Prevalence of food colors use in local and imported pre-packaged foods in Hong Kong

Food colors are food additives that are used to add or change the colors of food and beverages (European Food Safety Authority, 2022). The use of food colors in food and beverages is tightly regulated in most jurisdictions (Stich, 2016; Harp and Barrows, 2015; Lehto et al., 2017; Magnuson et al., 2013). Acceptable daily intakes (ADI) for each of the permitted food colors are based on the recommendations of the Joint FAO/WHO Expert Committee on Food Additives (Food and Agriculture Organization of the United Nations, 2021), and category-specific usage limits are determined according to the dietary exposure assessment conducted by the respective regulatory agencies (Dixit et al., 2013; Bemrah et al., 2012; Lim et al., 2019; Vin et al., 2014) to minimize the risk of exceeding the ADI, particularly among high consumers.

While intakes below the ADI are generally considered to pose no significant risk to health according to the best available evidence (Renwick and Walker, 1993; Herrman and Younes, 1999), studies have suggested some artificial food colors may have adverse effects, such as increasing the risk of hyperactivity in preschool children (Bateman et al., 2004; McCann et al., 2007). Because of that, consumer awareness of food color use has increased (Shim et al., 2011; Bearth et al., 2014; Burrows, 2009). In most countries, it is a legal requirement to list the types of food color used in pre-packaged products, either by listing the chemical name (e.g. tartrazine) or respective E numbers (e.g. E102 for tartrazine) to assist consumers in making informed food choices. However, there are concerns that consumer understanding of food labels may be suboptimal (Shim et al., 2011; Zhong et al., 2018; Altu and Elmaci, 1995), with studies showing consumers tended to avoid products with E numbers (Bloeming, 2020; Paans, 2013). Together, these may impact their ability to make informed choices about food color.

Studies from around the world have reported that food colors are widely used in pre-packaged foods, particularly in snacks and confectionery (Diouf et al., 2014; Tennant, 2008; AbuKhader et al., 2021). In Hong Kong, Lok et al. (2010, 2011) also reported that several synthetic food colors, including tartrazine (E102) and sunset yellow FCF (E110), are widely consumed by primary school children. Nonetheless, none of these studies provides a comprehensive picture of the overall usage of different types of food colors in various food categories. Since Hong Kong relies heavily on imported food, which accounts for around 90% of the total food supply (Environment and Ecology Bureau (Hong Kong), 2010), the pre-packaged food supply in Hong Kong could provide insights into food additive use across products from different regions.

Therefore, we aimed to examine the use of food colors in a representative sample (by market share) of pre-packaged foods available for sale in Hong Kong. We examined the difference in the prevalence of food color use between food categories and products from different regions.

The 2019 cross-sectional FoodSwitch Hong Kong database was used in this study. FoodSwitch Hong Kong was a project that conducted annual data collection between 2017 and 2019 using the same methodology that was previously described in detail (Wong et al., 2020). In brief, the non-probabilistic sampling method was used to collect pre-packaged food items in five major chain supermarkets in Hong Kong, including the supermarkets which mainly sell local and imported groceries (Park’n Shop and Wellcome), targeted imported products from specific origins, including Japan and the UK respectively (AEON and Marks & Spencer) and global imported products (City’Super). More than 70% of the market share within the supermarket category is captured by these supermarket chains, with the rest contributed by smaller retailers, online stores and others (Euromonitor International, 2019; Li, 2018). Trained research assistants collect the data from these supermarkets. Pictures of the product packages were taken, and a full dataset including information representing the products (e.g. barcode, brand name and product name), ingredient list, nutrition information and country of origin were compiled. According to the Global Trade Item Number (GTIN) standards on barcodes (GS1, 2020), the countries of origin were identified by the prefix containing 2 or 3 characters, e.g. “489” for Hong Kong, “000-013” for USA and “754-755” for Canada. The declaration of country of origin, as stated on the pack, was used if no GTIN-standard barcode was shown on the package. The countries of origin were then grouped into five regions: Asia, Europe, North America, Oceania and others (including Africa, Middle East and South America).

The pre-packaged food products (n = 21,122) were categorized into 18 major (e.g. bread and bakery products) and 60 respective minor groups (e.g. biscuits and bread) according to the categorization system developed by The Global Food Monitoring Group (Dunford and Neal, 2017). Products with missing or incomplete ingredient lists (n = 113) or that are pack-size duplicates (n = 554) were excluded. Furthermore, products in the following food categories, “alcohols” (n = 27), “vitamins and supplements” (n = 5) and “uncategorized foods” (n = 41) were also excluded as they are not required to carry the standard nutrition information panel of Hong Kong, and/or are not expected to be a major dietary contributor of food color. The remaining food items (n = 20,382) from 15 major food categories were included in the main descriptive analysis.

A total of 35 types of common food color (Supplementary Table 1) were examined in this study, which were classified as either “synthetic” or “natural” as the former were more commonly associated with adverse health effects. Under the Food and Drug (Composition and Labeling) (Amendment) Regulation 2004, the Hong Kong Centre for Food Safety (CFS), who regulate the usage of food additives in foods, uses the International Numbering System (INS), which includes subgroups (e.g. 160a-160f), for food labeling (Government of the Hong Kong SAR, 2017). The INS system generally corresponds to the E-number, although it is not specified in the regulation. All imported foods from around the world are subject to the same regulations. The official and alternative names of these food color and their respective E-codes were used as keywords for the identification in the ingredient lists. In cases where a food color can also serve multiple functions, e.g. E170 can serve as a food color, anti-caking agent or acidity regulator, a researcher (SHYY) examined the full ingredient list to confirm its specific usage and only those products which used E170 specifically as a food color was considered to contain E170 in this study. Terms in the ingredients list, such as “food colorant” without specifying the exact type being used, were not coded.

Data were analyzed using SPSS (version 28.0, IBM Corp. Ltd, New York, US). Descriptive statistics were computed for the prevalence (%) of the use of food colors (total, natural and synthetic) and the number of food colors used in the included food items, stratified by food category and region of origin where appropriate. Pearson’s χ² test was used to compare the total number of food colors used in food items from different regions. We did not correct for multiple comparisons as this was an exploratory study, and a two-tailed p < 0.05 was considered statistically significant.

The prevalence of food color use in the audited sample was 19.8% overall (n = 4,044) and ranged from 1.4 to 37.5% across the major food categories (Figure 1). The use of natural food color was more than four times as prevalent than that of synthetic food color (17.2 vs 3.9%).“Snack foods” had the greatest proportion of food items containing at least one food color (37.5%), followed by “confectionery” (34.7%) and “convenience foods” (33.3%). In terms of the number of food colors used, among products with food colors, most used one to two types of food colors (89.5%). “Confectionery” had the highest proportion of items using three or more food color (10.7%), followed by “snack foods” (3.4%). Only five categories, namely “special foods” (9.4%), “edible oils and oil emulsions” (7.3%), “fruit and vegetables” (5.2%), “sugar, honey and related products” (2.4%), and “eggs” (1.4%) had a prevalence of food color use below 10%. Minor food categories that had the highest prevalence of food color use included “dessert additions” (77.8%), ” other frozen foods not otherwise specified“(66.7%), and “chewing gum” (64.9%). Notably, 0.9% of baby foods were found to contain at least one food color (Online Supplemental Figure 1).

Of all the food colors examined, E160 (carotenoids) was the most frequently used (8.4% of all foods, 42.4% of foods with food color), followed by E150 (caramel; 7.4% of all foods, 37.4% of foods with food color) and E102 (tartrazine; 2.1% of all foods, 10.8% of foods with food color) (Figure 2). Four out of the top 10 food colors used [E102 (tartrazine), E110 (sunset yellow FCF), E129 (Allura red AC) and E133 (brilliant blue FCF)] were synthetic. “Confectionery” (13.5%) and “snack foods” (30.1%) had the highest prevalence of synthetic and natural color use, respectively.

The use of different food colors in food items from different categories was summarized in Tables 1 and 2. E160 (carotenoids) was the top food color used in eight categories, while E150 (caramel – unspecified) was the most frequently used food color in five categories. The use of E133 (brilliant blue FCF; 9.7%) and E163 (anthocyanins; 9.5%) was high among “confectionery”. Large variations in the types of food color used were also observed within food categories (Online Supplemental Tables 2 and 3). Food colors not permitted in Hong Kong (E105, Fast Yellow AB; E107, Yellow 2G; E125, ponceau SX; and E156, carbon black) were not found in any of the audited food items.

The prevalence of use and the total number of food colors used varied between the region of origin (both p_χ2 < 0.001). Products from Asia had the highest prevalence of food color use (26.3%; Figure 3), followed by products from Oceania (17.5%) and North America (16.4%). The use of three or more types of food color was particularly prevalent in products from North America (2.5%), Asia (2.4%) and Oceania (2.4%). The top food color used among pre-packaged foods of different regions of origin also varied (Online Supplementary Figure 2).

In this first comprehensive audit of >20,000 local and imported pre-packaged food and beverages available in Hong Kong, we found that ∼20% of items contained at least one food color. Snack foods, confectionery and convenience foods were most likely to use food colors, and E160 (carotenoids), E150 (caramels) and E102 (tartrazine) were the most frequently used. Asian products were also found to be more likely to use food colors, and the use of natural food color was more prevalent than synthetic food color.

The high prevalence of food color use in snack foods, confectionery, and convenience foods is unsurprising as they are highly processed (Omena Messias et al., 2016). These three food categories are also among the top categories which utilized more than two food colors, with ∼8% of confectionery found to contain four or more food colors. Although the safety of food colors has been well-established (Bastaki et al., 2017), underpinning their regulatory approval for use in foods, such safety assessments were commonly conducted based on their individual use (Lehto et al., 2017). This is a concern given the effect and safety of concurrent use of different food colors in the same products remains largely unknown (Chazelas et al., 2020). A randomized, double-blinded, placebo-controlled crossover trial in 153 young children aged 3 and 8 showed a potential association between artificial colors and/or sodium benzoate (E211) and increased hyperactivity. These colors included E129 (Allura red), E102 (tartrazine), E110 (sunset yellow FCF), E122 (carmoisine), E124 (ponceau 4 R), and E104 (quinoline yellow) (McCann et al., 2007). Of concern, some of these food colors in question were among the top 10 most commonly used in pre-packaged foods in Hong Kong identified in our study. Although the EFSA (Aguilar et al., 2008) and US FDA (Cheeseman, 2012) decided no action is warranted based on the finding from the Southampton study, it highlights the need for future research to understand the health impact of different combinations of different food colors, particularly the common mixes found in the food supply.

Furthermore, it is surprising and concerning that ∼20% of dairy products surveyed also contained at least one food color. Around 7% of milk (n = 33), mostly flavored milk, was found to contain colors including E101 (riboflavin), E102 (tartrazine), E133 (brilliant blue), E150 (caramels), and E160 (carotenoids). In Hong Kong, food colors are considered natural if they come from fruits or vegetables or are synthetic analogues of the same color from fruits or vegetables (Government of the Hong Kong SAR, 2020). For example, E160a (carotenes) from natural extracts and synthetic methods are considered natural food colors. Some yoghurts also contained E102 (tartrazine), although natural food colors such as E160 (carotenoids) and E163 (anthocyanins) are more commonly used. While E101 (riboflavin) and E160 (carotenoids) are unlikely to be harmful, there may be adverse effects associated with chronic exposure to E102 (tartrazine) and E133 (brilliant blue), including genotoxicity, carcinogenicity and hyperactivity (Kobylewski and Jacobson, 2010). Among baby food products, baby fruit bites were found to contain food color more commonly, although they were largely natural colors, e.g. E160 (carotenoids) and E163 (anthocyanins).

The overall prevalence of specific types of food color was similar to that reported by Gilsenan et al. in the UK (Gilsenan et al., 2002). Driven by consumers’ demand for “clean label” (Brewster, 2021), as well as manufacturer’s deliberate attempts to use natural food color (commonly less stringently regulated) to prevent the need to create separate formulations for the export market with a different list of approved food colors (Lehto et al., 2017; Stich, 2016), it is not surprising to see that six out of the top 10 food colors used are natural. However, it is still concerning that the artificial food colors E102 (tartrazine), E110 (sunset yellow FCF) and E133 (brilliant blue FCF/FD&C blue 1), which are made from coal tar or petroleum (Kobylewski and Jacobson, 2010), are among the top 10 food colors used. Consumption below the acceptable daily intake (ADI) of these food colors is considered with reasonable certainty to be harmless, according to the U.S. Food and Drug Administration (FDA) (U.S. Food and Drug Administration, n.d.). However, a study showed changes in biochemical markers in the brain even below the ADI level of E102 at 7.5 mg/kg body weight/day (Bhatt et al., 2018). Thus, there is still a need to investigate the potential long-term toxicological effects of exposure to these synthetic colors, particularly given their high prevalence in the food supply.

Some natural food color may also pose harmful health effects on humans. Concerns regarding the carcinogenicity of E150 (caramel), which was the second most-used color (7.4%), have been raised (Hengel and Shibamoto, 2013; Smith et al., 2015; Jacobson, 2012). Caramel colors are produced from caramelization that may result in the formation of low molecular weight constituents such as 5-hydroxymethylfurfural (HMF), 4-methyllmidazole (4-Mei) and 2-acetyl-4(5)-(1,2,3,4-tetrahydroxybutyl)imidazole (THI) (EFSA Panel on Food Additives and Nutrient Sources added to Food, 2011), which were suggested to be carcinogenic (Hengel and Shibamoto, 2013; Smith et al., 2015; Jacobson, 2012). However, the EFSA concluded the risks related to exposure to these caramelization side products from the use of caramel color in a typical diet are negligible (EFSA Panel on Food Additives and Nutrient Sources added to Food, 2011).

The high variability in the number and types of food colors used between products from different regions could be driven by consumer demand and preference, as well as differences in regulations related to food color used (Lehto et al., 2017; Magnuson et al., 2013; McAvoy, 2014). For example, the U.S. FDA defines that all additives added to foods for the purpose of changing the color of a food must be considered a food color, and approved as such (McAvoy, 2014). In contrast, the EU allows additives added to food for other purposes (e.g. flavoring), but with a secondary effect of changing the color of the food (e.g. paprika oleoresin), to be exempted from being treated as a food additive (McAvoy, 2014). In Asia, China (one of the main countries that Hong Kong imports its pre-packaged foods from) requires that each color be assessed and approved for its use (and permitted levels) on an individual basis (McAvoy, 2014), while Japan maintains a prescribed list of allowed food color and other additives (McAvoy, 2014).

Currently, consumers can utilize the information provided on the ingredient lists to identify the presence of specific forms of food colors. However, consumers’ understanding of the E-number or the name of food colors may be suboptimal (van Gunst and Roodenburg, 2019). Further to that, we also observed that some food labels only included wordings such as “food colorant”, “food colors”, or “food colorings” without specifying the types being used. The lack of understanding and ambiguity of food color labeling will likely limit the ability of consumers to make informed choices, which is a concern for consumers with allergies or intolerances to food additives (Bahna and Burkhardt, 2018).

The main strength of this study is that it is the first study in Hong Kong investigating the overall use of food color in pre-packaged foods. While several studies had examined the dietary exposure to certain types of food colors such as E102 (tartrazine) and E110 (sunset yellow FCF), none of the studies analyzing the food supply in Hong Kong provided a comprehensive overview of all food color used. Moreover, we also examined the difference in food color use by region of origin, which is seldom reported in the literature. Lastly, where a food additive can serve multiple purposes, including as a food color, we have examined the full ingredients lists to ensure that the correct usage is captured.

However, this study has several limitations. First, while the ingredient lists entered into the FoodSwitch Hong Kong database were randomly audited for accuracy by the data manager, there might be missed data entry errors, especially for products with many ingredients. Another limitation is that the data used were collected in 2019. Due to the discontinuation of products and the emergence of new products on the markets, the results may not reflect the current Hong Kong food supply. However, there is often a time lag between data collection and analysis. Considering this, the 2019 Hong Kong FoodSwitch dataset represents the most recent comprehensive dataset with information about the ingredients of the pre-packaged foods sold in Hong Kong. Thus, the results of this study could provide baseline data to enable time-trend comparison studies in the future. This also highlights the importance of regular monitoring and surveillance of food color use to keep up to date with the current food supply and track the trend in food color use over time. Finally, we did not provide the estimated daily intake of food colors in Hong Kong. In two local studies, the intake of food colors among primary school children varies among different districts in Hong Kong. While the average daily intake of sunset yellow FCF (E110) among primary school children aged 8–9 years was 51% higher than the ADI threshold in some districts, the dietary exposure of the 11 synthetic food colors was considerably lower than the FAO/WHO and ESFA acceptable daily intakes in the local area of Shatin (Yuet-Wan Lok et al., 2010; Lok et al., 2011; Lok et al., 2010). Large cross-sectional studies on food additives consumption from a representative local population are expected to estimate the average daily intake of food colors. A standardized food frequency questionnaire (FFQ) specific to the local community could be used to capture the usual intake of pre-packaged foods.

The use of food colors was prevalent in a wide range of pre-packaged foods available in Hong Kong, with the highest prevalence found in “snack foods”, “confectionery”, and “convenience foods”. Four of the top 10 food colors used were synthetic and may be associated with adverse health effects. Given their prevalent use in the food supply, the risk assessment of their exposure to a typical diet may need to be updated. Regular monitoring of their use should also be conducted to enable longitudinal tracking.

The supplementary material for this article can be found online.