Open Access
Open Peer Review

This article has Open Peer Review reports available.

How does Open Peer Review work?

Do sugar-sweetened beverages cause adverse health outcomes in adults? A systematic review protocol

  • Candyce Hamel1,
  • Adrienne Stevens1,
  • Kavita Singh2,
  • Mohammed T Ansari1,
  • Esther Myers3,
  • Paula Ziegler4,
  • Brian Hutton1, 15,
  • Arya Sharma5,
  • Lise M Bjerre6,
  • Shannon Fenton7,
  • David CW Lau8,
  • Kathryn O’Hara9,
  • Robert Reid10,
  • Erinn Salewski11,
  • Ian Shrier12,
  • Noreen Willows13,
  • Mark Tremblay14 and
  • David Moher1, 15Email author
Systematic Reviews20143:108

https://doi.org/10.1186/2046-4053-3-108

Received: 15 July 2014

Accepted: 28 August 2014

Published: 23 September 2014

Abstract

Background

Chronic diseases, such as cardiovascular disease and type 2 diabetes, impose significant burden to public health. Most chronic diseases are associated with underlying preventable risk factors, such as elevated blood pressure, blood glucose, and lipids, physical inactivity, excessive sedentary behaviours, overweight and obesity, and tobacco usage. Sugar-sweetened beverages are known to be significant sources of additional caloric intake, and given recent attention to their contribution in the development of chronic diseases, a systematic review is warranted. We will assess whether the consumption of sugar-sweetened beverages in adults is associated with adverse health outcomes and what the potential moderating factors are.

Methods/Design

Of interest are studies addressing sugar-sweetened beverage consumption, taking a broad perspective. Both direct consumption studies as well as those evaluating interventions that influence consumption (e.g. school policy, educational) will be relevant. Non-specific or multi-faceted behavioural, educational, or policy interventions may also be included subject to the level of evidence that exists for the other interventions/exposures. Comparisons of interest and endpoints of interest are pre-specified. We will include randomized controlled trials, controlled clinical trials, interrupted time series studies, controlled before-after studies, prospective and retrospective comparative cohort studies, case-control studies, and nested case-control designs. The MEDLINE®, Embase, The Cochrane Library, CINAHL, ERIC, and PsycINFO® databases and grey literature sources will be searched. The processes for selecting studies, abstracting data, and resolving conflicts are described. We will assess risk of bias using design-specific tools. To determine sets of confounding variables that should be adjusted for, we have developed causal directed acyclic graphs and will use those to inform our risk of bias assessments. Meta-analysis will be conducted where appropriate; parameters for exploring statistical heterogeneity and effect modifiers are pre-specified. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach will be used for determining the quality of evidence for outcomes.

Systematic review registration

PROSPEROCRD42014009638

Keywords

Sugar-sweetened beverage Adult Chronic disease Obesity Cardiovascular disease Metabolic syndrome Type 2 diabetes Cancer Systematic review

Background

Chronic diseases, such as cardiovascular disease and type 2 diabetes, cause significant morbidity and mortality worldwide. To address the burden of chronic diseases on the health of populations, effective interventions and public health policies are required. Most chronic diseases are associated with underlying preventable risk factors, such as elevated blood pressure, high blood glucose or glucose intolerance, hyperlipidemia, physical inactivity, excessive sedentary behaviours, overweight and obesity, and tobacco usage. The development of chronic diseases may be prevented if these risk factors are addressed before they progress to overt disease. A simple, unidirectional schematic depicts the hypothesized pathways by which sugar-sweetened beverage (SSB) consumption may lead to the development of chronic cardiovascular/cerebrovascular and metabolic diseases, chronic kidney disease, cancer, and gout (Additional file1). These mechanisms have not been conclusively established by research studies, and several conflicting theories have been put forward[18]. However, as some of the SSB constituents, notably sugar but also caffeine and by-products of caramel colouring, where added, are postulated to be involved in the mechanisms of disease development, they are included in our depiction of the disease pathways.

Recent attention has focused on the contribution of SSBs to the chronic disease epidemic. In an attempt to curb rising rates of obesity, New York City has proposed a ban on the sale of large-sized SSB products greater than 16 oz, including sodas, sweetened teas and coffees, energy drinks, and fruit drinks in restaurants, delis, sports arenas, movie theatres, and food carts[9]. The proposed ban has been twice rejected but is being reviewed by the United States (US) Court of Appeals. In March 2014, the World Health Organization (WHO) released draft guidelines with recommendations on limiting sugar consumption (through food and beverage) to reduce public health problems like obesity and dental caries. They are recommending decreasing the total energy intake of sugar by day from 10% (recommended since 2002) to 5%[10].

Some researchers hypothesize that SSBs are significant sources of caloric intake. With the adjusted prevalence of SSB consumption at 73% for young adults (20–34 years) and 50% among adults (35 years or older) according to the 2007–2008 US National Health and Nutrition Examination Survey (NHANES) data, the mean adjusted SSB energy intake was 338 and 236 kcal/day, respectively[11]. Heavy consumption (≥500 kcal/day) occurred among 20% of young adult and 12% of adult consumers[11]. Soda was the most heavily consumed SSB[11].

Several reviews have presented evidence syntheses on associations between SSB consumption and metabolic syndrome/type 2 diabetes,[1214] weight,[12, 1417], and cardiovascular disease[12]. In a meta-analysis of 11 prospective cohort studies, the highest quantile of SSB consumption (one to two servings/day) was associated with a statistically significant increased risk of developing type 2 diabetes (risk ratio (RR) = 1.26, 95% confidence interval (CI): 1.12–1.41) and metabolic syndrome (RR = 1.20, 95% CI: 1.02–1.42) in adults[13]. Vartanian et al. found associations between SSB consumption and increased energy intake and body weight, lower intake of calcium and other nutrients, and increased risk of medical problems such as type 2 diabetes, hypocalcemia, dental caries, and elevated blood pressure[15]. A World Health Organization report concluded that the evidence to implicate high intake of sugar-sweetened drinks on weight gain is moderately strong[18]. In addition, in a systematic review conducted as part of the Dietary Guidelines Advisory Committee deliberations, the following conclusion statements were provided[19]:

  •  ‘Limited evidence shows that intake of sugar-sweetened beverages is linked to higher energy intake in adults’.

  •  ‘A moderate body of epidemiologic evidence suggests that greater consumption of sugar-sweetened beverages is associated with increased body weight in adults’.

  •  ‘A moderate body of evidence suggests that under isocaloric controlled conditions, added sugars, including sugar-sweetened beverages, are no more likely to cause weight gain than any other source of energy’.

Other reviews, however, have concluded that the association between SSB consumption and weight is close to zero[16, 20]. Some reviews report a dose-response relationship between SSB and weight status but no corresponding weight loss when SSB consumption was reduced[21]. A systematic review of epidemiological research analysed beverages by category (water, milk, soft drinks, sugary drinks, non-carbonated, fruit juices, carbonated beverages, hot beverages, and alcoholic beverages), and the authors concluded that the results were inconsistent and did not establish an association between beverage intake and subsequent weight gain[22]. Within their beverage categories, however, the authors did not consistently separate SSBs from other drinks; for example, sweetened hot beverages were not differentiated from non-sweetened. Others have reviewed the biological plausibility of SSBs to uniquely affect the physiological energy balance regulatory systems (e.g. satiety and post-prandial regulatory systems) and concluded that known biological mechanisms did not support the concept that SSBs were somehow different from other sources of energy[23].

It is unclear what role SSBs play in the context of total caloric intake. Although total caloric intake and caloric intake from other dietary sources are factors that may be thought of as confounders, some evidence suggests that they may be clustered with SSB consumption and general unhealthy eating habits[12, 17, 24]. Conflicting evidence exists as to whether SSB intake is associated with increased energy intake[15, 19] and may lower the intake of milk, calcium, and other nutrients[15]. Furthermore, given that SSB consumption may alter taste preferences and quality of diet, caloric intake may in fact mediate the effect of SSB consumption and health outcomes (i.e. it may lie in the causal pathway).

The available systematic review evidence is not only conflicting but fraught with several methodological issues, and it is difficult, therefore, to make any firm conclusions about the health effects of SSBs[25]. The definition of what constitutes an SSB, for example, may vary and may not be explicitly described[1517, 26, 27]. Reviews may not have accounted for all variables that can confound associations between SSB consumption and health outcomes. Most reviews have addressed singular outcomes of interest to public health. Further, new primary evidence[2830] on the topic is rapidly accumulating since the last systematic review was published in 2011. We will attempt to overcome some of the methodological challenges encountered in other reviews by carrying out a rigorous assessment of the internal validity of individual studies, which will include an evaluation of confounding and biasing factors using causal directed acyclic graphs (DAGs)[31].

Objectives

The objective of the systematic review is to answer the following research questions:

In adults, does the consumption of SSBs cause adverse health outcomes? If so, what potential moderating factors affect the causal association between SSB consumption and outcomes?

Methods/Design

Eligibility criteria

Studies will be selected according to the criteria outlined below.

Study designs

We will include randomized controlled trials (RCTs), including cluster RCTs, controlled (non-randomized) clinical trials (CCTs) or cluster trials, interrupted time series (ITS) studies with at least three data points before and after the intervention[32], controlled before-after (CBA) studies, prospective and retrospective comparative cohort studies, and case-control or nested case-control studies. Cluster randomized, cluster non-randomized, or CBA studies will be included only if there are at least two intervention sites and two control sites[32]. We will exclude cross-sectional studies, case series, and case reports.

Participants

We will include studies examining the general adult human population or healthy adult humans (18 years or older). We will also include studies on people who are overweight or obese, but will otherwise exclude studies of populations restricted to specific diseases, conditions, or metabolic disorders. We will include studies addressing both adults and children if data provided for adults are reported separately.

Interventions

Of interest are interventions addressing SSB consumption, taking a broad perspective. In addition to direct consumption studies, we would consider interventions that influence consumption, such as those addressing the level of access to SSBs (e.g. university/college policy) and educational interventions addressing consumption as relevant. Non-specific or multi-faceted behavioural, educational, or policy interventions may also be included subject to the level of evidence that exists for the aforementioned interventions/exposures. We will also consider other types of interventions on a case-by-case basis, subject to what exists in the literature.

In terms of defining an SSB, we view them as akin to a complex intervention because they are composed of several parts. For example, in addition to sugar, some beverages contain caffeine and the by-products of caramel colouring (2-methylimidazole, 4-methylimidazole), which may contribute independently to adverse health outcomes. The scope of the review, therefore, warrants an examination of SSB consumption as a whole, rather than the specific constituents as exposure variables. Otherwise, such evaluations would have necessarily required the inclusion of studies addressing those constituents and in foods and drinks other than SSBs.

We will use the Centers for Disease Control and Prevention (CDC) definition of SSB for drinks that should be included. According to the CDC, SSBs contain added caloric sweeteners[33], which would include natural sweeteners such as honey and concentrated fruit juice. We have developed a classification scheme based on the CDC definition for use during the review (see classification scheme for SSBs below). For beverages such as coffee, tea, and homemade lemonade, studies will be included in the review if they explicitly state that sugar was added. We will exclude artificially sweetened (e.g. with aspartame or sucralose) beverages, alcoholic beverages, and 100% fruit or vegetable juices as exposures/interventions.

We will classify SSBs described in studies according to the following broad categories:

  •  Sodas—caffeinated/non-caffeinated (soft drinks, soda, pop, soda pop)

  •  Other non-carbonated sweetened beverages (fruitades, fruit drinks, fruit punches, [iced] teas, coffees, non-dairy fruit smoothies)—caffeinated/non-caffeinated

  •  Fortified sweetened beverages (energy drinks, fortified waters, sports drinks)—caffeinated/non-caffeinated and containing vitamins, amino acids, herbal stimulants, or other ingredients

  •  Flavored/sweetened milk or milk alternative beverages (dairy, soy, almond, milkshakes, dairy-based fruit smoothies)—caffeinated/non-caffeinated

Comparators

Given the broad perspective for interventions of interest, several comparisons will be relevant to include. Some may be more likely to come from observational designs and others from experimental studies.

Direct consumption studies:
  1. 1.

    SSB consumption compared with consumption of non-SSB drink (e.g. 100% fruit juice, artificially sweetened beverage, water)

     
  2. 2.

    Higher level of SSB consumption versus lower level of SSB consumption for the same drink type (e.g. carbonated cola beverages)

     
  3. 3.

    Comparisons among different categories of SSBs (e.g. soft drinks compared with fruit drinks; see classification scheme for SSBs) consumed in similar amounts

     
Interventions that influence consumption:
  1. 4.

    One level of access to SSB compared with another level of access (e.g. university/college policy on beverages in vending machines)

     
  2. 5.

    Educational intervention to specifically promote lower or no SSB consumption compared with no educational intervention/regular curriculum coverage/general health-focussed intervention

     
  3. 6.

    Non-specific or multi-faceted educational, behavioural, or policy dietary intervention (may include component of SSB consumption) compared with no intervention

     
  4. 7.

    Other comparisons involving interventions that address our research question (interventions assessed on a case-by-case basis, as encountered in the literature)

     

For comparator groups 2 and 3, we anticipate that volume will be the most feasible to analyse; however, we will extract all measures in which consumption is reported (e.g. volume, caloric intake from sugar) in studies to see what analysis is possible.

For feasibility, category 6 comparisons (non-specific, multi-faceted interventions) will be coded at title/abstract screening and not put through to full-text screening. If sparse evidence exists in the other potential comparison types, we will revisit eligibility for comparison 6.

Outcomes

Endpoints important for decision-making are of primary interest. If reported on, these will be analysed and graded. If a given clinical endpoint is not reported on, we will analyse and grade their relevant surrogate outcome(s).

  •  Endpoints important for decision-making:
    • Adverse cardiovascular (including cerebrovascular) events

    • Cancer (excluding basal cell and squamous cell carcinoma)

    • Chronic kidney disease

    • Mortality

    • Overweight/obesity

    • Type 2 diabetes

    • Dental caries

    • Quality of life (generic, validated tools only, such as those in Additional file 2)

    • Gout

  •  Surrogate outcomes:
    • Pre-diabetes

    • Metabolic syndrome

    • Change in cardiovascular disease (CVD) risk

    • Progression of obesity

    • Dyslipidemia

    • Hypertension

As some outcomes may be reported as a composite measure, we will extract all composite and individual outcomes as reported in the studies.

Outcomes will be collected as reported, with the exception of quality of life, which will be collected only if assessed with generic (not disease-specific), validated tools. Due to possible variation in disease definitions over time, we will extract definitions of outcomes as reported in individual studies. We will extract outcomes in all data forms (e.g. dichotomous, continuous) as reported in the included studies.

Timing

Studies will be selected for inclusion based on the length of follow-up of outcomes. The following will be used as a guide for all study designs:

  •  For all decision-making endpoint outcomes, studies should have a follow-up time of at least 1 year.

  •  For all surrogate outcomes, studies should be at least 6 months duration for follow-up.

  •  For cancer, studies should be at least 1 year duration for follow-up. Some types of cancer may need longer than a 1-year follow-up, but this will be evaluated on a case-by-case basis.

Setting

There will be no restrictions by type of setting.

Language

We will include articles reported in the English and French languages. A list of possibly relevant titles in other languages will be provided as an appendix.

Search methods

Electronic searches

A comprehensive literature search using high-recall subject searches will be conducted in MEDLINE®, Embase, CINAHL, PsycINFO®, ERIC, and The Cochrane Library. Electronic search strategies (Additional file3), guided by an experienced information specialist, have been developed and peer-reviewed according to the peer review of electronic search strategies (PRESS) guidelines[34]. The search will not be restricted for time period or the language of publication. We will exclude comments, letters, and editorials.

Other sources

Grey literature sources, such as websites listed within the Canadian Agency for Drugs and Technologies in Health’s (CADTH) Grey Matters checklist, the Federation of American Societies for Experimental Biology (FASEB) meeting abstracts, the Obesity Society abstracts, the Food and Nutrition Conference and Expo (FNCE) abstracts, the International Diabetes Federation website, the Centers for Disease Control and Prevention website, the Rudd Center for Food Policy and Obesity website, the Obesity and Energetics Offerings website, the American Heart Association, the European Union, the Center for Science in the Public Interest, the 20th International Congress of Nutrition, the American Beverage Association, and Refreshments Canada will be searched. We will consult within the research team for relevant studies that may not be in the published literature. We will scan the references of included studies and relevant reviews for additional articles and perform a forward search on key articles. We will search ClinialTrials.gov, the WHO International Clinical Trials Registry Platform, and the International Association for the Study of Obesity (IASO) for completed and ongoing studies.

Study selection process

Literature search results will be de-duplicated in Reference Manager[35] before uploading to Distiller Systematic Review Software® (Distiller SR), an online program that facilitates screening and data extraction[36]. Screening questions will be developed and pilot-tested with a subset of records before implementation. All titles and abstracts of records will be screened by one person; those deemed not relevant will be verified by a second person for exclusion. Full-text reports for all potentially relevant records and those without an available abstract will be screened by two independent reviewers. Discrepancies will be resolved by consensus or a third person. The study selection process will be reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram[37], including reasons for excluding full-text articles.

Data collection and analysis

Data extraction and management

Feedback will be solicited from the research team on the draft list of data variables for extraction. Data extraction forms will be developed and pilot-tested in Distiller SR. One person will extract all information. A second person will verify 20% of studies for general characteristics information and 100% of studies regarding outcomes data. Disagreements will be resolved by consensus or by a third team member, if needed. Information on the descriptive and quantitative characteristics of studies will include the following:

  •  Publication details (e.g. year of publication, language, publication status)

  •  Characteristics of study (e.g. study design, methods, country, setting, sample size, number of centres [if applicable], duration of follow-up, source of funding)

  •  Characteristics of population (e.g. age, gender, ethnicity, co-interventions, information regarding respondent bias/representativeness of the included population)

  •  Details about the exposure/intervention (e.g. type and category of SSB [see classification scheme for SSBs]), brand name, amount of sugar/serving, type of added sugar, amount of caffeine, other specific ingredients and their quantification, frequency of use, amount consumed in millilitres, SSB caloric content, percent of total calories obtained from SSB consumption, method of assessing SSB consumption; type of educational or other interventions and description, type of professional delivering intervention)

  •  Details about comparator group (e.g. for beverages: identity, frequency of use, amount used, brand name, specific ingredients)

  •  Outcomes of interest for the longest duration of follow-up (definitions, measurement methods, data, adjusted and unadjusted effect estimates)

  •  Confounding factors that were taken into consideration

  •  Risk of bias items

Assessing the risk of bias

The risk of bias for each included study will be assessed by one member of the research team and verified by a second member. Disagreements will be resolved by consensus or by a third team member, if needed. Assessment tool questions were reorganized (as needed) to ensure that domains relating to selection, confounding (where applicable), performance, attrition, detection, reporting, and ‘other’ biases were addressed in all tools. A modified version of the Cochrane Risk of Bias tool will be used to evaluate RCTs (Additional file4)[38]. In addition to the standard domains of bias, we will assess cluster trials for the possibility of recruitment bias[39]. A modified version of the Academy of Nutrition and Dietetics’ Evidence Analysis Library (EAL) Quality Criteria Checklist will be used to evaluate observational studies and CCTs (Additional file4)[40]. We removed questions pertaining solely to reporting characteristics, and we added a few other relevant questions. To evaluate ITS and CBA studies, a modified Cochrane Effective Practice and Organisation of Care (EPOC) tool will be used (Additional file4)[41]. Study sponsorship will be assessed for all studies. Each domain within a tool will be judged as unclear, low, or high risk of bias, with supporting information provided from the report or reviewer interpretation to rationalize the judgement of bias[38]. Some domains are outcome-specific and will be assessed at the outcome level. The risk of bias for outcomes will be factored into grading the quality of evidence. The overall risk of bias for the body of evidence will involve a judgement of the relative importance of domains, guided by known empirical evidence of bias, the likely direction of bias, and the likely magnitude of bias[38]. We will follow the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidance for determining the extent of the risk of bias for the body of evidence[42].

Step 1—Assessment of risk of bias for individual studies for a given outcome

  •  Low risk of bias: when all key domains are at a low risk of bias

  •  Moderate risk of bias: crucial limitation for one domain or some limitations for multiple domains sufficient to lower confidence in the effect estimate

  •  High risk of bias: crucial limitation for one or more domains sufficient to substantially lower confidence in the effect estimate

Step 2—Assessment of risk of bias for the body of evidence across studies for a given outcome (incorporated into GRADE assessments as one of the required domains)

  •  No serious limitation, do not downgrade: body of evidence mostly from studies at a low risk of bias

  •  Serious limitations, downgrade one level: body of evidence mostly from studies at moderate risk of bias

  •  Very serious limitations, downgrade two levels: body of evidence mostly from studies at high risk of bias

Regarding confounding bias, causal DAGs will be used[31]. Causal DAGs are graphical models used in epidemiology to determine sets of confounders that should be adjusted for to obtain unbiased effect estimates. They also identify biasing paths associated with selection bias. By adjusting for only those confounders that are derived from a causal DAG, the potential for over-adjustment or the creation of selection bias by conditioning on colliders (i.e. variables that are common effects of both the exposure and outcome) are reduced, given that the causal DAG is correct[43, 44]. We will use the minimal sufficient adjustment sets generated from the causal DAGs as a guide to determine if studies have adequately accounted for confounding variables or created selection bias by over-adjustment of imbalances between exposure and control groups. For a given outcome, if studies differentially account for variables from the minimum adjustment sets, we may not pool those studies in a meta-analysis. Draft causal DAGs (Additional file5) have been developed using the DAGitty program[45]. The causal DAGs may be modified to include important and justifiable variables we encounter when reviewing included studies.

Dealing with missing data

If information or data are missing or incomplete, we will attempt to contact the study authors twice over 2 weeks by email. If feasible, we will incorporate loss-to-follow-up data. We will not impute effect estimates, but will impute missing standard deviations or standard errors using data from other similar studies in the review, using an approach suggested in the literature[46].

Data analyses

For dichotomous outcomes, the risk ratio or odds ratio and 95% confidence intervals will be used for pooling. For continuous outcomes, mean differences and 95% confidence intervals will be used for pooling for outcomes reported on the same scales or measured in the same units. Standardized mean differences will be used for pooling where continuous outcomes are reported using different scales or measures. Transformation of data to allow analyses with mean differences will be made, wherever possible. All formats of continuous outcome data will be extracted whether reported as post-intervention or change from baseline. We will consider using the r value, a correlation coefficient, for continuous variables when some studies analyse as a dichotomous outcome and others analyse as a continuous outcome. Statistics from individual studies will be converted to an r value before meta-analysing[47, 48]. The r value can be roughly interpreted as a small (r = 0.1), medium (r = 0.3), or large (r = 0.5) effect size. We plan to transform the pooled r to another statistic, such as an odds ratio, to aid in interpretation[47]. For time-to-event data, the hazard ratio, which is usually estimated from a Cox proportional hazards model, will be pooled using the generic inverse variance method[49]. Comprehensive Meta-Analysis will be used for meta-analyses[50]. SAS software will be used for meta-regression analyses[51].

Other statistical considerations

Sparse binary data and studies with zero events When studies report rare events, narrative synthesis will suffice. When event rates are less than 1%, the Peto odds ratio method will be used. However, when control groups are of unequal sizes, when large magnitude of effect is observed, or when events become more frequent (5%–10%), the Mantel-Haenszel method without correction factor will be employed for quantitative synthesis[49].

Data conversions Where needed, we will convert data (e.g. standard error to standard deviation) for use in analyses and to facilitate consistent presentation of results across studies.

Interrupted time series designs If interrupted time series studies are included, we will re-analyse data, where needed and if feasible, for change in level and slope according to time series regression analyses[52].

Evidence synthesis

Study characteristics will be summarized narratively in the text and shown in summary tables in the report. Before meta-analyses are performed, studies will be assessed for heterogeneity on clinical and methodological characteristics; we plan to review these decisions with the research team before conducting analyses. For outcomes that can be measured on various scales (e.g. quality of life), heterogeneity of outcome measurements will also be assessed before pooling. With sufficient homogeneity and quantity of data, we will pool studies using standard random effects meta-analytic methods[53, 54]. Any meta-analyses will be done separately for observational studies and experimental ones. Narrative synthesis of data will be conducted when quantitative pooling is considered inappropriate (team decision based on the aforementioned issues, such as disparate clinical characteristics of included patients/participants). When important clinical or methodological heterogeneity precludes pooling, we may still present forest plots without a pooled summary estimate to show individual study effects. Effect estimates from observational studies at a high risk of bias may be excluded from the evidence synthesis when their findings are inconsistent with studies at a lower risk of bias. If only observational studies at a high risk of bias exist for a given outcome, we will not synthesize those studies because they are unlikely to inform about causality.

Statistical heterogeneity and effect modifiers

Statistical heterogeneity will be assessed using Cochrane Q (considered statistically significant at p < 0.10) and I-squared statistics. For the interpretation of I-squared, a rough guide of low (0%–25%), moderate (25%–50%), substantial (50%–75%), and considerable (75%–100%) heterogeneity will be used[55, 56]. When a body of evidence is determined to be statistically heterogeneous, we plan to explore the impact of moderating factors using a combination of subgroup analysis and meta-regression techniques, where the optimal approach for each variable will be determined once we see how data are reported in studies. Some variables (*) will be investigated as potential effect modifiers regardless of heterogeneity tests. The distribution of several aspects of patient demographics, exposure, and other characteristics will be of interest and include the following:

  •  Participant age*

  •  Participants who are post-menopausal women

  •  Sex*

  •  Ethnicity*

  •  Region (to account for culture and life style)

  •  Overweight/obesity*

  •  CVD risk group*

  •  SSB category*

  •  Amount of SSB consumption*

  •  Type of comparator*

  •  Caffeine-containing drinks*

  •  Outcome definition

  •  Study design

  •  Duration of study

  •  Single versus multi-centre studies

  •  Risk of bias assessments

  •  Covariate adjustment

  •  Funding

  •  Adjustment for total caloric intake*

We will follow previously published guidance for meta-regression[49]. Meta-regression will be based on the random effects model to allow for residual unexplained heterogeneity. A p value <0.10 will characterize statistical significance. When the sizes of the included studies are moderate or large, there should be at least 10 studies for a continuous study-level variable. For a categorical subgroup variable, each subgroup should have a minimum of four studies. These numbers serve as the lower bounds for considering meta-regression[49]. When included studies are mostly small in size, univariable meta-regression will be used when an insufficient number of studies are available to conduct multivariable analyses.

Regarding caloric intake, if highly correlated with SSB consumption, it will be difficult to decipher the independent effect of SSB consumption. If caloric intake lies in the causal pathway, then adjusting for it would eliminate any true associations between SSB consumption and outcomes. However, a counter-argument could be made that failure to adjust for caloric intake would produce spurious positive associations. Due to this uncertainty, we reasoned that assessing effect estimates for this variable in a subgroup analysis was the most appropriate; studies will not be penalized for adjusting or not adjusting in assessments of confounding bias.

Sensitivity analysis

Sensitivity analyses may be used to restrict analyses to low risk of bias and any decisions made regarding data handling.

Assessing for small study effects

We will investigate small study effects by the performance of cumulative meta-analysis (studies ordered and synthesized from the most to the least precise) and/or other graphical or statistical techniques if the following criteria are met: there are at least 10 studies available, studies are of unequal sizes, there are no substantial clinical and methodological differences between smaller and larger studies, and quantitative results are accompanied with measures of dispersion[5760].

Discussion

Grading the quality of evidence and interpretation

We will use the GRADE approach to evaluate the quality of evidence for outcomes[42]. Quality of evidence is the level of confidence for a causal inference that authors place in the estimate of effect for an outcome (i.e. their judgement that the evidence reflects the true effect). As stated previously, if surrogate measures are analysed in lieu of decision-making endpoints, these will be graded.

When grading the evidence, reviewers will evaluate the domains of study limitations (risk of bias), inconsistency, indirectness, imprecision, and publication bias and downgrade where important limitations exist. Studies may also be upgraded based on a strong magnitude of effect that is not due to known biases, dose-response gradient, and residual confounding that would have reduced the observed effect. The overall quality of evidence grade will be designated as high (confident the true effect lies close to that of the estimate), moderate (moderately confident in the effect estimate but may be substantially different), low (confidence in the effect estimate is limited), or very low (very little confidence in the effect estimate)[42]. The results will be discussed in light of the strength of findings as well as their implications for research and public health.

Quality assurance

We used the PRISMA for Protocols (PRISMA-P) checklist for reporting this protocol[61]. This review will be reported according to the PRISMA statement[37] and using a Measurement Tool to Assess the Methodological Quality of Systematic Reviews (AMSTAR) tool for additional quality control[62]. This protocol does not update any previously conducted systematic review. Any amendments made to this protocol when conducting the review will be outlined in the review’s manuscript.

Abbreviations

AMSTAR: 

A Measurement Tool to Assess the Methodological Quality of Systematic Reviews

CADTH: 

Canadian Agency for Drugs and Technologies in Health

CBA: 

controlled before-after

CCTs: 

controlled clinical trials

CDC: 

Centers for Disease Control and Prevention

CI: 

confidence interval

CVD: 

cardiovascular disease

CINAHL: 

Cumulative Index to Nursing and Allied Health Literature

DAGs: 

directed acyclic graphs

Distiller SR: 

Distiller Systematic Review Software

EAL: 

Evidence Analysis Library

EPOC: 

Effective Practice and Organisation of Care

ERIC: 

Education Resources Information Center

FASEB: 

Federation of American Societies for Experimental Biology

FNCE: 

Food and Nutrition Conference and Expo

GRADE: 

Grading of Recommendations Assessment, Development and Evaluation

IASO: 

International Association for the Study of Obesity

ITS: 

interrupted time series

NHANES: 

National Health and Nutrition Examination Survey

PRESS: 

peer review of electronic search strategies

PRISMA: 

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PRISMA-P: 

Preferred Reporting Items for Systematic Reviews and Meta-Analysis for Protocols

RCTs: 

randomized controlled trials

RR: 

risk ratio

SSB: 

sugar-sweetened beverage

WHO: 

World Health Organization.

Declarations

Acknowledgements

We thank Geoff Ball, Elizabeth Ghogomu, Gunnar Kraag, Sally Morton, Paul Poirier, Howard Sesso, Patricia Splett, Liz Stevens, Valerie Taylor, and Leanne Ward for their feedback on the protocol or aspects thereof; Becky Skidmore for devising searches and Jessie McGowan for peer-reviewing those searches; and Raymond Daniel for the assistance in preparing this manuscript. We have received funding from the Canadian Institutes of Health Research (FRN-132039) to conduct the systematic review. The funder was not involved in the design of the protocol or decision to submit the protocol for publication, nor will they be involved in any aspect of the conduct of the systematic review.

Authors’ Affiliations

(1)
Clinical Epidemiology Program, Centre for Practice-Changing Research, Ottawa Hospital Research Institute
(2)
University of Ottawa
(3)
E F Myers Consulting, Inc
(4)
Research Evidence Analysis, Academy of Nutrition and Dietetics
(5)
Canadian Obesity Network, Royal Alexandra Hospital, MMC
(6)
C.T. Lamont Primary Health Care Research Centre, Bruyere Research Institute, Department of Family Medicine, University of Ottawa
(7)
Planning, Research and Analysis Branch, Ministry of Health and Long-Term Care
(8)
Julia McFarlane Diabetes Research Centre, University of Calgary
(9)
School of Journalism and Communication Journalism, Carleton University
(10)
Division of Prevention and Rehabilitation, University of Ottawa Heart Institute, University of Ottawa
(11)
Healthy Living Section, Chronic Disease and Injury Prevention, Ottawa Public Health
(12)
Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, Jewish General Hospital, McGill University
(13)
Department of Agricultural, Food and Nutritional Science, University of Alberta
(14)
Healthy Active Living and Obesity Research Group, Children’s Hospital of Eastern Ontario Research Institute
(15)
Department of Epidemiology and Community Medicine, University of Ottawa

References

  1. Pereira MA: The possible role of sugar-sweetened beverages in obesity etiology: a review of the evidence. Int J Obes (Lond). 2006, 30 (Suppl 3): S28-S36.View ArticleGoogle Scholar
  2. Basciano H, Federico L, Adeli K: Fructose, insulin resistance, and metabolic dyslipidemia. Nutr Metab (Lond). 2005, 2 (1): 5-View ArticleGoogle Scholar
  3. Elliott SS, Keim NL, Stern JS, Teff K, Havel PJ: Fructose, weight gain, and the insulin resistance syndrome. Am J Clin Nutr. 2002, 76 (5): 911-922.PubMedGoogle Scholar
  4. Hung CT: Effects of high-fructose (90%) corn syrup on plasma glucose, insulin, and C-peptide in non-insulin-dependent diabetes mellitus and normal subjects. Taiwan Yi Xue Hui Za Zhi. 1989, 88 (9): 883-885.PubMedGoogle Scholar
  5. Klurfeld DM, Foreyt J, Angelopoulos TJ, Rippe JM: Lack of evidence for high fructose corn syrup as the cause of the obesity epidemic. Int J Obes (Lond). 2013, 37 (6): 771-773.View ArticleGoogle Scholar
  6. Melanson KJ, Zukley L, Lowndes J, Nguyen V, Angelopoulos TJ, Rippe JM: Effects of high-fructose corn syrup and sucrose consumption on circulating glucose, insulin, leptin, and ghrelin and on appetite in normal-weight women. Nutrition. 2007, 23 (2): 103-112.View ArticlePubMedGoogle Scholar
  7. Angelopoulos TJ, Lowndes J, Zukley L, Melanson KJ, Nguyen V, Huffman A, Rippe JM: The effect of high-fructose corn syrup consumption on triglycerides and uric acid. J Nutr. 2009, 139 (6): 1242S-1245S.View ArticlePubMedGoogle Scholar
  8. Rippe JM, Kris Etherton PM: Fructose, sucrose, and high fructose corn syrup: modern scientific findings and health implications. Adv Nutr. 2012, 3 (5): 739-740.View ArticlePubMedPubMed CentralGoogle Scholar
  9. Grynbaum M: New York plans to ban sale of big sizes of sugary drinks. New York Times. New York ed. 2012; May 30. pA1. [http://www.nytimes.com/2012/05/31/nyregion/bloomberg-plans-a-ban-on-large-sugared-drinks.html?pagewanted=all&_r=0],
  10. World Health Organization: WHO opens public consultation on draft sugars guideline. [http://www.who.int/mediacentre/news/notes/2014/consultation-sugar-guideline/en/],
  11. Han E, Powell LM: Consumption patterns of sugar-sweetened beverages in the United States. J Acad Nutr Diet. 2013, 113 (1): 43-53.View ArticlePubMedPubMed CentralGoogle Scholar
  12. Malik VS, Popkin BM, Bray GA, Despres JP, Hu FB: Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation. 2010, 121 (11): 1356-1364.View ArticlePubMedPubMed CentralGoogle Scholar
  13. Malik VS, Popkin BM, Bray GA, Despres JP, Willett WC, Hu FB: Sugar-sweetened beverages and risk of metabolic syndrome and type 2 diabetes: a meta-analysis. Diabetes Care. 2010, 33 (11): 2477-2483.View ArticlePubMedPubMed CentralGoogle Scholar
  14. Hu FB, Malik VS: Sugar-sweetened beverages and risk of obesity and type 2 diabetes: epidemiologic evidence. Physiol Behav. 2010, 100 (1): 47-54.View ArticlePubMedPubMed CentralGoogle Scholar
  15. Vartanian LR, Schwartz MB, Brownell KD: Effects of soft drink consumption on nutrition and health: a systematic review and meta-analysis. Am J Public Health. 2007, 97 (4): 667-675.View ArticlePubMedPubMed CentralGoogle Scholar
  16. Gibson S: Sugar-sweetened soft drinks and obesity: a systematic review of the evidence from observational studies and interventions. Nutr Res Rev. 2008, 21 (2): 134-147.View ArticlePubMedGoogle Scholar
  17. Malik VS, Schulze MB, Hu FB: Intake of sugar-sweetened beverages and weight gain: a systematic review. Am J Clin Nutr. 2006, 84 (2): 274-288.PubMedPubMed CentralGoogle Scholar
  18. World Health Organization: Joint WHO/FAO Expert Consultation on Diet, Nutrition and the Prevention of Chronic Diseases. WHO Technical Report Series No. 916. 2002, Geneva: World Health OrganizationGoogle Scholar
  19. United States Department of Agriculture: USDA’s Nutrition Evidence Library. [http://www.nutritionevidencelibrary.gov/],
  20. Forshee RA, Anderson PA, Storey ML: Sugar-sweetened beverages and body mass index in children and adolescents: a meta-analysis. Am J Clin Nutr. 2008, 87 (6): 1662-1671.PubMedGoogle Scholar
  21. Mattes RD, Shikany JM, Kaiser KA, Allison DB: Nutritively sweetened beverage consumption and body weight: a systematic review and meta-analysis of randomized experiments. Obes Rev. 2011, 12 (5): 346-365.View ArticlePubMedPubMed CentralGoogle Scholar
  22. 5.3 Beverages. Int J Obes (Lond). 2009, 33: S28-S34.Google Scholar
  23. Anderson GH: Sugars-containing beverages and post-prandial satiety and food intake. Int J Obes (Lond). 2006, 30: S52-S59.View ArticleGoogle Scholar
  24. Malik VS, Hu FB: Sugar-sweetened beverages and health: where does the evidence stand?. Am J Clin Nutr. 2011, 94 (5): 1161-1162.View ArticlePubMedPubMed CentralGoogle Scholar
  25. Weed DL, Althuis MD, Mink PJ: Quality of reviews on sugar-sweetened beverages and health outcomes: a systematic review. Am J Clin Nutr. 2011, 94 (5): 1340-1347.View ArticlePubMedPubMed CentralGoogle Scholar
  26. Palmer JR, Boggs DA, Krishnan S, Hu FB, Singer M, Rosenberg L: Sugar-sweetened beverages and incidence of type 2 diabetes mellitus in African American women. Arch Intern Med. 2008, 168 (14): 1487-1492.View ArticlePubMedPubMed CentralGoogle Scholar
  27. Ruxton CH, Gardner EJ, McNulty HM: Is sugar consumption detrimental to health? A review of the evidence 1995–2006. Crit Rev Food Sci Nutr. 2010, 50 (1): 1-19.View ArticlePubMedGoogle Scholar
  28. 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. 2011, 93 (6): 1321-1327.View ArticlePubMedPubMed CentralGoogle Scholar
  29. Cohen L, Curhan G, Forman J: Association of sweetened beverage intake with incident hypertension. J Gen Intern Med. 2012, 27 (9): 1127-1134.View ArticlePubMedPubMed CentralGoogle Scholar
  30. Eshak ES, Iso H, Kokubo Y, Saito I, Yamagishi K, Inoue M, Tsugane S: Soft drink intake in relation to incident ischemic heart disease, stroke, and stroke subtypes in Japanese men and women: the Japan Public Health Centre-based study cohort I. Am J Clin Nutr. 2012, 96 (6): 1390-1397.View ArticlePubMedGoogle Scholar
  31. Hernan MA, Hernandez-Diaz S, Robins JM: A structural approach to selection bias. Epidemiology. 2004, 15 (5): 615-625.View ArticlePubMedGoogle Scholar
  32. Cochrane Effective Practice and Organisation of Care Group: What study designs should be included in an EPOC review and what should they be called?. [http://epoc.cochrane.org/sites/epoc.cochrane.org/files/uploads/05%20What%20study%20designs%20should%20be%20included%20in%20an%20EPOC%20review%202013%2008%2012.pdf],
  33. Centers for Disease Control and Prevention: The CDC guide to strategies for reducing the consumption of sugar-sweetened beverages. [http://www.cdph.ca.gov/SiteCollectionDocuments/StratstoReduce_Sugar_Sweetened_Bevs.pdf],
  34. Sampson M, McGowan J, Cogo E, Grimshaw J, Moher D, Lefebvre C: An evidence-based practice guideline for the peer review of electronic search strategies. J Clin Epidemiol. 2009, 62 (9): 944-952.View ArticlePubMedGoogle Scholar
  35. Reuters T: Reference Manager 12 [Computer Program]. 2011, New York: Thomson ReutersGoogle Scholar
  36. Partners E: DistillerSR [Computer Program]. 2011, Ottawa: Evidence PartnersGoogle Scholar
  37. Moher D, Liberati A, Tetzlaff J, Altman DG: Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009, 151 (4): 264-269. W64.View ArticlePubMedGoogle Scholar
  38. Higgins J, Altman DG: Chapter 8: assessing risk of bias in included studies. [http://handbook.cochrane.org/chapter_8/8_assessing_risk_of_bias_in_included_studies.htm],
  39. Higgins J, Altman DG: Chapter 16: special topics in statistics. [http://handbook.cochrane.org/],
  40. Academy of Nutrition and Dietetics: Appendix 8: quality criteria checklist: primary research. [http://www.andeal.org/evidence-analysis-manual],
  41. Cochrane Effective Practice and Organisation of Care Group: Suggested risk of bias criteria for EPOC reviews. [http://epoc.cochrane.org/sites/epoc.cochrane.org/files/uploads/Suggested%20risk%20of%20bias%20criteria%20for%20EPOC%20reviews.pdf],
  42. Balshem H, Helfand M, Schunemann HJ, Oxman AD, Kunz R, Brozek J, Vist GE, Falck-Ytter Y, Meerpohl J, Norris S, Guyatt GH: GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011, 64 (4): 401-406.View ArticlePubMedGoogle Scholar
  43. Shrier I, Platt RW: Reducing bias through directed acyclic graphs. BMC Med Res Methodol. 2008, 8: 70-View ArticlePubMedPubMed CentralGoogle Scholar
  44. Pearl J: The art and science of cause and effect. Causality: Models, Reasoning and Inference. 2000, Cambridge: Cambridge University, 331-358.Google Scholar
  45. Textor J, Hardt J, Knuppel S: DAGitty: a graphical tool for analyzing causal diagrams. Epidemiology. 2011, 22 (5): 745-View ArticlePubMedGoogle Scholar
  46. Wiebe N, Vandermeer B, Platt RW, Klassen TP, Moher D, Barrowman NJ: A systematic review identifies a lack of standardization in methods for handling missing variance data. J Clin Epidemiol. 2006, 59 (4): 342-353.View ArticlePubMedGoogle Scholar
  47. Rosenthal R, Rosnow R, Rubin D: Contrasts and Effect Sizes in Behavioral Research: a Correlational Approach. 2000, Cambridge: Cambridge University PressGoogle Scholar
  48. Chinn S: A simple method for converting an odds ratio to effect size for use in meta-analysis. Stat Med. 2000, 19 (22): 3127-3131.View ArticlePubMedGoogle Scholar
  49. Fu R, Gartlehner G, Grant M, Shamliyan T, Sedrakyan A, Wilt TJ, Griffith L, Oremus M, Raina P, Ismaila A, Santaguida P, Lau J, Trikalinos TA: Conducting quantitative synthesis when comparing medical interventions: AHRQ and the Effective Health Care Program. J Clin Epidemiol. 2011, 64 (11): 1187-1197.View ArticlePubMedGoogle Scholar
  50. Borenstein M, Hedges L, Higgins J, Rothstein H: Comprehensive meta-analysis version 2. [http://www.meta-analysis.com/pages/downloads/Meta-Analysis-Manual.pdf],
  51. SAS. [http://www.sas.com],
  52. Ramsay CR, Matowe L, Grilli R, Grimshaw JM, Thomas RE: Interrupted time series designs in health technology assessment: lessons from two systematic reviews of behavior change strategies. Int J Technol Assess Health Care. 2003, 19 (4): 613-623.View ArticlePubMedGoogle Scholar
  53. DerSimonian R, Laird N: Meta-analysis in clinical trials. Control Clin Trials. 1986, 7 (3): 177-188.View ArticlePubMedGoogle Scholar
  54. Arends LR, Hamza TH, van Houwelingen JC, Heijenbrok-Kal MH, Hunink MG, Stijnen T: Bivariate random effects meta-analysis of ROC curves. Med Decis Making. 2008, 28 (5): 621-638.View ArticlePubMedGoogle Scholar
  55. Deeks JJ, Higgins J, Altman DG: Analysing data and undertaking meta-analyses. Cochrane Handbook for Systematic Reviews of Interventions: Cochrane Book Series. Edited by: Higgins JPT, Green S. 2008, Chichester: Wiley, 243-296.View ArticleGoogle Scholar
  56. Higgins JP, Thompson SG, Deeks JJ, Altman DG: Measuring inconsistency in meta-analyses. BMJ. 2003, 327 (7414): 557-560.View ArticlePubMedPubMed CentralGoogle Scholar
  57. Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, Carpenter J, Rucker G, Harbord RM, Schmid CH, Tetzlaff J, Deeks JJ, Peters J, Macaskill P, Schwarzer G, Duval S, Altman DG, Moher D, Higgins JP: Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ. 2011, 343: d4002-View ArticlePubMedGoogle Scholar
  58. Egger M, Davey SG, Schneider M, Minder C: Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997, 315 (7109): 629-634.View ArticlePubMedPubMed CentralGoogle Scholar
  59. Rucker G, Schwarzer G, Carpenter J: Arcsine test for publication bias in meta-analyses with binary outcomes. Stat Med. 2008, 27 (5): 746-763.View ArticlePubMedGoogle Scholar
  60. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR: Introduction to Meta-analysis. 2009, New York: WileyView ArticleGoogle Scholar
  61. Moher D, Shamseer L, Clarke M: Reporting guidelines for systematic review protocols. 19th Cochrane Colloquium: 19–22. 2011, October ; MadridGoogle Scholar
  62. Shea BJ, Grimshaw JM, Wells GA, Boers M, Andersson N, Hamel C, Porter AC, Tugwell P, Moher D, Bouter LM: Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol. 2007, 7: 10-View ArticlePubMedPubMed CentralGoogle Scholar
  63. Chow AW: Epidemiology, pathogenesis and clinical manifestations of odontogenic infections. [http://www.uptodate.com/contents/epidemiology-pathogenesis-and-clinical-manifestations-of-odontogenic-infections],
  64. Bray GA, Nielsen SJ, Popkin BM: Consumption of high-fructose corn syrup in beverages may play a role in the epidemic of obesity. Am J Clin Nutr. 2004, 79 (4): 537-543.PubMedGoogle Scholar
  65. Johnson RJ, Sanchez-Lozada LG, Nakagawa T: The effect of fructose on renal biology and disease. J Am Soc Nephrol. 2010, 21 (12): 2036-2039.View ArticlePubMedGoogle Scholar
  66. Gagliardi AC, Miname MH, Santos RD: Uric acid: a marker of increased cardiovascular risk. Atherosclerosis. 2009, 202 (1): 11-17.View ArticlePubMedGoogle Scholar
  67. van Dam RM: Coffee consumption and risk of type 2 diabetes, cardiovascular diseases, and cancer. Appl Physiol Nutr Metab. 2008, 33 (6): 1269-1283.View ArticlePubMedGoogle Scholar
  68. Butt MS, Sultan MT: Coffee and its consumption: benefits and risks. Crit Rev Food Sci Nutr. 2011, 51 (4): 363-373.View ArticlePubMedGoogle Scholar
  69. Center for Science in the Public Interest: Petition to bar the use of caramel colorings produced with ammonia and containing the carcinogens 2-methylimidazole and 4-methylimidazole. [http://cspinet.org/new/pdf/caramel_coloring_petition.pdf],
  70. Tremblay MS, Colley RC, Saunders TJ, Healy GN, Owen N: Physiological and health implications of a sedentary lifestyle. Appl Physiol Nutr Metab. 2010, 35 (6): 725-740.View ArticlePubMedGoogle Scholar
  71. Ware JE: SF-36® Health Survey Update. [http://www.sf-36.org/tools/SF36.shtml],
  72. World Health Organization: The World Health Organization Quality of Life Assessment (WHOQOL): development and general psychometric properties. Soc Sci Med. 1998, 46 (12): 1569-1585.View ArticleGoogle Scholar

Copyright

© Hamel et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Advertisement