16
Soft drinks and sweetened beverages and the risk of cardiovascular disease and mortality: a systematic review and meta-analysis
Running title: Soft drinks and adverse cardiovascular disease and mortality
Aditya Narain,1 Chun Shing Kwok MBBS MSc BSc,1,2 Mamas A Mamas BM BCh DPhil1,2,3
1. Keele Cardiovascular Research Group, Keele University, Stoke-on-Trent, United Kingdom
2. University Hospital of North Midlands, Stoke-on-Trent, UK
3. Farr Institute, University of Manchester, Manchester, UK
Corresponding author:
Aditya Narain
Keele Cardiovascular Research Group
Institute for Science & Technology in Medicine,
Keele University,
Guy Hilton Research Centre,
Thornburrow Drive, Hartshill,
Stoke-on-Trent, UK, ST4 7QB
Tel: +44 (0) 1782 671653
Fax: +44 (0) 1782 674467
Mail to:
Keywords: Sweetened beverages; Acute Myocardial infarction; Stroke, Mortality; Meta-analysis
Disclosures: None.
Word count: 3,251 (must re-count)
Abstract
Background: Soft drink consumption is associated with adverse health behaviours whichthat predispose to adverse coronary heartcardiovascular diseaserisk factor profiles; however, it is unclear whether their intake independently leads to an increased risk of cardiovascular events and mortality. It is controversial whether consumption of sweetened beverages is associated with an increased risk of adverse cardiovascular events and mortality. AWe conducted a systematic review and meta-analysis was conducted to evaluate evaluate the relationship between sweetened beverages and cardiovascular events and mortalitythis.
Methods: Medline and EMBASE were searched in July 2015 for studies that considered soft drink intake their association with and risk of mortality, myocardial infarction (MI) or stroke. Pooled risk ratios for adverse outcomes were calculated using inverse variance with a random effects model, and heterogeneity was assessed using the I2 statistic.
Results: A total of 87 prospective cohort studies with 308,810420 participants (age range 34-75 years) were included in the review. The pooled results suggest a significant increase in stroke greater risk of stroke (RR 1.13 95%CI 1.02-1.24), and MI (RR 1.22 95%CI 1.14-1.30), but not vascular events with incremental increase in sugar-sweetened beverage (SSB) consumption. For stroke, there were significant increases with greater incremental increase in artificially sweetened beverage (ASB consumption (stroke RR With incremental increase in artificially sweetened beverage (ASB) consumption, there was a greater risk of stroke (RR 1.08 95%CI 1.03-1.14), but not vascular events or MI. In the evaluation of high versus low SSB, there was a significant increase a greater risk of MI (RR 1.19 95%CI 1.09-1.31) but not for stroke, vascular events or mortality. For ASB, there was a significantly increase in greater risk of stroke (RR 1.14 95/%CI 1.04-1.26) and vascular events (RR 1.44 95%CI 1.02-2.03) but not MI or mortality.
Conclusions: Our results suggest an possible association between increased cardiovascular risk and greater consumption of sugar-sweetened and artificially sweetened beverages and cardiovascular risk, although consumption may be a surrogate for adverse health behaviors. However, these findings are based on observational studies with methodological limitations so further research is warranted to clarify this relationship.
Review Criteria
Inclusion criteria: Studies which evaluated the association between soft drink consumption and adverse cardiovascular outcomes were selected. No restrictions were placed on definition or method of ascertaining soft drink consumption. Studies had to evaluate two or more groups of beverage consumers and adverse outcomes (myocardial infarction, stroke, cardiovascular events or mortality). There was no restriction based on study design, study cohort and language of study report.
Message for the Clinic
Our results suggest that soft drink consumption is associated with greater increased risk of cardiovascular events. The risk increases with dose and applies to is greater for each extra serving consumed per day of both sugar- and artificially sweetened beverages. People who are at risk of, or already have, cardiovascular disease should consider limiting consumption of sugar-sweetened beverages.
Introduction
The burden associated with cardiovascular disease is significant. Since it is the number one cause of mortality worldwide, there is an increasing emphasis on primary and secondary prevention; expenditure within the European Union totals nearly €196 billion a year [1-3]. Soft drinks can be categorisedcategorized into sugar-sweetened beverages (SSB) or artificially sweetened beverages (ASB). The impact of SSB intake on cardiovascular disease and mortality is garnering interest in the media and amongst policy makers, with several international societies recommending a reduction in SSB consumption [4,5]. Currently, there are no guidelines related to artificially sweetened beverage consumption.
Sugar-sweetened beverage consumption Consumption of both sugar-sweetened and artificially sweetened beverages has been is associated with a greater risk of developing an adverse with the development of cardio-metabolic risk factor profiles. A meta-analysis of 7 prospective cohort studies found one daily serving increment of SSBs to be associated with 0.22kg (95%CI 0.09-0.34kg) weight gain over one year [6]. Both SSB and ASB intake has been associated with a greater risk of incidencet of type-2 diabetes independently of adiposity; a meta-analysis of 17 prospective cohort studies found a higher consumption of SSB/ASB by one serving per day conferred a greater risk of developing type-2 diabetes (RR 1.28 95%CI 1.12-1.46; RR 1.29 95%CI 1.08-1.54 respectively). such as weight gain and type-2 diabetes [6,[7]. Similarly, associations between artificially sweetened beverage intake and type-2 diabetes mellitus have been characterised [7]. In addition, a meta-analysis of 6 prospective cohort studies including 240,726 participants found SSB intake to be positively associated with risk of hypertension (RR 1.10 95%CI 1.06-1.15) [8]. Similarly, an analysis of three large prospective cohorts found SSB/ASB consumption to confer a greater risk of incident hypertension (aHR 1.13 95%CI 1.09-1.17; aHR 1.14 95%CI 1.09-1.18 respectively) both types of sweetened beverage were found to be associated with an increased risk of incident hypertension in a study investigating three large prospective cohorts [89].
Despite the well-characterisedcharacterized association between sweetened beverage intake and development of an adverse cardio-metabolic risk factorsfactor profile, the relationship between sweetened beverage intake and cardiovascular disease and mortality is more controversial. A study of 13,624 individuals from a Californian retirement community found those in the highest tertile of ASB, but not SSB, consumption to have a greater an increased risk of mortality with artificially sweetened beverage intake (RR 1.18 95%CI 1.07-1.30), but not with sugar-sweetened beverage intake [910]. These findings were not reflected in 2 prospective cohort studies (the Health Professionals’ Follow-up Study and Nurses’ Health Study), where individuals who consumed ≥2 servings of SSB per day had a greater risk of mortality (HR 1.18 95% CI 1.04-1.33) than those who consumed <1/month; consumption of ASBs had no effect on mortality risk. However, conflicting results were observed in 2 prospective cohort studies (the Health Professionals’ Follow-up Study and Nurses’ Health Study) where an increased risk of mortality with sugar-sweetened beverage intake (HR 1.18 95% CI 1.04-1.33), but not with artificially sweetened beverage consumption was observed [1011]. There are further studies, such as the Singapore Chinese Health Study, which included 52,584 participants and reported no significant association between sugar-sweetened beverages intake and mortality [1112].
With reference to cardiovascular disease, a prospective cohort study of 42,883 malesby de Koning et al. in 2012 found found individuals SSB, but not ASB, consumption to be associated with a greater risk of an increased risk of coronary heart disease with SSB intake (RR 1.18 95%CI 1.06-1.31), but not with ASB consumption [1213]; ]. whereas In contrast, a prospective cohort study of 39,786 Japanese participants by Eshak et al. in 2012 found no association between sweetened beveragesoft drink consumption and ischemic heart disease after 18 years of follow-up [1314].
In view of the inconsistent findings regarding the association betweenwhether sweetened beverage consumption is independently associated with and adverse cardiovascular outcomes, we conducted a systematic review and meta-analysis to study the impact of both sugar- and artificially sweetened beverage intake on the risk of adverse cardiovascular adverse outcomes and mortality.
Methods
Eligibility criteria
Prospective Sstudies which evaluated the association between soft drink consumption and adverse cardiovascular outcomes were selected. No restrictions were placed on definition or method of ascertaining soft drink consumption. Studies had to evaluate two or more groups of beverage consumers and adverse outcomes (myocardial infarction (MI), stroke, cardiovascular events or mortality). There was were no restrictions based on based on study design (e.g. retrospective, prospective), study cohort and or language of study report.
Search strategy
MEDLINE and EMBASE were searched using Ovid SP with no date or language restriction in July 2015 using the broad search terms in Figure 1. Furthermore, researchers screened bibliographies of potentially relevant studies and reviews for additional studies.
Study selection and data extraction
Two reviewers (AN and CSK) screened all titles and abstracts for studies that met the inclusion criteria and excluded any articles that did not clearly fulfil the selection criteria. Full-texts of potentially relevant studies were downloaded and assessed. Final decisions on inclusion or exclusion were made with discussion with a third reviewer (MAM). Two extractors (AN and CSK) independently collected data from included studies on study design, study date, country of origin, sample size, participant age, gender and inclusion criteria, population characteristics, risk of bias, definition of sweetened beverage or soda consumption, follow-up and results.
Risk of bias assessment
Risk of bias was evaluated by considering 5 categories: reliable exposure ascertainment, reliable outcome ascertainment, lost loss-to follow up, use of adjustment for potential confounders and generalisabilitygeneralizability of findings.
Publication bias was considered using asymmetry testing if there were more than 10 studies in a particular the meta-analysis, and if statistical heterogeneity was <50% [1415].
Data analysis
RevMan 5.025 (Nordic Cochrane Centre) was used to conduct random effects meta-analysis using inverse variance method for pooled risk ratios (RR). Where possible, adjusted risk estimates from the primary studies were pooled, otherwise raw outcome data was used to yield unadjusted risk estimates. As continuous continues variables were available, analysis evaluating risk of adverse outcomes per unit or serving increase of soft drink intake was performed. Furthermore, where multiple groups were reported, extreme groups (i.e. the highest quantity of consumption of sweetened beverages or soft drinks compared to the lowest) were pooled, as this increased likelihood of detecting potential associations. The analysis was split into sugar-sweetened beverage consumption and artificially sweetened (diet or low-calorie) beverage consumption. A sensitivity analysis was performed excluding studies which did not adjust for body mass index.
Statistical heterogeneity
Statistical heterogeneity was assessed using I2 statistic, with I2 values of 30%–60% representing a moderate level of heterogeneity [1516]. If I2 was above 60% for the pooled analysis, we sought to explore sources of heterogeneity in subgroups of studies.
Results
From the total 303 studies screened for this review, 10 9 articles.[910-1314,1617-2021] were included. These reported results from 8 7 unique studies (Nurse’s Health Study and Health Professionals Follow-up Study were reported on multiple occasions for different outcomes) [6,1213,1617,1718]. The process of study selection is shown in Figure 1.
Study design and participant characteristics
All included studies were prospective cohort studies except for the case-control study by Nicknam et al [20]. Although the majority of studies were conducted in the USA, some were conducted in Japan, Sweden, Iran and Singapore. These studies took place between 1980 and 2011. The total number of participants from the 8 7 unique studies was 308,810420. The number of participants in each study ranged from 390 2,564 to 88,520 and ages of participants in the 5 studies that reported age ranges was between 34 and 75 years. The percentage of male participants in the studies ranged from 32% to 5452%, aside from two studies, one comprising of all males and the other all females. The study design and participant characteristics are shown in Table 1. Supplementary Table 1 shows participant characteristics according to beverage consumption. Participants who consumed higher amounts of sugar-sweetened beverages were more likely to be smokers and exercise less.
Risk of bias
Table 2 shows the risk of bias of included studies. All 8 7 studies used a food frequency questionnaire or survey to evaluate beverage intake. Outcome ascertainment was reliable in all studies; methods used include direct contact with patients, evaluation of medical records and checking death registries. Outcomes studied were MI, stroke and mortality. Of the 8 7 unique studies, the percentage of participants lost to follow-up was unclear in 54; the remaining three studies reported a loss to follow-up <10%. All studies adjusted for age, sex, smoking and physical activities in their analysis. Other frequently adjusted variables include body mass index, alcohol intake, diet and family history. One study which did not adjust for body mass index was excluded in a sensitivity analysis. A total of 3 studies were conducted in general populations, whereas the remainder used specific cohorts such as male health professionals or female nurses, or ethnicities such as Chinese, Iranian orChinese or Japanese populations.
Study results
Table 3 shows the results of included studies. Sweetened beverages were categorized most commonly into sugar-sweetened and artificially sweetened. Artificially sweetened beverages included low-calorie and diet sodas. A total of two studies use MI as an outcome, while two other studies use coronary heart disease (defined as fatal or nonfatal MI). Stroke featured as an outcome of 4 studies, and mortality in 3. Follow-up ranged from a case-control 9.8study to 24 years. 3 studies provided data to show the effects of an incremental increase in sweetened beverages on MI, stroke and mortality. All 8 7 studies provided a high versus low table in their results.
The Nurse’s Health Study and the Health Professionals Follow-up Study cohorts were investigated in more than one included study. These articles were included despite studying the same cohorts because each study evaluated different outcomes.
Studies of incremental increase in sugar sweetened beverages or artificially sweetened beverages and adverse outcomes
Figure 2 shows the risk of adverse outcomes with incremental increase in sugar sweetened beverage intake. The pooled results suggest a one-serving per day increase in SSB consumption was associated with a greater risk of stroke (RR 1.13 95%CI 1.02-1.24) and MI (RR 1.22 95%CI 1.14-1.30), but not of vascular events. significant increase in stroke RR 1.13 95%CI 1.02-1.24, and MI RR 1.22 95%CI 1.14-1.30, but not vascular events with sugar sweetened beverage consumption. For these analyses there was little evidence of statistical heterogeneity as the respective I2 values for stroke and coronary heart disease MI were 0% and 8% respectively.