Introduction

In response to the World Health Assembly Resolution on Oral Health (WHA74.5) a Draft Global Strategy on Oral Health has been developed [1]. This includes the Strategic Objective ‘to enable all people to achieve the best possible oral health and target and reduce the social and commercial determinants and risk factors of oral diseases and conditions.’ The Global Strategy on Oral health, aims to recommend cost-effective oral health interventions by 2024. This will form part of the updated Appendix 3 of the WHO Global action plan on the prevention and control of non-communicable diseases (NCDs). WHO Member States will be guided in developing national responses to promote oral health and reduce oral health inequalities and diseases including dental caries; globally.

Dental caries is the most prevalent oral disease [2]. The role of dietary sugars in its aetiology is well established [3, 4]. A WHO recommendation, for minimising lifelong risk of dental caries, is to limit free sugars intake to below 5% of energy intake. This puts forth an important strategy for caries prevention [5]. Addressing the universally high free sugars intake, is an important part of the Global Strategy on Oral Health [1], yet public health measures to reduce sugars consumption are rare [6, 7]. To make a tangible difference, oral health policy and action plans need to move away from approaches that rely entirely on individual dietary behaviour change for free sugars reduction. Creating social and economic structures to support people to make behaviour changes will require taking bold action to focusing on upstream interventions to limit free sugars intake [8,9,10]. One such measure is the implementation of taxation on sugar-sweetened beverages (SSB).

SSB are a common source of dietary free sugars, with an average global consumption by children of 326 ml/day (ranging from 115 ml/day in Australia to 710 ml/day in China) [11]. In European countries SSB contribution to free sugars varies between 0–65% [12] and in the US SSB contribute to added sugars intake by 16–23% [13]. These data suggest that lowering SSB intake could significantly impact on free sugars intake. The WHO recognises, and provides guidance on, the implementation of taxes on SSB as an evidence-based policy to prevent obesity and non-communicable diet-related diseases [6]. SSB taxes have been introduced in several countries around the world to incentivise healthy beverage choice [14]. A wealth of evidence suggests that this has a positive impact on obesity prevention [15], though little attention has focused on the impact on the global impact on dental caries.

Within-country-based modelling studies aiming to predict the impact of taxation of SSB on subsequent levels of dental caries are largely from high income countries (HIC) [16,17,18,19]. However, a diversity of approaches and assumptions has led to vast between-study differences in effect sizes and irreproducible results. The impacts of SSB taxation on development of dental caries in low- and middle-income countries (LMIC) have not been broadly reported [20]. The challenges in publishing in mainstream journals of HIC faced by authors from LMIC, may explain this observation. With a view to assisting WHO in the identification of cost-effective interventions, an independent assessment of the evidence pertaining to the predicted impact of SSB taxation on dental caries in both HIC and LMIC on prevention of dental caries was conducted. For this purpose, an Umbrella Review approach was chosen. An umbrella review is a review of systematic reviews [21] that captures the vast amount of evidence contained in systematic reviews and studies to access research evidence and inform decision-making. They provide a summary of existing research syntheses related to a given topic or question and are also applied when there is a need for “fast” evidence in reduced timeframes. Umbrella reviews can be used to summarize more than one research synthesis e.g., for different populations or geographic locations. Using an umbrella review approach, the aims of this study were first, to identify the available data pertaining to the impact of SSB taxation on consumption of SSB (i.e., the price elasticity of demand (PED) and sugars, and of the impact of the reduced sugars intake on the development of dental caries; and second, to utilise these data as model parameters to assess the impact of SSB taxation on caries prevention in both LMIC and HIC. The objectives were first, to conduct an umbrella review to identify the best available evidence pertaining to the PED of SSB and the impact of taxation on sugars consumption, and of sugars consumption on the development of dental caries; and second, to use these data to estimate the potential impact of introducing a 20% SSB volumetric tax on averting dental caries.

Methods

A preliminary search for previous umbrella reviews on the topic was conducted in PubMed and Web of Science and no existing reviews were identified. The protocol for this umbrella review was registered on PROSPERO in January 2022 (CRD 42022293187) [22]. The review was guided by the Methodology for Joanna Briggs Institute (JBI) Umbrella Reviews [21]. The following questions were addressed:

  1. 1.

    What are the effects of SSB taxation on SSB on PED/ consumption?

  2. 2.

    What is the effect of SSB taxation on consumption of sugars?

  3. 3.

    What are the effects of decreasing sugars consumption on levels of dental caries?

  4. 4.

    What is the likely effect of a 20% volumetric tax on averting dental caries over a 10-year period?

Population

Studies of healthy populations (i.e., reviews that do not specifically target participants with disease), all age groups, race, gender and geographic locations were included. Data pertaining to the above questions was explored by age group (adults, children (for dental caries outcomes children were further classified according to whether outcomes related to the primary dentition, the permanent dentition or both), geographic location (country), income classification of included countries (high, middle, and low), socioeconomic status of participants and type of tax. To facilitate best-possible context-specificity, where possible, the evidence was mapped according to the WHO Regions.

Intervention/exposure and outcomes

For Questions 1 and 2, systematic reviews of studies that measured the impact of any type of tax to SSB including ad valorem, volumetric tax, or nutrient based tax (based on the sugars content of the drink) were included. Different levels of taxes, including excise tax and sales tax, applied by government, manufacturer, or retailer, were considered.

For Question 1, the outcome was measure of SSB consumption measured as sales, household consumption level, and data on consumption from surveys of dietary intake. SSB included beverages with added sugars, e.g., carbonated and still beverages and sweetened fruit juices, excluding beverages sweetened exclusively with artificially sweeteners. Reports on change in consumption expressed as amount (e.g., mL/day, week, or year, or change in energy intake (EI) from SSB) or units of frequency from which amount can be derived, were included. Data on change in consumption of SSB derived from data on PED including own price elasticity (OPE) and cross price elasticity (CPE) of SSB were included.

For Question 2, the outcome measure was a quantitative change in the intake of sugars (free sugars) expressed as grams, kilograms or ounces per day, week, or year, or the amount of sugars expressed as a percent of EI. Outcomes also included data on the change in energy intake when expressed as Kcal or kJ, MJ per day, month or year, which allowed for conversion to a quantified amount of sugars using the Atwater Factor of 4 kcal/gram sugars [23].

For Question 3, systematic reviews were included if they reported data pertaining to a quantified measure of sugars intake or change in sugars intake. Sugars intake included total sugars (and any component of) i.e., free sugars, added sugars, sucrose, non-milk extrinsic sugars, expressed as g or kg/day or /year or as a percentage of EI, or a per capita population intake or availability. Systematic reviews that reported solely on the frequency of sugars consumption were excluded.

For Question 3, the outcome was a measure of dental caries. This included decayed, missing and filled teeth (DMFT (for permanent teeth), dmft (for primary teeth)), decayed, missing and filled surfaces (DMFS (for permanent teeth), dmfs (for primary teeth)), decayed, extracted due to caries, filled teeth (deft), or comparisons between caries and no caries or higher caries vs. lower caries.

Sources of information

Online biomedical databases, including PubMed, Embase, Web of Science, Scopus, CINAHL, and Dentistry and Oral Sciences Source were searched. Registries for systematic reviews, including Cochrane Library, the JBI Systematic Review Register and the PROSPERO Library, were searched. Databases were selected from available sources through The University of Adelaide library, aiming for broad geographical coverage. There were no restrictions placed on language and study duration. The date limit was from 2000 to end 2021 (to broadly cover the period since the introduction of taxes on SSBs [24]. The search strategy is presented in Additional file 1. In addition to the search, known experts were contacted by email to identify any further systematic reviews. The inclusion/exclusion criteria and outcome measures for each review question are described below.

Types of study included

The umbrella review included exclusively systematic reviews with or without evidence synthesis, e.g., meta-analysis. For Questions 1 and 2, systematic reviews which examined the effects of SSB taxes on SSB consumption, included both naturalistic (explore the impact of real-world taxes) and modelling studies (hypothetical tax) that used cross sectional or longitudinal (before after) or time series data on price and consumption, or data from experimental intervention studies. Systematic reviews of studies with exclusively qualitative data were excluded.

For Question 3, systematic reviews which exclusively examined the effects of amount of sugars intake or changing sugars intake on dental caries levels of change in dental caries development were included. Reviews that incorporate text and opinion as their primary source of evidence were excluded.

Study selection

Articles identified by searches were initially screened independently in duplicate by two authors to exclude systematic reviews clearly outside of scope. Potentially eligible reviews underwent full text review by independent duplicate assessment for inclusion. Differences between reviewers’ results were resolved by discussion. Reviews that were excluded at full-text screening are presented in Additional file 2. One reviewer extracted relevant data using a modified version of the JBI data extraction form [25] and a second reviewer checked data. Completed data extraction forms for included reviews are presented in Additional file 3 (Questions 1 and 2) and Additional file 4 (Question 3). Where pooled analysis was not available to answer a review question, data from original studies within systematic reviews was extracted and summarised.

Quality of included data

The quality of eligible systematic reviews and evidence syntheses was assessed in duplicate using AMSTAR (A Measurement Tool to Assess Systematic Reviews) which is based on an 11-point scale. An AMSTAR score of 8–11 is categorised as high quality, 4–7 is moderate quality and 0–3 low quality [26]. Any disagreement between reviewers was resolved by involvement of a third reviewer. Assessments were conducted independently in duplicate by two reviewers without conflicting interest (e.g., not an author of the review) and any discrepancy between reviewers were resolved by discussion.

Data summary

The approach to the presentation of findings from included systematic reviews included a tabulated summary of identified reviews that was presented in reverse chronological order and included information on AMSTAR rating. This was supported by a narrative summary that used a ‘best available evidence’ approach where data from the most recent high quality systematic review was described first, followed by a comparison with data from moderate quality reviews or earlier published reviews. Evidence synthesis from the most recent best quality systematic review was used to inform subsequent outputs. To enable a more detailed assessment and interpretation of the evidence when meta-analysis from SRs were not available, further data extraction of the characteristics and findings of primary studies included in the SRs was carried out. In this instance, data extracted included: author year, country of data collection, sample size, age, objectives, intervention or exposure, outcome, and quality assessment.

Estimated impact on levels of dental caries

The impact of SSB taxation on intake of free sugars was estimated in two ways. First, from the identified estimates of the impact of SSB taxation on the consumption of SSB (PED data from Question 1) together with available information on the level of SSB consumption in HIC and LMIC [11]. Second, using the identified estimated of the impact of SSB taxation on intake of energy and sugars (Question 2). In the absence of data from meta-analysis pertaining to sugars intake and dental caries, estimates of sugars reduction and the best available data were used. This included identified data on the dose response between sugars and development of dental caries, (identified in Question 3) to estimate the impact of a 20% volumetric SSB tax on the number of caries prevented in both children and adults over a period of 10 years (Question 4). The identified data for HIC and for LMIC were applied separately.

Results

Figure 1 presents the PRISMA flow diagram. For Questions 1 & 2, from all databases combined, 419 systematic reviews were identified following de-duplication. Following title and abstract screening, 48 systematic reviews were retained for full text screening. Following full text screening, 14 systematic reviews were included. Four were rated as high quality [15, 27,28,29] nine as moderate quality [30,31,32,33,34,35,36,37,38] and one as low quality [39]. Two reviews did not make a declaration of conflict of interest [38, 39]. For question 3, from all databases combined, 100 systematic reviews were identified following de-duplication. Following title and abstract screening 21 reviews were retained for full text screening from which two systematic reviews were included; a further three systematic reviews were identified through known experts in the field giving five systematic reviews. Two reviews were rated as high quality [3, 4] and three as moderate quality [12, 40, 41]. One review did not make a declaration of conflicts of interests [40]. Reasons for exclusions of full texts are provided in Fig. 1 and Additional file 2.

Fig. 1
figure 1

PRISMA flowchart of study selection and inclusion process

Full size image

Question 1: Impact of taxation on SSB consumption and PED

Of the 14 systematic reviews identified that had data relevant to Question 1, eight included a narrative account only and six conducted evidence synthesis by pooling data or meta-analysis. Systematic reviews included data from the African, European, Southeast Asian, and Western Pacific WHO Regions as well as The Region of The Americas. A summary of included reviews along with the AMSTAR rating is provided in Table 1. Details of the AMSTAR rating for each included review is provided in Additional file 5. A summary of the original studies included in the 14 reviews, including the quality appraisal for these studies, is provided in Additional file 6.

Table 1 Summary of included systematic reviews with data pertaining to Question 1a
Full size table

Data from high-income countries

Of the systematic reviews with quantitative evidence synthesis, Teng et al. [27], which included data from HIC, and had a high AMSTAR rating, showed that a 10% tax on SSB reduced consumption by 10.0% (95% CI: -5.0, 14.7%) (PED 1.0 (95% CI: -0.5, -1.47)). The analysis by Afshin et al. [28], also with a high AMSTAR rating, was based on pooled data from studies in the US only and showed a 10% tax would reduce consumption by 7.0% (95% CI: -3.0,-10.0%). The review of Powell et al. [37] had a moderate AMSTAR rating and was based on US-based price elasticity studies and showed the overall OPE was -1.21 (95% CI: -0.7, -2.26) implying that a 10% tax would reduce consumption by 7.1–22.6% (average reduction 12.1%), and a 20% tax by 14.2–45.2% (average reduction 24.2%). The review of Andreyeva et al. [38], with a moderate AMSTAR rating and based on US data only, showed, based on mean price elasticity estimates (95% CI: -0.8, 1.0) that a 10% tax would reduce SSB consumption by 8–10%.

Data from low- and middle-income countries

The best available evidence synthesis for LMIC was provided by the review of Nakhimovsky et al. [29] which had a Moderate AMSTAR rating. This synthesis standardised data across studies and showed that a 10% tax would reduce consumption by 6–12% (average reduction 9.0%, PED 0.90 (range: -0.6 to -1.2)). The review by Cabrera Escobar et al. [35], which had a moderate AMSTAR rating, included both HIC and LMIC and showed in a meta-analysis of data from the USA, Mexico, Brazil, and France that overall OPE was -1.3 (95% CI: -1.085, -1.509), thus indicating that a 10% tax would reduce consumption by 10.9, 15.1% (average reduction 13.0%). A summary of the data on percent change in SSB consumption is provided in Table 2.

Table 2 Summary of quantitative findings of the impact of SSB taxation on percentage change in consumption
Full size table

Data by age group and SES

Only the review by Teng et al. [27] presented pooled data by age group, finding no significant difference. In an earlier review, Thow et al. [34] reported that the impact of taxes ranging from 5 to 30% on SSB consumption was proportional to tax applied. However, the impact on EI was higher in adults (range: 10.0 to 48.0%) compared with children (range: 5.0 to 8.0%) in due of considerable substitution (e.g. with milk). The systematic reviews of Maniadakis et al. [36], Thow et al. [34] (HIC), and of Nakhimovsky et al. [29] (LMIC) showed SSB taxes to be more regressive in lower income groups. However, difference by SES was not found in the analysis of HIC by Teng et al. [27], which included some studies where a 5% tax led to a greater reduction in consumption in higher SES groups. Further data on the impact of SSB tax on consumption by age and SES can be found in the summary of original studies in Additional file 6.

Only one systematic review provided data on percentage reduction of SSB consumption by type of tax, which varied but was not statistically significant [27]. Average reductions with a 10% tax level were 2.3% (95% CI: -11.2, 7.4) for an ad valorem tax, 10.2% (95% CI: -4.1, -15.9%) for volumetric; and 14.0% (95% CI: -7.5%, -20.1%) for a nutrient-based tax based on a sugars concentration threshold.

Question 2: Systematic reviews that enabled estimation of the impact of SSB tax on sugars consumption

Two included systematic reviews reported data on the impact of SSB tax on EI, which enabled the estimation of the impact of SSB tax on sugars intake [29, 36]. Nakhimovsky et al. [29], by using data from LMIC reported that a 10% tax would reduce EI by a median of 18.0 (range: -5.0, -39.0) KJ/person/day or by 4.3 ( range: -1.2 to -9.3) Kcal/person/day. Based on the Atwater conversion factor (4.0 kcal/ (17.0 kJ)/g sugar) [23], this reduction is equivalent to 1.1 (range: 0.3 to 2.3) g sugars/person/day. Maniadakis, et al. [36] based on data from HIC, estimated that a 10% increase in price/tax would reduce EI by a maximum of 50 kcal/person/day or, 450 per month. This reduction is equivalent to 12.5 g/d and 112.5 g/month decrease in sugars intake. Of the original studies included in the identified systematic reviews, 16 had data that enabled the impact of SSB taxation on intake of sugars to be determined (summarised in Additional file 7). Nine of these studies provided data that enabled the impact of a 20% tax on SSB on free sugars intake to be estimated, showing decreases ranging from 1.8 g to 11.0 g grams sugars/person/day, with the average decrease in LMIC being 4 g/d and in HIC 4.4 g/d.

Question 3. The effects of decreasing sugars consumption on levels of dental caries

The five included systematic reviews are summarised in Table 3. One review [4] was an update of an earlier review [3]. None of the included systematic reviews provided a meta-analysis pertaining to the impact of amount of sugars intake on the development of dental caries. However, data from included original studies relating to the dose response relationship are summarised in Table 4. The best available data, based on study design came from cohort studies with low risk of bias (RoB). For adults, this was provided by the analysis of Bernabé et al. [104], which showed that for each 10 g intake of total sugars DMFT increased 0.09 over the 11-year follow up period. For children, the best available data for the primary dentition came from the analysis of Turck et al. [12] (Moderate RoB), which showed an increase in dmft of 1.64 between ages 3 and 6 for each 10 g/day sucrose consumption. The best available data for the permanent dentition of children came from the Michigan Study (low RoB) [105, 106] which showed a 1% increase for each 5 g of sugar in children over a three-year follow-up period.

Table 3 Summary of systematic reviews with data pertaining to Question 3a
Full size table
Table 4 Identified studiesa with data enabling estimation of the dose response between sugars and dental caries
Full size table

Question 4. The likely effect of a 20% volumetric tax on averting dental caries over a 10-year period

Table 5 tabulates the best available data pertaining to the impact of taxation on SSB consumption, sugars intake, and the dose response between amount of sugars intake and development of caries. Estimates for the impact of a 20% volumetric SSB tax on caries development for both HIC and LMIC are presented.

Table 5 Summary of the best available evidence and estimated impact of a 20% volumetric tax on the development of dental caries over a ten-year period
Full size table

Discussion

Through an umbrella review of the best available evidence, the findings of this study suggest that over a ten-year period, a 20% volumetric tax to SSB would reduce the per capita caries count (number of teeth affected by caries) in adults in both HIC and LMIC by approximately 0.03. In children the per capita caries count (number of tooth surfaces affected by caries) would reduce by 0.16 and 0.18, and the caries occurrence by 2.7% and 2.9% in LMIC and HIC respectively.

It has been recognised that no single action will be effective in reducing sugars intake to recommended threshold levels and that this is unlikely to be achieved by interventions that rely on individuals changing behaviour alone [118]. This umbrella review has indicated that SSB taxation alone would have a modest impact on disease levels. Moreover, these reductions in dental disease would likely have notable cost benefits due to the high direct costs incurred in treating dental caries and the indirect costs associated with the disease [119].

Findings in context of previous findings

Most modelling studies of the effect of SSB taxation on caries have reported on the impact on consumption of SSB but not reported the impact on quantitative measures of dietary sugars consumption per se [18,19,20]. However, Schwendicke et al. [17] using consumption data from the German National Nutrition Survey indicated a 20% tax had variable impacts across gender and SES groups. The most affect sugars intake being in lower income males (up to an average of 13.7 g/day in males aged 20–29 and 5.7 g/day in females aged 15–19 years). Three [29, 34, 36] of the four [27, 29, 34, 36] systematic reviews identified in this umbrella review that included analyses by SES also showed taxes to be more regressive in lower SES groups.

It has been suggested that taxes are most effective when price change is passed to the consumer [8]. However, few data were identified on the impact of different types of taxation. Only one high quality systematic review showed that nutrient based taxes were most effective, an observation also noted by the WHO [120]. Nutrient based taxes are not, however, always passed to the consumer; SSB taxation in the UK [121], which was a tiered system, based on different sugars thresholds, drove product reformulation to lower the sugars content of drinks available for purchase. The UK SSB has thus resulted in a dual benefit of lowering sugars of products available and deterring consumers from buying drinks with higher sugars content.

Limitations

Despite the advantage of umbrella reviews in being a method of review to capture large amounts of evidence in a short time frame, there are several limitations to be acknowledge. First, despite a broad search strategy, potentially relevant studies may have been omitted for example the databases selected did not include all databases available in LMIC e.g. African Index Medicus (AIMS), Index, Medicus for the Eastern Mediterranean Region (IMEMR). Moreover, the process of screening, data extraction and analysis take time meaning that any relevant systematic reviews published after the search cut-off date and before publication of the umbrella review will be omitted. In this review, screening, extracting and quality appraisal of systematic reviews were carried out independently in duplicate. However, inadvertent systematic error during selection, appraisal or extraction cannot be ruled out. Umbrella reviews, by nature, are subject to limited coverage of evidence because if a study has not been included in a published systematic review, an umbrella review will not include it. In extracting data from this umbrella review, the estimates for caries reduction for LMIC may be conservative because the identified data on the dose response between sugars intake and development of dental caries came exclusively HIC. Sugars exposure may have a greater effect on caries in populations with less exposure to fluoride [122] and increased likelihood of undernutrition [123]. This umbrella review identified systematic reviews with data from most regions of the world, including the African, European, Southeast Asian, Western Pacific and The Region of the Americas. The data sources used in included systematic reviews covered a broad range of databases including those capturing publications from both HIC and LMIC for example Web of Science, SciElo and LILACS, However, some database capturing data from LMIC e.g. the African Index Medicus, Index Medicus for the Eastern Mediterranean Region were not searched by any review. Future systematic and umbrella reviews should select databases to ensure optimum geographic representation and to ensure all available data from LMIC are captured.

Most systematic reviews reported the impact of SSB taxation on EI with few studies reporting the direct impact on intake of sugars. Restricting included data to original studies identified in the included systematic reviews only may have missed some original data on the impact of taxation on sugars intake. However, to conduct a systematic review of original studies was beyond the scope of this analysis. In the present analysis, Atwater Factors were used to derive intake of sugars from reported changes in EI. This approach assumes that any change in EI is accounted for by a change in intake of sugars, which would underestimate the impact on sugars in scenarios where SSB were replaced with drinks containing energy from nutrients other than sugars (e.g., with milk substitution). For example, although Fletcher et al. [42] reported no impact of taxation on the intake of energy, due to substitution with other sources of energy, this does not equate to no reduction in intake of sugars.

In addition to data from the original studies on the average effect of SSB taxation on sugars intake, published pooled estimates of the mean consumption for HIC and for LMIC [11] were also used to estimate the impact on sugars intake. However, it must be noted that there is considerable variation in SSB intake between HIC and LMIC, which was not captured in this analysis.

Future research

Based on published data, this umbrella review has provided an indication of the reduction in dental caries that might result from SSB taxation in HIC and LMIC. Further cost-effective analysis based on these data will determine the likely impact on cost benefit to healthcare services in LMIC and HIC. The current study has also identified few data from LMIC exploring the dose response between amount of sugars intake and development of dental caries showing that identified data from LMIC pertaining to amount of sugars and dental caries are cross sectional. There is a need for more, well-conducted cohort studies, especially from LMIC, to explore the dose–response relationship between amount of sugars intake and development of dental caries.

Conclusion

Through an umbrella review, the best available evidence pertaining to the impact of SSB taxation on sugars intake and levels of dental caries in both HIC and LMIC has been identified. Evidence indicates a 20% tax would reduce sugars intake by 20.0% in HIC and 18.0% in LMIC, and per capita sugars intake by 4.0 g or more a day. This one intervention alone has a modest positive impact on oral health by reducing caries counts in both adults in both HIC and LMIC by 0.03, and caries prevalence in children by 2.7% and 2.9% for HIC and LMIC respectively.