Screening level intake estimates of low and no-calorie sweeteners in Argentina, Chile, and Peru

ABSTRACT

As obesity rates increase, several countries in Latin America have implemented strategies to curb the consumption of sugars, resulting in reformulations of products with low and no-calorie sweeteners (LNCS). The increased availability of LNCS-containing products raises concerns about the potential risk of exceeding the Acceptable Daily Intake (ADI). Information on the intake of LNCS among Latin American countries are limited by the lack of publicly available national consumption data. Using the Budget Method, screening level intake estimates of six LNCS (acesulfame potassium, aspartame, cyclamate, saccharin, steviol glycosides, and sucralose) were derived for Argentina, Chile, and Peru based on national sales data and product labels. Four tiered assessments were conducted where assumption of LNCS use ranged from the most conservative Tier 1 to the more refined yet conservative LNCS use and concentrations in subsequent tiers. The estimated intakes, applicable to the total population as well as children, were below their ADIs for all tiers. For Tier 2 where average LNCS concentrations were assumed present in all LNCS-containing products, intakes were <60% of the ADI. Estimates for the more refined tiers were comparable to published estimates based on select subpopulations in these countries, validating the approach used in this study.

KEYWORDS:

• Argentina
• Chile
• Peru
• LNCS
• Budget Method
• non-nutritive sweeteners
• high-intensity sweeteners
• acesulfame potassium
• aspartame
• cyclamate
• saccharin
• steviol glycosides
• sucralose

Introduction

In response to concerns about the potential association between sugar intakes and high obesity rates and guidelines from the World Health Organization (WHO) that recommend reducing the level of sugar in the diet to less than 10%, or even less than 5% (conditional recommendation based on very low and moderate evidence in the case of dental health) (WHO 2015), several countries have implemented regulatory strategies to curb the consumption of sugars, such as restricting the sales of sugar-sweetened beverage (SSB) in schools, establishing restrictions on advertising for SSB, imposing higher taxes on SSB, while reducing taxes on other beverages with no or low sugar content, and implementing front-of-pack (FOP) labelling strategies to better inform consumers (Bergallo et al. 2018; WCRFI 2018, 2019, 2020; Carriedo et al. 2021). Food and beverage producers have consequently reformulated products to include less sugar or replaced sugars with low and no-calorie sweeteners (LNCS) to comply with guidance to reduce sugar intakes and regulatory strategies implemented by individual governments. Similarly, consumers have turned to LNCS in lieu of sugar as one possible alternative to decrease their sugar intake.

LNCS are some of the most studied food additives; their uses in food and beverages are typically regulated and maximum allowable levels have been set by government agencies to ensure public safety. However, there are concerns about the increased uses of LNCS (Sylvetsky and Rother 2016) and the potential risk of exceeding the respective LNCS Acceptable Daily Intake (ADI) (Serra-Majem et al. 2018). A comprehensive review of the global intake data on the major LNCS (acesulfame potassium, aspartame, saccharin, sucralose, cyclamate, and thaumatin) using literature published during an 10-year period from 2008 to 2017 did not show any exceedance of individual sweetener ADIs among the general population (Martyn et al. 2018). It was however noted in this review that LNCS intake data for Latin America were based on small studies that were not nationally representative. The majority of recent studies investigating LNCS intakes in Latin American countries published after the review by Martyn et al. (2018) have been focused on specific subpopulations, e.g. pregnant women in Chile (n = 601) (Fuentealba Arevalo et al. 2019), children and adolescents living in urban areas in Chile (n = 250) (Martínez et al. 2020), low-medium income pre-schoolers in Chile (n = 959) (Venegas-Hargous et al. 2020), and children and adolescents living in urban areas in Argentina (n = 2664) (Garavaglia et al. 2018).

Two recent studies have used nationally representative data from the Pesquisa de Orcçamentos Familiares (POF), 2008–2009 conducted in Brazil by the Instituto Brasileiro de Geografia e Estatistica (IBGE). The first study (Silva Monteiro et al. 2018) reported only on the prevalence of use of table-top LNCS and did not estimate the amount of LNCS consumed, while the second (Barraj et al. 2021) used LNCS use rates and market share data provided by industry to derive nationally representative estimates of six LNCS intakes from table-top uses as well as from pre-packaged foods. We are not aware of any other similar studies reporting on nationally representative intake of LNCS in Latin America, in part because the representative data needed to evaluate the intakes of LNCS in Latin America are often not available at a national level. In the absence of data from national surveys, model diets such as the Budget Method (Hansen 1966, 1979; WHO 2001) can be used. The method is based on default assumptions regarding consumption amounts (per kg body weight) of solid foods and of non-milk beverages (calculated based on the caloric needs and intakes of young children), maximum concentration of the additive, and the proportion of solid foods and of non-milk beverages that may contain the additive. The Budget Method has been used to assess the theoretical maximum daily dietary exposure (TMDE) to food additives (IPCS 2009). Refinements to the method, such as using more realistic consumption estimates, sales data to estimate the fraction of the food supply that could contain the food additive and concentration data obtained from product labels or provided by industry, result in more refined, albeit still conservative, screening level intake estimates (Tran et al. 2020, 2021).

The purpose of this study was to conduct tiered screening level intake assessments of six LNCS using the budget method approach and country-specific sales and LNCS use data for three Latin American countries with no publicly available nationally representative LNCS intake data available, including Argentina, Chile, and Peru. All three countries have LNCS labelling regulation in place or under development. In addition, there are several recent publications summarising LNCS intake in select subpopulations in Chile, hence estimated intakes in these publications could be used to evaluate our proposed approach. The six LNCS covered in these assessments were acesulfame potassium, aspartame, cyclamate, saccharin, steviol glycosides, and sucralose. The sales data were obtained from Euromonitor International’s global research database known as Passport (Euromonitor International 2020) and the LNCS use level data were obtained from Mintel’s Global New Products Database (GNPD) (Mintel 2020).

Materials and methods

Data

Euromonitor International Passport database

Euromonitor International Passport database (Euromonitor International 2020) collects data on sales of packaged foods and beverages (henceforth, the Market Size database) as well as on ingredients used by the producers of these products (henceforth, the Ingredients database). The Market Size database includes data on the total volume of sales, differentiated by retail (supermarkets, retailers, etc.) and restaurants. The Ingredients database contains information on the volume of select ingredients, including LNCS (referred to as “high intensity sweeteners” in the database), used in each product category. The LNCS information in the Ingredients database can be accessed either via a listing that identifies the total volume of LNCS used in the various food and beverage categories as a percentage of all LNCS used in the country (“ingredients by product” listing), or via a listing that identifies all ingredients for each product category as well as their contribution to the total volume of that category (“products by ingredients” listing).

Actual and projected sales data for the period 2017–2023 from the Market Size and Ingredients databases were extracted from the Euromonitor website for use in these analyses.

Mintel Global New Products Database

The Mintel GNPD monitors consumer packaged goods globally. The range of information compiled on each product includes, among others, the ingredient list and nutritional information (as they appear on the package) (Mintel 2020). The database typically includes several pictures of each product, including pictures of the ingredients list, nutrition label and preparation instructions, if applicable. Data for the last three complete years (2017–2019) were downloaded. The ingredient lists, nutrition label data, and amount of sweetener per serving (Chile only) in the Mintel GNPD provide a listing of which sweeteners, if any, are used in the various products as well as the corresponding amounts, when available. Information from this database was used to estimate the fraction of LNCS-containing products that contain the specific sweeteners and the associated concentrations for each relevant beverage and food category.

Methods

The Budget Method

The Budget Method is based on default assumptions to assess the TMDE to food additives. Parameters of the Budget Method include the estimates of consumption of solid foods and non-milk beverages, maximum concentration of the additive in foods and non-milk beverages, and the proportion of solid foods and non-milk beverages that may contain the additive. The TMDE (mg/kg bw/day) is derived in the Budget Method using Equation (1) (IPCS 2009):

(1) $TMDE=\left[Ma{x}_{(FA-B)}xCon{s}_{B}xPc{t}_{(FA-B)}\right]+\left[Ma{x}_{(FA-F)}xCon{s}_{F}xPc{t}_{(FA-F)}\right]$(1)

where:

Max_(FA-B) (or Max_(FA-F)): Maximum level of food additive in non-milk beverages (mg/L) (or solid foods (mg/kg))

Cons_B (or Cons_F): Amount of non-milk beverages (L/kg bw/day) (or solid foods (kg/kg bw/day)) consumed per day

Pct_(FA-B) (or Pct_(FA-F)): Percent of non-milk beverages (or solid foods) that contain the food additive.

The default consumption estimates used in the calculation of the TMDE are 0.1 litre of non-milk beverages/kg bw/day and 100 kcal/kg bw from solid foods, which is equivalent 0.05 kg/kg bw/day, assuming an energy density of 2 kcal/g (Hansen 1966, 1979). The default consumption amounts of 0.05 kg/kg bw/day and 0.1 L/kg bw/day used in the Budget Method are based on the caloric requirements of young children and the range of potential beverage intakes of infants and young children, and hence are conservative as those needs (on a kg bw basis) decrease with increasing age.

In the default application of the method to estimate the TMDE, the concentration of the additive in solid foods and non-milk beverages is typically assumed to be the maximum reported use level of the additive. The proportions of solid foods and non-milk beverages that may contain an additive can either be arbitrary estimates, e.g. default proportions of 12.5% for solid foods and 25% for non-milk beverages are often used in Europe, although higher proportions and even 100% was used for beverages when estimating exposures for children (WHO 2012), or alternatively, country and additive specific market share and label data can be used (WHO 1999).

Processing the Euromonitor and Mintel data

In the current study, the fraction of foods and beverages that contain the specific sweeteners, i.e. Pct_(FA-B) and Pct_(FA-F) in Equation (1) was estimated as the product of two quantities derived from the Euromonitor and Mintel databases: (i) the fraction of all beverages and foods that could contain LNCS and (ii) the fraction of LNCS-containing beverages and foods that use the specific sweeteners. The concentrations of each LNCS in various food and beverage categories were obtained from the Mintel database and combined to obtain weighted concentration estimates in all beverages and all foods.

The concentrations reported in the Mintel GNPD in foods and beverages were expressed on an as consumed basis by reviewing individual product labels and applying the dilution factors listed on the labels. Concentrations for stevia and steviol glycosides were re-expressed on a steviol basis as was cyclamate on a cyclamic acid basis, as needed.

The information extracted from the Euromonitor and Mintel databases and the approach used to process and combine the data are described below and summarised in Figure 1 (Additional information is provided in the Supplemental Materials).

Figure 1. Process to combine data from the Euromonitor and Mintel databases

Beverages

The list of beverage categories in the Euromonitor database includes Alcoholic drinks, Hot beverages, and Soft drinks which are further broken down into subcategories. (a more detailed listing of the beverage categories is provided in the Supplemental Materials).

The Euromonitor ingredients by product listing was reviewed to identify which beverage categories could contain LNCS, and the LNCS volume data for those beverage categories was then obtained from the Euromonitor products by ingredient listing. Since greater than 85% of the LNCS volume used in food, beverages, and non-food and beverage products (e.g. cosmetics, oral care products) was reported to be used in the Soft drinks category, we focused our analysis on the Soft drinks category.

The Soft drinks category includes, among others, a Carbonates category, that was further broken down into Cola Carbonates and Non-Cola Carbonates. The Cola Carbonates category is also broken down into a Low Calorie Cola Carbonates category and a Regular Cola Carbonates category. Only the Low Calorie Cola Carbonate category in the list of beverage categories available from the Euromonitor database was identified as low calorie and hence, the total volume of this category was assumed to contain at least one LNCS. Since all other beverage categories can include LNCS-containing products as well as “regular” products, the volume of LNCS-containing products within each category was back-calculated from the total volume of that beverage category and an estimated total LNCS concentration that was anchored on the total LNCS concentration in the Low Calorie Cola Carbonate category (see Supplemental Materials).

Beverage categories in the Mintel GNPD database (see Supplemental Materials) were re-grouped, as needed, to categories matching those in the Euromonitor database. The concentration in ppm (or mg/L or mg/kg) of each of the six sweeteners of interest was derived based on the concentrations reported in ingredient lists or from the amount per serving and serving size information reported on nutrition labels. It should be noted that in some instances products that were identified as containing a LNCS of interest based on the ingredient listing did not always contain information on the sweetener concentration or the amount per serving on the product labels.

The estimate of the volume of LNCS-containing products in each beverage category obtained from the Euromonitor database, and the estimated fractions of LNCS-containing products that use specific sweeteners from the Mintel GNPD in each beverage category were combined to derive an estimated fraction of all beverages combined that contain the specific sweeteners for use in the application of the Budget Method. For each sweetener, the average and maximum concentrations for each beverage category were obtained from the Mintel database and were used to derive a weighted overall average and maximum concentrations across all beverages. The concentrations were weighted based on the estimated volume of the category in the Euromonitor database (see Supplemental Materials).

Packaged foods

Data from the Euromonitor ingredients by product listing indicated that 95% (Argentina), 96% (Chile), and 86% (Peru) of all LNCS used in food and beverage products are used in beverages, and in each country, at least 90% of the LNCS volume used in packaged foods was used in four categories: chewing gum, cookies, dessert mixes, and yoghurt. The sum of the ingredients in the products by ingredient listing for packaged foods was different from the total volume of that category in the Market Size database, indicating that, in the case of packaged foods, the products by ingredient may not include all ingredients. Further, LNCS concentrations in foods are likely to be different depending on the food category (e.g. LNCS concentrations in candy are likely very different from concentrations in desserts or other food products like cookies and baked goods). Hence, the approach used for beverages to estimate the amount (and fraction) of LNCS-containing beverages by anchoring on the LNCS concentration in the Low Calorie Cola Carbonates category is not feasible, and data from the Mintel GNPD for chewing gum, cookies, desserts, and ice cream (mapped to the dessert mixes category in the Euromonitor database), and yoghurt, were used to estimate the fraction of products within each of these categories that could contain LNCS. Specifically, the estimate of the fraction of foods that contain LNCS was derived as the ratio of the total number of products in each of the four categories to the total number of food products in the database. The potential uncertainty associated with only using these four food categories for the overall estimated intake of sweeteners is likely to be minimal, not only because 90% of the LNCS volume used in packaged foods was present in these four categories but also because, as mentioned above, 86% to 96% of the LNCS used in foods and beverages were used in beverages. Following recommendations from the Food and Agriculture Organization of the United Nations and the World Health Organisation (IPCS 2009) to exclude categories with particularly high concentrations, the overall average and maximum concentration estimates for foods were derived excluding chewing gum since a review of the sweetener concentrations across the four categories indicated that the concentrations in chewing gum were much higher than those for the other categories. The overall average and maximum concentration levels for the food categories (cookies, desserts and ice cream, and yoghurt) were derived as weighted averages of the average and maximum concentrations for each product category weighted by the number of products in the Mintel database category (see Supplemental Materials).

Exposure scenarios

The TMDE of the various sweeteners were derived using Equation 1 under four tiered exposure scenarios that were conducted for each sweetener. As mentioned above, parameters used to calculate the TMDE include default consumption estimates of 0.1 litre of non-milk beverages/kg bw/day and 0.05 kg/kg bw/day from solid foods based on WHO (2001); the average or maximum percentage of LNCS-sweetened beverages calculated from Euromonitor data; default 12.5% solid foods that can contain the specific sweetener for Tiers 1 and 2 based on WHO (2012) with Tiers 3 and 4 assuming the average percentage of solid foods calculated from the Mintel data; worst case assumption of 100% of specific sweetener within LNCS-sweetened foods and beverages or a weighted average percentage based on a combination of data from Mintel and Euromonitor; and the weighted average or maximum concentration of the specific sweetener in foods and beverages based on a combination of data from Mintel and Euromonitor. Table 1 lists the four scenarios and assumptions used in each.

Table 1. Exposure scenarios and assumptions

Parameter	Tier 1	Tier 2	Tier 3	Tier 4
Amount of beverage consumed (L/kg bw/day)^1	0.1	0.1	0.1	0.1
Amount of food consumed (kg/kg bw/day)^1	0.05	0.05	0.05	0.05
% LNCS-sweetened beverages^2	Maximum	Maximum	Average	Average
% LNCS-sweetened foods^3	12.5%	12.5%	Average	Average
% of specific sweetener within LNCS-sweetened foods and beverages^4	100%	100%	Weighted Average	Weighted Average
Specific LNCS concentration in foods and beverages^5	Weighted Maximum	Weighted Average	Weighted Maximum	Weighted Average

¹WHO default (WHO 2001). ² As derived from Euromonitor data. ³ WHO default for Tiers 1 & 2 (WHO 2012); as derived from Mintel data for Tiers 3 & 4. ⁴ Worst case (100%) for Tiers 1 & 2, as derived from Mintel and Euromonitor data for Tiers 3 & 4. ⁵ As derived from Mintel and Euromonitor data.

Results

Beverages: fraction containing LNCS and sweeteners concentrations

Estimated fraction of beverages that contain LNCS

Table 2 summarises the volume of sales of the various beverages and the calculated percent of beverages that contain LNCS. The percent of beverages containing LNCS ranged from 1.1% for energy drinks in Peru to 33.5% for cola carbonates in Argentina. LNCS containing beverages constituted 15.2%, 167.4%, and 5.8%, respectively of all beverages in Argentina, Chile, and Peru.

Table 2. Volume of sales and estimated fraction of beverages that contain LNCS

Country and beverage category	Volume of sales (000 tonnes)		% Containing LNCS
	All ^1	LNCS-containing^2
Argentina
Cola carbonates	3190.6	1064.7	33.5%
Low calorie cola carbonates^1	1064.7	1064.7	100%
Non cola carbonates	2296.7	332.1	14.5%
Bottled water	4714.5	253.7	5.4%
Juice	NA	NA	NA
Concentrates (liquid)	256.6	16.2	6.3%
Concentrates (powder)	32.2	3.5	11.0%
Energy drinks	NA	NA	NA
Sports drinks	66.3	9.8	14.8%
Total^3	11064.7	1680.0	15.2%
Chile
Cola carbonates	1718.0	423.5	24.7%
Low calorie cola carbonates	423.5	423.5	100%
Non cola carbonates	809.5	154.5	19.1%
Bottled water	769.3	49.3	6.4%
Juice	417.0	8.5	2.0%
Concentrates (liquid)	32.6	2.1	6.3%
Concentrates (powder)	8.3	0.2	2.9%
Energy drinks	NA	NA	NA
Sports drinks	58.2	8.5	14.7%
Total^3	3933.6	646.6	16.4%
Peru
Cola carbonates	922.7	63	6.8%
Low calorie cola carbonates	62.9	62.9	100%
Non cola carbonates	799.0	52.1	6.5%
Bottled water	837.2	17.4	2.1%
Juice	377.6	34.0	9.0%
Concentrates (liquid)	NA	NA	NA
Concentrates (powder)	2.0	0.1	2.9%
Energy drinks	38.6	0.4	1.1%
Sports drinks	144.2	21.3	14.7%
Total^3	3250.9	188.2	5.8%

¹Extracted from the Euromonitor database. ² Derived by anchoring on the total LNCS concentration in the low calorie cola carbonate category (see Supplemental Materials). ³ The “All” Total does not match the sum for the individual categries as it includes additional beverage categories that were not reported to contain LNCS in the Euromonitor database.

LNCS-containing beverages that use specific sweeteners and sweetener concentrations

Table 3 summarises the percent of LNCS-containing beverages that use the specific sweeteners and the weighted average and maximum concentrations across beverage products, where the average and maximum concentrations per beverage category were extracted from the Mintel database, and the volume of sales from Euromonitor were used to derive the weighted average and maximum across beverage categories (see Methods Section and Supplemental Materials). Acesulfame potassium and aspartame are the most used beverage sweeteners in Argentina, followed by cyclamate and saccharin, and finally by sucralose and steviol glycosides. Acesulfame potassium and aspartame are also the most used sweeteners in Chile, followed by sucralose, then steviol glycosides. In the Mintel database, no beverage products were identified as containing cyclamate or saccharin in Chile, hence it was arbitrarily assumed that the fraction of beverages that contain these sweeteners is one half the smallest fraction among the other sweeteners (i.e. 11% for steviol glycosides). While the Mintel database included LNCS containing beverage products in Peru, the LNCS concentration data were limited, hence, the averages of the LNCS concentrations from Argentina and Chile were used to derive the average and maximum concentrations assigned to beverages in Peru (additional details provided in the Supplemental Materials).

Table 3. Fraction of LNCS-containing beverages and concentration of sweeteners by country

	LNCS
	Acesulfame Potassium	Aspartame	Cyclamate^1	Saccharin	Steviol Glycosides	Sucralose
Argentina
% Total LNCS beverages^2	60%	53%	40%	40%	8%	18%
Concentration (ppm)^3
Mean	119	256	171	26	50	28
Maximum	197	389	261	53	69	56
Chile^4
% Total LNCS beverages^2	67%	57%	6%	6%	11%	28%
Concentration (ppm)^3
Mean	125	319	171	26	38	102
Maximum	166	428	261	53	48	146
Peru
% Total LNCS beverages^2	71%	44%	0.10%	5%	10%	42%
Concentration (ppm)^3,5
Mean	123	259	209	20	60	51
Maximum	184	400	327	42	65	76

¹Cyclamate concentrations expressed as cyclamic acid. ² Based on the sum of the total volume of beverages that use the specific LNCS and total volume containing any sweetener across all beverage categories. ³ Overall concentration (mean and maximum) across all beverages were derived as the weighted average concentration for each beverage category based on concentration data from the Mintel database and total volume from the Euromonitor database (refer to text and the Supplemental Materials for more details). ⁴ No beverages were identified as containing cyclamate or saccharin in the Chile Mintel GNPD, hence it was assumed that the fraction of LNCS-containing beverages in Chile that use these two sweeteners is 6% (or ½ the minimum estimate from the other sweeteners). In addition, cyclamate and saccharin concentrations from Argentina were assumed. ⁵ Concentration data for Peru were limited, hence the average concentrations from Argentina and Chile were assumed.

Packaged foods: fraction containing LNCS and sweeteners concentrations

Table 4 lists the number of food products in the Mintel database and the fraction that contain LNCS. Per the Euromonitor data described above, less than 10% of the LNCS tonnage used in packaged foods was used in categories other than Gum, Cookies, Dessert and ice cream, and Yoghurts, hence the LNCS contribution of these “Other” products was considered negligible. The calculated percent of food products that contain LNCS was 2% in Argentina and Peru, and 8% in Chile.

Table 4. Number of food products in the Mintel database that contain LNCS

Country and Food Category	Total N	With LNCS
		N	%
Argentina
Cookies	772	20	3%
Desserts & Ice Cream	212	100	47%
Gum	28	26	93%
Yoghurts	161	44	27%
Other^1	6,275	NA	NA
All^2	7,420	164	2%
Chile
Cookies	179	43	24%
Desserts & Ice Cream	138	44	32%
Gum	12	11	92%
Yoghurts	189	131	69%
Other^1	2,200	NA^1	NA
All^2	2,706	218	8%
Peru
Cookies	338	16	5%
Desserts & Ice Cream	177	45	25%
Gum	17	12	71%
Yoghurts	87	16	18%
Other^1	3,531	NA	NA
All^2	4,133	77	2%

¹Per Euromonitor data described above, less than 10% of the LNCS tonnage used in packaged foods was used in categroies other than Gum, Cookies, Dessert mixes, and Yoghurts. The LNCS contribution of these “Other” prodcuts was considered negligible. ² Excludes gum.

Table 5 summarises the percent of LNCS-containing food products that use the specific sweeteners and the weighted average and maximum concentrations across food products, where the average and maximum concentrations per food product category, excluding chewing gum, were extracted from the Mintel database, and the volume of sales from Euromonitor were used to derive the weighted average and maximum across food product categories (see Methods Section and Supplemental Materials). In Argentina, acesulfame potassium and aspartame are the most used sweeteners in food products, followed by sucralose, then saccharin and steviol glycosides, and finally by cyclamate. In Chile, sucralose and steviol glycosides are also the most used sweeteners, followed by acesulfame potassium, and 1% or less of LNCS food products contain aspartame, cyclamate, or saccharin. No foods products were identified as containing cyclamate in the Chile Mintel GNPD, hence it was arbitrarily assumed that the fraction of LNCS-containing foods in Chile that use this sweetener is one half the smallest fraction for the other sweeteners (i.e. 0.5% for aspartame). No foods products were identified as containing cyclamate or saccharin in the Peru Mintel GNPD, hence it was arbitrarily assumed that the fraction of LNCS-containing foods in Peru that use these sweetener is one half the smallest fraction for the other sweeteners (i.e. 39% for aspartame) (additional details provided in the Supplemental Materials).

Table 5. Fraction of LNCS-containing food products and concentration of sweeteners by country

	LNCS
	Acesulfame Potassium	Aspartame	Cyclamate^1	Saccharin	Steviol Glycosides	Sucralose
Argentina
% Total LNCS food products^2	69%	50%	12%	13%	13%	42%
Concentration (ppm)^2
Mean	112	219	131	178	34	149
Maximum	284	547	267	227	100	290
Chile^3,4
% Total LNCS food products^2	7%	0.5%	0.2%	1%	55%	85%
Concentration (ppm)^2
Mean	119	270	131	126	71	146
Maximum	195	270	267	147	186	346
Peru^3,5
% Total LNCS food products^2	61%	39%	19%	19%	44%	43%
Concentration (ppm)^2
Mean	115	244	131	152	53	148
Maximum	240	408	267	187	143	318

¹Cyclamate concentrations expressed as cylcamic acid. ² The fractions of LNCS-containing foods that contain the specific sweeteners and the overall concentration (mean and maximum) across all packaged foods were derived using the data from the Mintel GNPD for individual food categories weighted by the volume of sales from the Euromonitor database (refer to the Methods Section and the Supplemental Materials for more details). ³ No foods products were identified as containing cyclamate in the Chile Mintel GNPD, or cyclamate or saccharin in the Peru Mintel GNPD, hence it was assumed that the fraction of LNCS-containing foods in Chile that use cyclamate is 0.2% (or ½ the minimum estimate from the other sweeteners), and the fraction of LNCS-containing foods in Peru that use these cyclamate or saccharin is 19% (or ½ the minimum estimate from the other sweeteners). ⁴ The overall cyclamate concentration from Argentina were assumed. ⁵ Concentration data for Peru were limited, hence the average concentrations from Argentina and Chile were assumed.

Estimated LNCS intakes

Estimated intakes by source

Estimated intakes of the six sweeteners from beverages, foods, and beverages and foods combined for Argentina, Chile and Peru are summarised in Tables 6–8, respectively.

Table 6. Screening level estimated intakes of LNCS in Argentina using the Budget Method approach by assessment tier (mg/kg bw/day)

LNCS (JECFA ADI)	Food Supply	Estimated Intakes in Argentina
		(mg/kg bw/day)
		Tier 1	Tier 2	Tier 3	Tier 4
Acesulfame Potassium (15 mg/kg bw/day)	Beverages	6.6	4	1.78	1.08
	Foods	1.78	0.7	0.22	0.09
	Total	8.37	4.7	2	1.17
	Total (% ADI)	56%	31%	13%	8%
Aspartame (40 mg/kg bw/day)	Beverages	13.07	8.6	3.15	2.07
	Foods	3.42	1.37	0.3	0.12
	Total	16.48	9.97	3.45	2.19
	Total (% ADI)	41%	25%	9%	5%
Cyclamate^1 (11 mg/kg bw/day)	Beverages	8.76	5.74	1.58	1.03
	Foods	1.67	0.82	0.03	0.02
	Total	10.42	6.56	1.61	1.05
	Total (% ADI)	95%	60%	15%	10%
Saccharin (5 mg/kg bw/day)	Beverages	1.79	0.86	0.32	0.15
	Foods	1.42	1.11	0.03	0.03
	Total	3.21	1.97	0.36	0.18
	Total (% ADI)	64%	39%	7%	4%
Steviol Glycosides (4 mg/kg bw/day)	Beverages	2.33	1.68	0.08	0.06
	Foods	0.63	0.21	0.01	0.005
	Total	2.96	1.9	0.09	0.06
	Total (% ADI)	74%	47%	2%	2%
Sucralose (15 mg/kg bw/day)	Beverages	1.87	0.95	0.16	0.08
	Foods	1.81	0.93	0.13	0.07
	Total	3.68	1.88	0.29	0.15
	Total (% ADI)	25%	13%	2%	1%

¹Expressed as cylcamic acid for comparison with the ADI.

Table 7. Screening level estimated intakes of LNCS in Chile using the Budget Method approach by assessment tier (mg/kg bw/day)

LNCS (JECFA ADI)	Food Supply	Estimated Intakes in Chile
		(mg/kg bw/day)
		Tier 1	Tier 2	Tier 3	Tier 4
Acesulfame Potassium (15 mg/kg bw/day)	Beverages	4.1	3.1	1.84	1.39
	Foods	1.22	0.74	0.06	0.04
	Total	5.32	3.84	1.9	1.42
	Total (% ADI)	35%	26%	13%	9%
Aspartame (40 mg/kg bw/day)	Beverages	10.59	7.88	3.98	2.96
	Foods	1.69	1.69	0.005	0.005
	Total	12.28	9.57	3.98	2.97
	Total (% ADI)	31%	24%	10%	7%
Cyclamate^1 (11 mg/kg bw/day)	Beverages	6.46	4.24	0.24	0.16
	Foods	1.67	0.82	0.002	0.001
	Total	8.12	5.05	0.25	0.16
	Total (% ADI)	74%	46%	2%	1%
Saccharin (5 mg/kg bw/day)	Beverages	1.32	0.63	0.05	0.02
	Foods	0.92	0.79	0.01	0.01
	Total	2.24	1.42	0.06	0.03
	Total (% ADI)	45%	28%	1%	1%
Steviol Glycosides (4 mg/kg bw/day)	Beverages	1.19	0.94	0.09	0.07
	Foods	1.16	0.45	0.41	0.16
	Total	2.35	1.39	0.5	0.23
	Total (% ADI)	59%	35%	13%	6%
Sucralose (15 mg/kg bw/day)	Beverages	3.61	2.53	0.67	0.47
	Foods	2.16	0.91	1.18	0.5
	Total	5.77	3.44	1.85	0.97
	Total (% ADI)	38%	23%	12%	6%

¹Expressed as cylcamic acid for comparison with the ADI.

Table 8. Screening level estimated intakes of LNCS in Peru using the Budget Method approach by assessment tier (mg/kg bw/day)

LNCS (JECFA ADI)	Food Supply	Estimated Intakes in Peru
		(mg/kg bw/day)
		Tier 1	Tier 2	Tier 3	Tier 4
Acesulfame Potassium (15 mg/kg bw/day)	Beverages	2.71	1.82	0.75	0.5
	Foods	1.5	0.72	0.14	0.07
	Total	4.21	2.54	0.89	0.57
	Total (% ADI)	28%	17%	6%	4%
Aspartame (40 mg/kg bw/day)	Beverages	5.9	3.81	1.02	0.66
	Foods	2.55	1.53	0.15	0.09
	Total	8.45	5.34	1.16	0.74
	Total (% ADI)	21%	13%	3%	2%
Cyclamate^1 (11 mg/kg bw/day)	Beverages	4.82	3.07	0.002	0.001
	Foods	1.67	0.82	0.05	0.02
	Total	6.49	3.89	0.05	0.03
	Total (% ADI)	59%	35%	0%	0%
Saccharin (5 mg/kg bw/day)	Beverages	0.63	0.3	0.01	0.01
	Foods	1.17	0.95	0.03	0.03
	Total	1.79	1.25	0.05	0.03
	Total (% ADI)	36%	25%	1%	1%
Steviol Glycosides (4 mg/kg bw/day)	Beverages	0.95	0.89	0.04	0.03
	Foods	0.89	0.33	0.06	0.02
	Total	1.85	1.22	0.1	0.06
	Total (% ADI)	46%	30%	2%	1%
Sucralose (15 mg/kg bw/day)	Beverages	1.12	0.75	0.19	0.12
	Foods	1.99	0.92	0.13	0.06
	Total	3.11	1.67	0.31	0.18
	Total (% ADI)	21%	11%	2%	1%

¹Expressed as cylcamic acid for comparison with the ADI.

For Argentina (see Table 6), the total estimated intakes of acesulfame potassium, aspartame, and cyclamate ranged from 1.17 to 8.37 mg/kg bw/day, 2.19 to 16.48 mg/kg bw/day, and 1.05 to 10.42 mg/kg bw/day, respectively, across the four tiered scenarios. Total estimated intakes of saccharin, steviol glycosides, and sucralose were less than the other sweeteners and ranged from 0.18 to 3.21 mg/kg bw/day, 0.06–2.96 mg/kg bw/day, and 0.15–3.68 mg/kg bw/day, respectively.

For Chile (see Table 7), the total estimated intakes of acesulfame potassium, aspartame, cyclamate, and sucralose ranged from 1.17 to 8.37 mg/kg bw/day, 2.19 to 16.48 mg/kg bw/day, 1.05 to 10.42 mg/kg bw/day, and 0.97–5.77 mg/kg bw/day, respectively, across the four tiered scenarios. Total estimated intakes of saccharin and stevia were less than the other sweeteners and ranged from 0.03 to 2.24 mg/kg bw/day and 0.23 to 2.35 mg/kg bw/day, respectively.

For Peru (see Table 8), the total estimated intakes of acesulfame potassium, aspartame, cyclamate, and sucralose ranged from 0.57 to 4.21 mg/kg bw/day, 0.74 to 8.45 mg/kg bw/day, 0.03 to 6.49 mg/kg bw/day, and 0.18 to 3.11 mg/kg bw/day, respectively, across the four tiered scenarios. Total estimated intakes of saccharin and stevia were less than the other sweeteners and ranged from 0.03 to 1.79 mg/kg bw/day and 0.06 to 1.85 mg/kg bw/day, respectively.

Figures 2–5 summarise the percent contributions of beverages and foods to the total estimated intakes for each Tier. In Tier 1 (Figure 2), for all sweeteners, contribution to total intake from beverages were generally comparable across countries, and typically higher than the contribution from food, except for saccharin and sucralose in Peru. Similar results were observed in Tier 2 (Figure 3), except for saccharin, where foods had higher contributions to total intakes than beverages in all three countries. In the refined Tiers 3 (Figure 4) and 4 (Figure 5) assessments, beverages contributed at least 85% of total intake for acesulfame potassium and aspartame, cyclamate (as sodium cyclamate), saccharin and stevia in Argentina. The same was true for Chile except for stevia where beverage contribution was lower. For Peru, beverages also contributed at least 85% to the total intake of acesulfame potassium and aspartame, but their contribution to the total intake of cyclamate (as sodium cyclamate), saccharin and stevia were lower. For sucralose, beverages contributed 36% to 68% of total intake, depending on the country and exposure scenario.

Figure 2. Percent contribution of beverages and foods to total estimated intakes – Tier 1

Figure 3. Percent contribution of beverages and foods to total estimated intakes – Tier 2

Figure 4. Percent contribution of beverages and foods to total estimated intakes – Tier 3

Figure 5. Percent contribution of beverages and foods to total estimated intakes – Tier 4

Comparison of estimated LNCS intakes vs. the JECFA ADI

The estimated intakes for each sweetener were compared to their respective JECFA ADI (WHO 2017, 2019). Figure 6 summarises the estimated intakes presented in Tables 6–8 for the six sweeteners as a percent of their respective JECFA ADIs. The estimated intakes of each of the six sweeteners are well below their respective JECFA ADI even for Tier 1, the most conservative assessment scenario. Estimated intake of the six sweeteners as a percent of their respective JECFA ADI ranged from 21% (for aspartame and sucralose in Peru) to 95% (for cyclamate in Argentina) for Tier 1. For the still conservative Tier 2 assessment that used average LNCS concentrations but assumed all the sweeteners are present in all LNCS containing foods and beverages, the estimated intakes ranged from 11% (for sucralose in Peru) to 60% (for cyclamate in Argentina) of their respective JECFA ADI. Estimated intakes in Tiers 3 and 4 were all below 15% and 10% respectively of their JECFA ADI.

Figure 6. Estimated intake of sweeteners as percent of the associated JECFA ADI. (a) Acesulfame potassium (ADI: 15 mg/kg bw/day). (b) Aspartame (ADI: 40 mg/kg bw/day). (c) Cyclamate (ADI: 11 mg/kg bw/day). (d) Saccharin (ADI: 5 mg/kg bw/day). (e) Steviol glycosides (ADI: 4 mg/kg bw/day). (f) Sucralose (ADI: 15 mg/kg bw/day)

Discussion

The current study derived screening level intake estimates of six LNCS using the Budget Method. The default consumption amounts used in the method, namely 0.1 L/kg bw/day for beverages and 0.05 kg/kg bw/day for solid foods, are conservative. The default consumption amount for beverages translates to 6 L of non-milk beverages for a 60 kg person. Reviews of the method by Douglass et al. (1997) and Bär and Würtzen (1990, as cited by Douglass et al, 1997) indicate that the 0.1 L/kg bw/day is too conservative and suggested that half that amount (or 50 mL/kg bw/day) may be appropriate. The default food consumption amount translates to 3 kg of food for a 60 kg person, which is a conservative estimate, since consumption data from the US National Health and Nutrition Examination Survey indicate that the total daily intake of both foods and beverages is about 3 kg (data not shown). Hence, the intake estimates derived in this study are conservative, even in the case of the refined Tiers 3 and 4 assessments. Further, since the default consumption amounts used in the Budget Method were derived by Hansen (Hansen 1966, 1979) using conservative assumptions on the caloric needs of young children and on the range of potential beverage intakes of infants, children, and adults in hot climates, the estimated intakes derived in the current study are not only applicable to the general population, but also applicable to the subpopulation of children.

Argentina

Garavaglia et al. (2018) reported on results of a survey conducted in 2011 to estimate intake of five LNCS (acesulfame potassium, aspartame, cyclamate, saccharin, and sucralose) among school children in Buenos Aires, Argentina. While the percent consumers of cyclamate was the lowest as compared to the other sweeteners, ranging from 3% to 8%, depending on the age group, the estimated intakes of cyclamate were typically higher than those for the other sweeteners. With the exception of cyclamate, Tier 3 and 4 intake estimates derived in this study are in the range of the 50th to 75th percentile intakes derived by Garavaglia et al. (2018). In the case of cyclamate, Tiers 1 and 2 estimates are more in line with the estimates from Garavaglia et al. (2018). Similar to the findings in the current study, beverages were also the major contributors to the total estimated intakes derived by Garavaglia et al. (2018).

Chile

A law recently enacted in Chile requires a warning on the front package if the food or beverage product exceeds pre-set limits on calories, sugars, saturated fats, and sodium has led food and beverage producers to reformulate their products and, in particular, to replace sugars with LNCS, thereby lowering the calorie and sugar content of their products. Several studies have evaluated LNCS intakes pre- and post-law enactment (Venegas-Hargous et al. 2020; Martínez et al. 2020). Martínez et al. (2020) reported estimated intakes of LNCS for school children in Chile. The 25th, 50th and 75th percentile estimated intakes of acesulfame potassium reported by Martínez et al. (2020) were 0.25 mg/kg bw/day, 0.88 mg/kg bw/day and 2.10 mg/kg bw/day, respectively. In the current study, the estimated intakes of acesulfame potassium from Tiers 3 and 4 were 1.90 mg/kg bw/day and 1.42 mg/kg bw/day, respectively, which overlap with the estimates reported by Martínez et al. (2020), albeit they are in the upper range, as would be expected given the conservative assumptions inherent in the Budget Method. The maximum intake reported by Martínez et al. (2020) for acesulfame potassium was 7.60 mg/kg bw/day, slightly higher than the Tier 1 estimate of 5.32 mg/kg bw/day derived in this study.

Similar comparisons of the estimates reported by Martínez et al. (2020) and the estimates derived using the Budget Method for the other sweeteners show similar concordance, except for cyclamate and saccharin, with the maximum estimated intake derived by Martínez et al. (2020) and colleagues were at or above the estimates obtained in Tier 1, with estimates in Tiers 3 and 4 overlapping with estimates within the 50th to 75th percentiles range. In the case of cyclamate and saccharin, the Tier 1 and 2 estimates were higher than the maximum intakes reported by Martínez et al. (2020). This difference is driven by the assignment of cyclamate and saccharin concentrations to beverages based on data reported in Argentina and the assumption used in those two assessment tiers that 100% LNCS-containing beverages contain cyclamate and saccharin, a conservative assumption that is contradicted by the data from Mintel GNPD and the Martínez et al. (2020) study that found that only 2% of the LNCS products contained these sweeteners.

The contribution of beverage products to the total intake of acesulfame potassium (94%), aspartame (96%), and steviol glycosides (36%), and sucralose (48%) reported by Martínez et al. (2020) are also consistent with the contribution of beverages to the estimated total intakes derived using the Budget Method and which ranged from 77% to 98% for acesulfame potassium, from 82% to 99.9% for aspartame, from 18% to 68% for steviol glycosides and 36% to 73% for sucralose (see Figures 2–5). Larger differences were observed between the Martínez et al. (2020) estimates and the estimates derived using the Budget Method for cyclamate and saccharin, most likely driven by the assignment of concentrations based on Argentina.

A recent study of LNCS intakes among university students (Duran Aguero et al. 2017) shows a high prevalence of weekly consumption of acesulfame potassium, aspartame, and sucralose (ranging from around 85% to 92%) among the 472 participants from Chile, with lower prevalence of cyclamate and saccharin (23% and 26%, respectively), and relatively low prevalence of steviol glycosides (2%).

Acesulfame potassium, aspartame, and sucralose estimated intakes from the same study (Duran Aguero et al. 2017) were 0.64 mg/kg bw/day, 1.96 mg/kg bw/day, and 1.38 mg/kg bw/day, respectively among male students, and 0.53 mg/kg bw/day, 1.45 mg/kg bw/day, 0.67 mg/kg bw/day, respectively for female students (Duran Aguero et al. 2015). These estimates are below the estimated intakes derived in the current assessment for acesulfame potassium and aspartame, and in the range of the estimates obtained from the refined Tiers 3 and 4. These differences reflect the conservative assumptions used in the current study.

Peru

All estimated intakes were below their respective ADI. However, as discussed above, the concentration data for Peru in the Mintel GNPD were limited, foods and beverages were assigned concentration data based on the other two countries (except for cyclamate and saccharin where only data from Argentina were available). Data from the study of LNCS intakes among university students (Duran Aguero et al. 2017) showed a high prevalence of weekly consumption of acesulfame potassium, aspartame, and sucralose (ranging from around 67% to 79%) among the 204 participants from Peru, and relatively low prevalence of steviol glycosides (9%). None of the students reported consumption of saccharin or cyclamate.

Acesulfame potassium, aspartame, and sucralose-estimated intakes from the same study were 0.72 mg/kg bw/day, 2.13 mg/kg bw/day, and 0.88 mg/kg bw/day, respectively among male students, and 0.82 mg/kg bw/day, 2.12 mg/kg bw/day, 0.73 mg/kg bw/day, respectively for female students (Duran Aguero et al. 2015). These estimates are in the range of intakes estimated for Tiers 3 and 4 for acesulfame potassium and Tiers 2 and 3 for aspartame and sucralose. These differences reflect the potential uncertainty in the default assignment of surrogate concentrations (based on the Argentina and Chile data).

Impact of the uncertainty in the assessment

The four tiered assessments illustrate the impact of the potential uncertainty in the LNCS concentrations and the fraction of the foods and beverages that use the particular LNCS. Tiers 1 and 2 make the same assumption regarding the fraction of food and beverages that use the specific sweetener, but differ with respect to the concentration assigned; the same is true for Tiers 3 and 4. On the other hand, Tiers 1 and 3 make the same assumption regarding the concentration levels but differ with respect to the fraction of food and beverages that use LNCS; the same is true for tiers 2 and 4. A comparison of the estimates summarised in Tables 6–8 shows that smaller relative differences are observed between the estimates derived in Tiers 1 and 2 (and Tiers 3 and 4) as compared to the relative differences observed between Tiers 1 and 3 (and Tiers 2 and 4). That is, the uncertainty in the estimate of the fraction of food or beverages that use LNCS has a bigger impact on the estimated intakes than the potential uncertainty in LNCS concentrations. Hence, it may be more beneficial to focus resources on refining the estimates on the fraction of food and beverages that use LNCS than obtaining more refined estimates of LNCS concentrations.

Conclusion

Strengths of the current study include the use of a model diet and country-specific food and beverage sales data, and country and product-specific labels to estimate LNCS intake when actual consumption data from national surveys are not available.

The estimated LNCS intakes were below their respective JECFA ADIs, even for Tier 1 assessments that used the most conservative assumptions. Further, the estimated intakes for the refined Tiers 3 and 4 were generally in the range of those observed in studies conducted in the same countries on select subpopulations, further validating the results of these assessments and the methodology used to derive the intake estimates.

The default consumption values used in the Budget Method are derived using conservative assumptions on the caloric needs of young children and on the range of potential beverage intakes of infants, children, and adults in hot climates, hence, the estimated intakes derived in the current study for Argentina, Chile, and Peru are not only applicable to the general consumer population, but also applicable to the subpopulation of children in these countries.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by the International Sweeteners Association.