 2017 Feb 21];78:1039. Available from: http://www.jstor.org/stable/3462013?origin=crossref Pereira MA, O’Reilly E, Augustsson K, Fraser GE, Goldbourt U, Heitmann BL, Hallmans G, Knekt P, Liu S, Pietinen P, et al. Dietary Fiber and Risk of Coronary Heart Disease. Arch Intern Med [Internet]. 2004 [cited 2018 Apr 9];164:370. Available from: http://www.ncbi.nlm.nih.gov/pubmed/14980987 Liu S, Buring JE, Sesso HD, Rimm EB, Willett WC, Manson JE. A prospective study of dietary fiber intake and risk of cardiovascular disease among women. J Am Coll Cardiol [Internet]. 2002 [cited 2018 May 1];39:49–56. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11755286 Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major Dietary Protein Sources and Risk of Coronary Heart Disease in Women. Circulation [Internet]. 2010 [cited 2018 Oct 15];122:876–83. Available from: https://www.ahajournals.org/doi/10.1161/CIRCULATIONAHA.109.915165 Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr [Internet]. Oxford University Press; 1997 [cited 2018 Apr 9];65:1220S–1228S. Available from: https://academic.oup.com/ajcn/article/65/4/1220S/4655744 Rosen SE, Henry S, Bond R, Pearte C, Mieres JH. Sex-Specific Disparities in Risk Factors for Coronary Heart Disease. Curr Atheroscler Rep [Internet]. Springer US; 2015 [cited 2017 Nov 28];17:49. Available from: http://link.springer.com/10.1007/s11883-015-0523-8 Winkelmayer WC, Stampfer MJ, Willett WC, Curhan GC. Habitual Caffeine Intake and the Risk of Hypertension in Women. JAMA [Internet]. American Medical Association; 2005 [cited 2017 Jun 5];294:2330. Available from: http://jama.jamanetwork.com/article.aspx?doi=10.1001/jama.294.18.2330 Swithers SE. Artificial sweeteners produce the counterintuitive effect of inducing metabolic derangements. Trends Endocrinol Metab [Internet]. 2013 [cited 2018 Apr 9];24:431–41. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23850261 Elwood PC, Pickering JE, Hughes J, Fehily AM, Ness AR. Milk drinking, ischaemic heart disease and ischaemic stroke II. Evidence from cohort studies. Eur J Clin Nutr [Internet]. Nature Publishing Group; 2004 [cited 2017 Mar 16];58:718–24. Available from: http://www.nature.com/doifinder/10.1038/sj.ejcn.1601869 Chrysant SG, Chrysant GS. An Update on the Cardiovascular Pleiotropic Effects of Milk and Milk Products. J Clin Hypertens [Internet]. 2013 [cited 2017 Jun 7];15:503–10. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23815539 Popkin BM, D’Anci KE, Rosenberg IH. Water, hydration, and health. Nutr Rev *Internet+. NIH Public Access; 2010 [cited 2017 Jun 26];68:439–58. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20646222

-p

26.

17

Journal Pre-proof

of ro -p re lP na

40.

Smith-Warner SA, Spiegelman D, Ritz J, Albanes D, Beeson WL, Bernstein L, Berrino F, van den Brandt PA, Buring JE, Cho E, et al. Methods for Pooling Results of Epidemiologic Studies. Am J Epidemiol [Internet]. 2006 [cited 2018 Apr 9];163:1053–64. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16624970 Han E, Powell LM. Consumption Patterns of Sugar-Sweetened Beverages in the United States. J Acad Nutr Diet [Internet]. 2013 [cited 2018 Apr 9];113:43–53. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23260723

Jo ur

39.

18

Journal Pre-proof

19

Table 1: Baseline characteristics of participants from included studies CHD cases Study and sex

SSB intake

Baseline cohort1

Year of FFQ

Mean age, y

Median Follow-up, person-y

Total CHD events

CHD deaths

Median (ml/d)

Nondrinkers %

Male

5,238

1987

54.6

269

52

127.3

21.0

Female2

6,481

9.2 45,860 9.2 58,199

41.4

36.5

Male

21,141

1985

57.3

6.0 121,813

534

47.1

32.1

Male

41,684

1986

53.4

9.7 382,573

420

52.1

33.9

Female

29,528

1986

61.4

291

29.8

42.4

Female

81,412

1980

Female

61,700

Female

37,161

ARIC

53.9

123

ro

1339

1272

-p

HPFS

WHS

6.5 513,930

397

97

26.0

38.7

1986

52.6

10.0 606,997

696

208

29.6

47.1

1993

53.9

5.3 190,198 8.2 2,207,823

152

10

29.8

43.8

4248

1630

47.9

37

na

284,345 (23.9)

10.0 288,253

46.9

Jo ur

NHS4 1986

lP

NHS4 1980

re

IWHS3

All Studies5

of

ATBC

54.3

1

Sample size after exclusion of participants with baseline cardiovascular diseases, diabetes mellitus, cancer, very high or very low

caloric intake and missing information on SSB intake. Events censored at 10 years follow-up. 2

Insufficient events of coronary death for analysis among women

3

IWHS only had self-reported data on CHD events, therefore only CHD deaths were used

4

Number of participants only counted once

5

Percentage of males in parentheses

Journal Pre-proof

20

Table 2: Combined hazard ratio (HR) for coronary events for one serving daily increase of SSB intake All

Women 2

Men 2

HR (95%CI)

I test for betweenstudies heterogeneity (p-value)

N

HR (95%CI)

I test for betweenstudies heterogeneity (p-value)

N

HR (95%CI)

Crude Model

1.16 (1.11,1.22)

0.0% (0.70)

284,289

1.18 (1.10,1.27)

0.0% (0.90)

216,226

Model 1

1.08 (1.03, 1.14) 1.09 (1.04, 1.15) 1.08 (1.02, 1.14) 1.08 (1.02,1.13)

0.0% (0.46) 0.0% (0.74) 0.0% (0.61) 0.0% (0.67)

284,289

1.06 (0.98, 1.15) 1.08 (1.00, 1.17) 1.06 (0.97, 1.15) 1.06 (0.97,1.15)

24.4% (0.27) 0.0% (0.41) 0.0% (0.41) 0.0% (0.40)

216,226

Model 2 Model 3 Model 4

284,289 280,886 274,754

P l a

re

213,759 143,184

N

P-value, test for effect modification by sex

1.15 (1.08,1.23)

33.8% (0.22)

68,063

0.63

1.10 (1.02, 1.18) 1.10 (1.03, 1.18) 1.10 (1.02, 1.17) 1.09 (1.02,1.17)

0.0% (0.52) 0.0% (0.80) 0.0% (0.55) 0.0% (0.65)

68,063

0.53

68,063

0.68

67,127

0.51

131,570

0.61

f o

ro

-p

216,226

2

I test for betweenstudies heterogeneity (p-value)

n r u

Age at baseline (y) and the calendar year in which the baseline questionnaire was returned were entered into the model through the strata statement. Within each study, HRs with 95% CIs for the incidence of a coronary events was calculated by using Cox proportional hazards regression with time in study (y) as the time metric. The study-specific logs of HRs were

o J

weighted by the inverse of their variances, and a combined estimate of the HRs was computed by using a fixed-effects model. A crude model included only SSB and the stratification variables (smoking, physical activity, education and alcohol); model 1 considered confounding by diet and included quintiles of cereal fibers; quintiles of trans-fat; quintiles of polyunsaturated fat/saturated fat ratio; model 2 further included total energy; model 3 further included BMI; model 4 then included intermediate variables: baseline hypertension and high cholesterol. One serving = 355ml.

Journal Pre-proof

21

Table 3: Combined hazard ratio (HR) for coronary events by substitution of one serving sugar-sweetened beverages (SSB) with one serving of coffee, tea, milk, fruit juice or artificially-sweetened beverages (ASB) All HR (95%CI)

Caffeinated coffee for SSB

Crude model Model 1 Model 2 Model 3 Model 4

Total Coffee for SSB

Crude model Model 1 Model 2 Model 3 Model 4

Women 2

I test for betweenstudies heterogeneity (p-value)

N

0.91 (0.86, 0.97) 0.90 (0.84, 0.96) 0.98 (0.91, 1.04) 0.92 (0.86, 0.99) 0.94 (0.87, 1.00)

0.0% (0.46) 7.3% (0.37) 63.6% (0.02) 0.0% (0.42) 16.2% (0.31)

284,345

0.92 (0.86, 0.98) 0.97 (0.91, 1.04) 0.94 (0.87, 1.00) 0.93 (0.87, 1.00) 0.95 (0.88, 1.01)

10.9% (0.34) 36.5% (0.19) 6.3% (0.36) 10.1% (0.34) 13.1% (0.33)

2

HR (95%CI)

284,345 284,345

274,754

n r u

272,626 272,626 272,626 269,173 212,864

I test for betweenstudies heterogeneity (p-value)

N

HR (95%CI)

0.0% (0.92) 5.2% (0.37) 17.2% (0.31) 0.0% (0.56) 0.0% (0.52)

216,282

re

0.93 0.82, 1.06) 1.00 (0.87, 1.14) 1.01 (0.88, 1.16) 0.91 (0.84, 0.98) 1.02 (0.89, 1.17)

0.0% (0.59) 0.0% (0.63) 30.9% (0.23) 25.0% (0.25) 26.4% (0.24)

209,801

-p

216,282 216,282 213,759 143,184

209,801 209,801 207,283 102,422

2

I test for betweenstudies heterogeneity (p-value)

N

P-value, 2 I test for between -groups heterogeneity

0.88 0.81, 0.96) 0.89 (0.81, 0.97) 0.89 (0.81, 0.98) 0.90 (0.81, 1.01) 0.91 (0.83, 1.00)

62.8% (0.10) 52.4% (0.15) 66.5% (0.08) 61.5% (0.12) 66.3% (0.09)

68,063

0.22

68,063

0.72

68,063

0.01

67,127

0.57

131,570

0.38

0.91 (0.85, 0.98) 0.96 (0.889, 1.040) 0.91 (0.84, 0.99) 1.00 (0.87, 1.15) 0.92 (0.851, 1.00)

66.4% (0.09) 76.6% (0.04) 0.0% (0.56) 0.0% (0.44) 0.0% (0.42)

62,825

0.75

62,825

0.64

62,825

0.24

61,890

0.24

110,442

0.23

f o

ro

0.95 (0.87, 1.03) 0.91 (0.83, 0.99) 1.07 (0.97, 1.17) 0.94 (0.86, 1.03) 0.96 (0.88, 1.06)

P l a

280,886

o J

Men

Journal Pre-proof

22

Table 3: (continued) All HR (95%CI)

Tea for SSB

N

Women HR (95%CI)

2

I test for betweenstudies heterogeneity (p-value)

Men HR (95%CI)

N

f o

o r p

Crude model

0.87 (0.81, 0.94)

0.0% (0.60)

283,862

0.84 (0.75, 0.93)

7.7% (0.35)

Model 1

0.97 (0.89, 1.05) 0.94 (0.87, 1.01) 0.94 (0.87, 1.01) 0.95 (0.88, 1.02)

0.0% (0.87) 0.0% (0.54) 0.0% (0.72) 0.0% (0.69)

283,862

1.06 (0.91, 1.22) 0.94 (0.84, 1.05) 0.95 (0.84, 1.06) 0.95 (0.84, 1.06)

e

0.88 (0.82, 0.94) 1.01 (0.95, 1.09) 0.97 (0.91, 1.04) 0.96 (0.90, 1.03)

31.9% (0.19) 0.0% (0.69) 0.0% (0.69) 0.0% (0.76)

284,344

0.0% (0.75) 0.0% (0.66) 0.0% (0.61) 0.0% (0.68)

216,282

o J

0.79 (0.70, 0.89) 0.96 (0.85, 1.09) 0.97 (0.86, 1.10) 0.97 (0.85, 1.09)

0.96 (0.90, 1.03)

0.0% (0.78)

274,754

0.96 (0.84, 1.08)

0.0% (0.62)

143,184

Model 2 Model 3 Model 4

Low fat milk for SSB

2

I test for betweenstudies heterogeneity (p-value)

Crude model Model 1 Model 2 Model 3

Model 4

283,862 280,403

l a

273,788

n r u

284,344 284,344 280,885

r P

0.0% (0.95) 35.1% (0.20) 7.8% (0.35) 17.0% (0.31)

2

2

I test for betweenstudies heterogeneity (p-value)

N

P-value, I test for betweengroups heterogeneity

215,799

0.90 (0.82, 0.99)

0.0% (0.80)

68,063

0.36

215,799

0.93 (0.85, 1.03) 0.94 (0.85, 1.03) 0.94 (0.85, 1.03) 0.94 (0.86, 1.04)

0.0% (0.93) 0.0% (0.81) 0.0% (0.81) 0.0% (0.86)

68,063

0.16

68,063

0.95

67,127

0.88

131,570

0.98

0.92 0.85, 1.00) 1.04 (0.95,1.13) 0.97 (0.90,1.06) 0.96 (0.89,1.05)

33.7% (0.22) 0.0% (0.50) 1.9% (0.36) 0.0% (0.40)

68,062

0.03

68,062

0.33

68,062

0.98

67,126

0.98

0.97 (0.88,1.05)

0.0% (0.49)

131,570

0.89

215,799 213,276 142,218

216,282 216,282 213,759

Journal Pre-proof

23

Table 3: (continued) All HR (95%CI)

Whole fat milk for SSB

Crude model Model 1 Model 2 Model 3 Model 4

Total milk for SSB

Crude model Model 1 Model 2 Model 3 Model 4

2

Women HR (95%CI)

I test for betweenstudies heterogeneity (p-value)

N

1.03 (0.95, 1.12) 1.05 (0.96, 1.15) 0.94 (0.86, 1.02) 0.95 (0.87, 1.04) 0.97 (0.88, 1.06)

62.2% (0.01) 49.0% (0.07) 42.6% (0.11) 46.2% (0.08) 39.9% (0.13)

284,345

0.93 (0.87, 0.99) 0.97 (0.90, 1.03) 0.97 (0.90, 1.03) 0.96 (0.90, 1.03) 0.97 (0.90, 1.03)

10.2% (0.35) 0.0% (0.46) 0.0% (0.45) 0.0% (0.50) 0.0% (0.65)

284,345 284,345 280,886

J

u o

rn

284,344 284,344 284,344 280,885 274,754

Men HR (95%CI)

N

0.0% (0.87) 0.0% (0.83) 0.0% (0.81) 0.0% (0.84) 0.0% (0.93)

e

o r p

0.91 (0.81, 1.01) 1.00 (0.89, 1.13) 1.000 (0.89, 1.12) 1.01 (0.90, 1.13) 1.01 (0.89, 1.13)

0.0% (0.79) 0.0% (0.83) 0.0% (0.80) 0.0% (0.90) 0.0% (0.88)

216,282

r P

216,282 216,282 216,282 213,759 143,184

216,282 216,282 213,759 143,184

2

I test for betweenstudies heterogeneity (p-value)

N

P-value, I test for betweengroups heterogeneity

0.94 (0.85,1.04) 1.03 (0.93,1.15) 0.88 (0.79,0.98) 0.89 (0.80,1.00) 0.90 (0.81,1.00)

59.4% (0.09) 80.7% (0.01) 59.1% (0.09) 57.7% (0.09) 41.6% (0.18)

68,063

0.001

68,063

0.48

68,063

0.03

67,127

0.02

131,570

0.01

0.94 (0.87,1.02) 0.95 (0.87,1.03) 0.95 (0.87,1.03) 0.94 (0.87,1.02) 0.95 (0.87,1.03)

62.9% (0.07) 52.4% (0.12) 52.5% (0.12) 49.6% (0.14) 31.5% (0.23)

68,062

0.62

68,062

0.44

68,062

0.46

67,126

0.36

131,570

0.42

f o

1.28 (1.09, 1.49) 1.11 (0.94, 1.31) 1.10 (0.93, 1.29) 1.13 (0.96, 1.34) 1.16 (0.98, 1.37)

l a

274,754

2

I test for betweenstudies heterogeneity (p-value)

2

Journal Pre-proof

24

Table 3: (continued) All HR (95%CI)

Fruit juice for SSB

Crude model Model 1 Model 2 Model 3 Model 4

ASB for SSB

0.79 (0.71, 0.89) 0.96 (0.86, 1.08) 0.98 (0.88, 1.10) 0.97 (0.85, 1.07) 1.11 (1.11, 1.11)

2

Women HR (95%CI)

I test for betweenstudies heterogeneity (p-value)

N

0.0% (0.51) 3.0% (0.31) 27.6% (0.22) 0.0% (0.44) 58.6% (0.03)

284,277 284,277 284,277 280,818

l a

274,620

n r u

Crude model

0.79 (0.66, 0.95) 1.03 (0.87, 1.23) 1.08 (0.91, 1.28) 1.02 (0.85, 1.22) 0.97 (0.81, 1.16)

2

I test for betweenstudies heterogeneity (p-value) 0.0% (0.57) 0.0% (0.53) 15.1% (0.32) 0.0% (0.58) 0.0% (0.51)

f o

o r p

e

r P

Men HR (95%CI)

N

216,215 216,215 216,215 213,692 143,050

0.79 (0.69, .92) 0.92 (0.79,1.06) 0.92 (0.79, .06) 0.91 (0.79,1.06) 1.11 (1.11, .11)

2

I test for betweenstudies heterogeneity (p-value)

N

P-value, I test for betweengroups heterogeneity

39.2% (0.19) 32.4% (0.23) 27.4% (0.25) 34.0% (0.22) 80.0% (0.01)

68,062

0.98

68,062

0.31

68,062

0.16

67,126

0.35

131,570

0.14

0.84 0.0% 284,289 0.836 0.0% 216,226 0.84 0.0% 68,063 0.97 (0.78, 0.90) (0.73) (0.76, 0.93) (0.43) (0.75, 0.93) (0.79) Model 1 0.92 0.0% 284,289 0.94 27.6% 216,226 0.90 0.0% 68,063 0.61 (0.85, 0.99) (0.48) (0.84, 1.04) (0.25) (0.81, 1.01) (0.74) Model 2 0.94 19.7% 284,289 0.96 43.1% 216,226 0.91 0.0% 68,063 0.48 (0.86, 1.01) (0.29) (0.86, 1.07) (0.15) (0.81, 1.02) (0.51) Model 3 0.89 27.7% 280,830 0.91 49.4% 213,703 0.87 0.0% 67,127 0.60 (0.83, 0.97) (0.23) (0.82, 1.02) (0.12) (0.78, 0.98) (0.40) Model 4 0.88 26.8% 305,480 0.90 50.4% 173,910 0.87 0.0% 131,570 0.63 (0.82, 0.96) (0.23) (0.81, 1.01) (0.11) (0.77, 0.97) (0.46) Age at baseline (y) and the calendar year in which the baseline questionnaire was returned were entered into the model through the strata statement. Within each study, HRs with 95% CIs for the incidence of a coronary events was calculated by using Cox proportional hazards regression with time in study (y) as the time metric. The study-specific logs of HRs were weighted by the inverse of their variances, and a combined estimate of the HRs was computed by using a fixed-effects model. A crude model included only SSB and the stratification variables (smoking, physical activity, education and alcohol); model 1 further included confounding by diet and included quintiles of cereal fibers; quintiles of trans-fat; quintiles of poly-unsaturated fat/saturated fat ratio; model 2 further included total energy; model 3 further included BMI; model 4 further included baseline hypertension and high cholesterol. One serving = 355ml

o J

2

Journal Pre-proof

25

Figure 1: Association between one serving daily increase of SSBs (355ml) and the risk of developing coronary events

of

Age at baseline (y) and the calendar year in which the baseline questionnaire was returned were entered into the model through the strata statement (Model 3). Within each study, HRs with 95% CIs for the incidence of a coronary event was calculated by using Cox proportional hazards regression with time in study (y) as the time metric. The studyspecific logs of HRs were weighted by the inverse of their variances, and a combined estimate of the HRs was computed by using a fixed-effects model. Analyses were stratified by sex and adjusted for smoking, physical activity, alcohol intake, education, alcohol, diet (quintiles of cereal fiber; quintiles of trans-fat; quintiles of poly-unsaturated fat/saturated fat ratio), energy and BMI. The diamonds represent the combined hazard ratios and 95% CI. ARIC, Atherosclerosis Risk in Communities Study; ATBC, Alpha-Tocopherol and Beta-Carotene Cancer Prevention Study; HPFS, Health Professionals Follow-Up Study; IWHS, Iowa Women’s Health Study; NHS80, Nurses’ Health Study 1980; NHS86, Nurses’ Health Study 1986; WHS, Women’s Health Study

ro

Figure 2: Hazard ratio (HR) for all coronary events by substitution of one serving sugar-sweetened beverages (SSB) with one serving of coffee

-p

Age at baseline (y) and the calendar year in which the baseline questionnaire was returned were entered into the model through the strata statement (Model 3). Within each study, HRs with 95% CIs for the incidence of a coronary events was calculated by using

re

Cox proportional hazards regression with time in study (y) as the time metric. The study-specific logs of HRs were weighted by the

lP

inverse of their variances, and a combined estimate of the HRs was computed by using a fixed-effects model. Analyses were

stratified by sex and adjusted for smoking, physical activity, alcohol intake, education, alcohol, diet (quintiles of cereal

na

fiber; quintiles of trans-fat; quintiles of poly-unsaturated fat/saturated fat ratio), energy and BMI. The diamonds represent the combined hazard ratios and 95% CI. ARIC, Atherosclerosis Risk in Communities Study; ATBC, Alpha-

Jo ur

Tocopherol and Beta-Carotene Cancer Prevention Study; HPFS, Health Professionals Follow-Up Study; IWHS, Iowa Women’s Health Study; NHS80, Nurses’ Health Study 1980; NHS86, Nurses’ Health Study 1986; WHS, Women’s Health Study. One serving = 355ml

Figure 3: Hazard ratio (HR) for all coronary events by substitution of one serving sugar-sweetened beverages (SSB) with one serving of artificially-sweetened beverages Age at baseline (y) and the calendar year in which the baseline questionnaire was returned were entered into the model through the strata statement (Model 3). Within each study, HRs with 95% CIs for the incidence of a coronary events was calculated by using Cox proportional hazards regression with time in study (y) as the time metric. The study-specific logs of HRs were weighted by the inverse of their variances, and a combined estimate of the HRs was computed by using a fixed-effects model. Analyses were

stratified by sex and adjusted for smoking, physical activity, alcohol intake, education, alcohol, diet (quintiles of cereal fiber; quintiles of trans-fat; quintiles of poly-unsaturated fat/saturated fat ratio), energy and BMI. The diamonds represent the combined hazard ratios and 95% CI. ARIC, Atherosclerosis Risk in Communities Study; ATBC, Alpha-

Journal Pre-proof

26

Tocopherol and Beta-Carotene Cancer Prevention Study; HPFS, Health Professionals Follow-Up Study; IWHS, Iowa Women’s Health Study; NHS80, Nurses’ Health Study 1980; NHS86, Nurses’ Health Study 1986; WHS, Women’s Health Study. One serving = 355ml

Highlights Pooled analysis of four large cohort studies and two intervention studies



Higher consumption of sugar-sweetened (SSB) beverages is associated with greater risk of coronary

of



ro

events

Replacing SSB by artificially-sweetened beverage decreases the risk of developing coronary events



Replacing SSB by coffee decreases the risk of developing coronary events

Jo ur

na

lP

re

-p



Figure 1

Figure 2

Figure 3