Labelling changes in response to a tax on sugar-sweetened beverages, United Kingdom of Great Britain and Northern Ireland

Abstract Objective To evaluate the changes in sugar and energy labelling of carbonated sugar-sweetened soft drinks after the implementation of a tax on sugar-sweetened drinks in the United Kingdom of Great Britain and Northern Ireland. Methods We visited nine main supermarkets before (May 2014) and after (April 2018) the tax came into effect and obtained data from product packaging and nutrition information panels of carbonated sugar-sweetened soft drinks. We used the paired t-test to assess differences in sugar and energy content of the same products between 2014 and 2018. Findings We obtained data from 166 products in 2014 and 464 products in 2018, of which 83 products were the same in both years. Large variations in stated sugar content were found between the different carbonated sugar-sweetened soft drinks in both 2014 and 2018 for all products and for the 83 products. The mean sugar content of the 83 products decreased by 42% between 2014 and 2018, from 9.1 g/100 mL (standard deviation, SD: 3.3) to 5.3 g/100 mL (SD: 3.5; P < 0.001). The mean energy content decreased by 40%, from 38 kcal/100 mL (SD: 13) in 2014 to 23 kcal/100 mL (SD: 15) in 2018 (P < 0.001). Conclusion The significant decreases in the labelling of sugar and energy content of carbonated sugar-sweetened soft drinks after the levy came into effect suggest this tax has been effective. The sugar content of drinks still varied considerably in 2018, suggesting further reductions in sugar content of these drinks is possible. The levy thresholds should be reduced and the tax increased to drive further reformulation of soft drinks to reduce their sugar content.


Introduction
In July 2015, the Scientific Advisory Committee on Nutrition of the United Kingdom of Great Britain and Northern Ireland recommended that average free sugars (sugar) intake across the country's population should not exceed 5% of total energy intake. 1 The committee defined free sugars as all monosaccharides and disaccharides added to foods by the manufacturer, cook or consumer, as well as sugars naturally present in honey, syrups (e.g. high fructose corn syrup, glucose syrup and maple syrup) and unsweetened fruit juices. Free sugars do not include lactose when naturally present in milk and milk products, nor sugars contained within the cellular structure of foods (whole fruits and vegetables). 1 However, this definition is new and not aligned with current nutrition labelling on food packaging and manufacturers' claims about sugar content. Manufacturers can still claim that their product contains "no added sugar" even though the product contains fruit or vegetable juice or juice concentrate.
The committee on nutrition also advised that consumption of sugar-sweetened drinks be minimized in children and adults, 1 because high intake of sugar is contributing to obesity, type 2 diabetes and dental caries, 2-10 all of which are major public health problems in the United Kingdom. [11][12][13][14][15][16][17][18] The average sugar intake in 2014 in the United Kingdom exceeded the recommendations of the committee on nutrition in all age groups. 19 The mean sugar intake in adults was 60 g per day, accounting for 12% of daily energy intake. The mean sugar intake in children was 54 g (13.4% of energy intake) per day in 4-10-year-olds and 73 g (14.1%) per day in 11-18-year-olds. 19 Soft drinks, which include all drinks and fruit juices apart from alcoholic drinks, are the main contributors to sugar intake in children (4-10 years) and teenagers (11-18 years), and the second main contributor in adults (18-64 years). 20 Within soft drinks, carbonated soft drinks were the single largest category of the soft drinks market in 2016 with a 38% market share by volume. 21 The volume of carbonated soft drinks sold was 5 billion litres, with an average person consuming 78 L of carbonated soft drinks per year, of which 50% were high sugar drinks. 22 In 2016, the British government announced a tax on sugar-sweetened drinks, the Soft Drinks Industry Levy, to tackle childhood obesity. This tax came into effect in April 2018. The drinks liable for this tax are those that have had sugar added during production, including honey or syrups. Drinks are not liable if they have fruit juice, vegetable juice and milk in them unless these are in addition to sugar. 23 Companies producing sugar-sweetened soft drinks will have to pay 0. 24 British pounds (£) (equivalent to 0.32 United States dollars, US$, in 2018) per litre of drink if the drink contains 8 g or more of sugar per 100 mL or £0.18 (US$ 0.24) per litre of drink if it contains between 5 and 8 g of sugar per 100 mL. Companies can reformulate their soft drink to reduce the sugar content, which may reduce or remove the tax liability.
Product reformulation is commonly described as a voluntary or mandatory effort by government to get manufacturers to lower the unhealthy components (e.g. saturated fat, trans fats, sugar, salt) of food and drink products at the time of production, without making the profile of the overall product worse for consumers (e.g. increasing calorie content). 24  reformulated products often replace an existing product (e.g. the same brand of drink with less sugar). The British government aims to motivate manufacturers of sugar-sweetened soft drinks with 8 g of sugar per 100 mL or more to reduce the levels to below 8 g and pay a lower tax, and manufacturers of drinks with 5-8 g of sugar per 100 mL to lower the sugar levels to less than 5 g and pay no tax. 23 This study aimed to assess changes in the sugar and energy content on labels of carbonated sugar-sweetened soft drinks between 2014 and 2018 after the implementation of the tax on sugarsweetened drinks in United Kingdom. 23 The 2014 data, which were previously published, 25 found that the sugar content of similar products and flavours varied substantially, which suggests that products with lower sugar content were available and that those products with higher content could be reformulated to reduce the amount of sugar.

Study design and setting
The same study design and procedures were used in both 2014 and 2018. 25 We assessed all carbonated sugar-sweetened soft drinks available in the nine main supermarkets in the United Kingdom in May 2014 and April 2018. The retailers were Aldi, Asda, Lidl, Marks and Spencer, Morrisons, Sainsbury's, Tesco, Co-op® Food and Waitrose, which together have over 93.2% of the grocery market share in the country. 26 We checked the products in one large outlet of each of the retailers because the bigger stores tend to stock more own-label and branded products than their smaller branches. Data were collected from the product packaging and nutrient information panels on the products.

Data collection
We defined carbonated sugar-sweetened soft drinks as any carbonated or sparkling drink with added sugar that are not described as energy or sport's drinks on the pack.
Since we focused on the sugar and energy content of drinks, we excluded carbonated sugar-sweetened soft drinks in 2014 labelled as zero or no added sugar. However, in 2018, we included all carbonated sugar-sweetened soft drinks, regardless of their claims. We included all drinks because some manufacturers were claiming that their products contained no added sugar, even when they contained fruit juice or juice concentrates. We manually collected nutrition information from product packaging and nutrition labelling using the data collector applications of Foodswitch United Kingdom, which can be used to take images of food packaging for manual data entry on a database, from which the data can be extracted for analysis. 27 For each product, we collected data on the name of the company that manufactured the product (e.g. The Coca-Cola® Company), brand name of the soft drink (e.g. Fanta™), product name (e.g. Fanta™ Orange, Fanta™ Lemon), pack size, serving size, amounts of sugar (g) and energy (kcal) in the product, per 100 mL and per serving.
We checked all the data again after entry and 5% of the entries against the product label in a random selection of products. We also screened each flavour category for outliers and checked the lowest and highest values in each category against the product labels and corrected, if necessary.
We divided the products into categories of flavour (Box 1), supermarket own label and brand, and manufacturer. If there were more than four products with a particular flavour, we created a category. Products that did not fit into a flavour category were included in the all-products analysis.

Per 100 mL
Some brands sell the same formulation in different serving sizes, e.g. Coca-Cola® comes in 250 mL and 330 mL cans and 500 mL bottles. The formulation per 100 mL is the same. Therefore, for the data per 100 mL, we only included an example of one formulation regardless of the serving size.

Per serving
The data per serving included all the different serving sizes available except for 1 L bottles. We excluded 1 L products from the per serving analysis since we could not accurately quantify how much a consumer would drink if they were drinking from a 1 L bottle, but they are likely to drink more than the industrystandard serving of 250 mL. We considered any product with a can or bottle size up to 500 mL as one serving, regardless of what is stated on the product as a serving size. For example, often a bottle of 500 mL is split into two servings, but we consider that most consumers drink these drinks as one serving.

Level of sugar
We compared the sugar content of the products to the United Kingdom front-of-pack colour-coded labelling for sugar in drinks: red/high sugar > 13.5 g/portion or > 11.25 g/100 mL; a m b e r / m e d i u m s u g a r > 2 . 5 t o ≤ 11.25 g/100 mL; and green/low sugar ≤ 2.5 g/100 mL. 28

Maximum sugar intake
We also compared the sugar content of the products with the maximum daily recommendation for sugar intake for adults (30 g) and 7-10-year-old children (24 g). 1

Soft Drinks Industry Levy
We assessed the products against the levy thresholds: £0.24 per litre of drink if the drink contains 8 g of sugar or more per 100 mL or £0.18 per litre if the drink contains between 5 and 8 g of sugar per 100 mL. 23

Manufacturers
We analysed the sugar and energy content by manufacturer for each manufacturer with five or more products in the sample in both 2014 and 2018. Where a product was imported, but produced by a recognized international manufacturer, we categorized it under the international manufacturer, e.g. a product produced by Coca-Cola® New Zealand was categorized under Coca-Cola®. This categorization helps show the contribution of each manufacturer's products to sugar content in soft drinks and helps in tracking companies' reformulation progress over time.

Statistical analysis
We used the independent-sample t-test to compare the levels of sugar between supermarket own-label and brand products in the entire samples of products in 2014 and 2018. We assumed that supermarkets are more willing to reformulate their own-label products than brand manufacturers who may be more risk averse and reluctant to reformulate products. This reluctance may be because brand manufacturers are influenced by loyalty of their customers to the brand and so they want to continue using old recipes that customers are used to, while supermarkets are constantly changing products and ranges and so are more willing to take risks. We used the paired t-test to examine whether the sugar content of drinks changed significantly between 2014 and 2018. For assessing reformulation, we only included products with data available in both 2014 and 2018 in this analysis.
Descriptive statistics (mean, standard deviation, SD, and range) were calculated. P-values of less than 0.05 were considered statistically significant. All data were analysed using SPSS, version 25 (SPSS Inc., Chicago, United States of America).

Results
We collected 166 and 464 products with nutrition information in 2014 and 2018, respectively. Of these products, 166 and 381 carbonated sugar-sweetened soft drinks met the inclusion criteria per 100 mL in 2014 and 2018, respectively. In 2018, the sample was larger because we included products that claimed to have zero sugar and no added sugar. The mean sugar content per 100 mL was 9.  (Table 1 and Fig. 1). In an analysis of the 249 products without claims of zero sugar and no added sugar in 2018, the mean sugar content was still lower than in 2014, 6.4 g/100 mL (SD: 3.4).
In 2014, on average, grape drinks contained the highest amount of sugar, ranging from 11.5 to 13.5 g/100 mL, followed by ginger beer (5.6 to 16.0 g/100 mL). Cherry-flavoured drinks contained the lowest amount of sugar (Table 1). Similarly, in 2018, on average, grape drinks contained the highest amount of sugar, ranging from 4.3 to 17.9 g/100 mL, followed by elderflowerflavoured drinks. Table 2

Manufacturer comparison
As shown in Table 3, there was a significant difference in sugar content between supermarket own-label and brand products in 2014 (8.4 g vs 9.6 g/100 mL, P = 0.023) and 2018 (3.0 g vs 5.8 g/100 mL, P < 0.001). In 2014, the Princes Ltd product range contained the highest average sugar content per 100 mL (13.3 g; SD: 1.2) and the Co-op® Food range contained the lowest (5.8 g; SD: 3.4; Table 3). In 2018, the Tropical Sun range contained the highest average sugar content per 100 mL (12.3 g; SD: 3.8) and Tesco had the lowest (1.4 g; SD: 2.0). In all the products, Sainsbury's made the biggest reduction in sugar content on average between 2014 and 2018, a reduction of 7.9 g/100 mL.
In 2018, the Tropical Sun product range contained the highest energy content per 100 mL and Tesco and Aldi the lowest (Table 4). In all the samples Cott Beverages Ltd and Sainsbury's made the biggest reduction in energy content on average between 2014 and 2018, a reduction of 32 kcal/100 mL.

Sugar content per serving
We included 43 and 169 products in the analysis per serving in 2014 and 2018, respectively. The serving size

Flavour categories
Individual products

Reformulation
We included the same 83 products in both 2014 and 2018. The mean sugar content per 100 mL for these products was 9.1 g (SD: 3.3) in 2014 and 5.3 g (SD: 3.5) in 2018, a reduction in sugar content of 42% (P < 0.001 by paired t-test). The mean energy content per 100 mL was 38 kcal (SD: 13) in 2014 and 23 kcal (SD: 15) in 2018, a 40% reduction in energy content (P < 0.001 by paired t-test). These averages are slightly different from those when all products were included in each year. This analysis reflects reductions made in the same products rather than the overall products available.

Combined manufacturers
Supermarket own-labels

Discussion
Our study shows early changes in the sugar content on labels of carbon-ated sugar-sweetened soft drinks in the United Kingdom. The comparisons made between the same products over time showed significant reductions in the sugar content of these drinks. These results illustrate the substantial reformulation manufacturers have made so far. The reductions seen are likely to have been prompted by the Soft Drink Industry Levy. 23 Before the announcement of this tax, only Tesco and ASDA had publicly announced their intention to gradually reformulate their products to reduce the sugar content but the tax seems to have speeded up the process. [29][30][31] After the announcement of the Soft Drink Industry Levy with the two-year window given for reformulation, many companies made public announcements about reformulating their products to avoid the tax. [31][32][33][34][35] These announcements support our belief that most of the reductions in sugar seen in our study can be attributed to the Soft Drink Industry Levy.
Since the implementation of the Soft Drink Industry Levy, sales of carbonated sugar-sweetened soft drinks have also fallen, resulting in a decrease in overall consumption of these drinks, 36 as was speculated. 37 Other countries have imposed taxes on sugar-sweetened drinks and some have also seen reductions in sales. 38,39 Despite this reduction, sugar content remains high: many of the drinks in 2018 would still receive a red (high) label for sugars per serving. The sugar content of similar products varied considerably, judging by the wide ranges, which demonstrates that further reformulations are possible. If further reductions are needed, and if they are based on the 2018 data on sugar and energy content, the thresholds of the Soft Drink Industry Levy could be reduced and the tax paid increased to drive further incremental reformulation.
We also analysed the sugar content of soft drinks by manufacturer, which showed that the products that were high in sugar were generally produced by a few main brand manufacturers. These manufacturers tend just to pay the tax instead of reformulating their products to reduce sugar content. Reformulation is the responsibility of the manufacturers and they need to be targeted with measures that push them to reformulate their products, such as higher taxes and/ or lower thresholds.
Manufacturers of the reformulated products in our study have either reduced the total sugar or reduced sugar by replacing it with non-caloric sweeteners without changing the product name. Replacing sugar with non-caloric sweeteners has been associated with a lower risk of obesity. 40 Nevertheless, there is insufficient evidence on the relationship between the use of non-caloric sweeteners and long-term weight control. 41 Future research can focus on continuing the collation and documentation of sugar and energy content of products. Our study provides categories of products that can be monitored in the future. The data collected can be combined with sales data of best-selling products to monitor change over time and the effect on the intake of sugar in the population.
Our study has limitations. We used sugar and energy content data provided on the nutrition labels of drinks products that were available in stores on the days the products were collected. Therefore, our data are dependent on the accuracy of the data provided on the label and the availability of products in stores. Manufacturers are assumed to provide accurate and up-to-date information on their packaging in line with European Union regulations. Moreover, most products are required by law to have nutrition information on packaging in the United Kingdom and manufacturers will likely provide the correct figures. Therefore, the data collected from the label are likely to match the true sugar content.
Products available on supermarket shelves are not always strictly in line with the descriptions used by the industry. For instance, carbonated sugar-sweetened soft drinks in 2014 labelled as zero sugar or no added sugar were excluded. However, in 2018, we included all these drinks, regardless of such claims. We decided to include these drinks because some manufacturers claimed their products contained no added sugar, but they still added juice or juice concentrates to them which contain sugar. To minimize the potential influence of the products claiming zero sugar and no added sugar in 2018, we did separate analyses that (i) included only the products surveyed in both 2014 and 2018 and (ii) excluded products claiming zero and no added sugar in both years. These analyses showed a consistent and significant reduction in sugar in carbonated sugar-sweetened soft drinks.
Reformulations to reduce sugar in carbonated soft drinks are contributing to achieving the recommendation that sugar intake should not exceed 5% of total energy intake. 1

Research
Sugar and energy in carbonated sugar-sweetened soft drinks Kawther M Hashem et al.