Trans fatty acids are found in industrially produced partially hydrogenated vegetable oils (PHVO), as well as in small quantities in fats from ruminants(Reference Oteng and Kersten1). Compared with non-hydrogenated vegetable oils, PHVOs have increased stability and shelf life, as well as a flavour profile that is more similar as SFA, at a fraction of costs of the latter(Reference Skeaff2). Hence until recently, PHVOs were a popular choice of ingredient among food industry to manufacture foods with desired textures and flavours at a lower cost(Reference Uauy, Aro and Clarke3).
Trans-fat, particularly those industrially produced, had been consistently linked to negative health effects(Reference Uauy, Aro and Clarke3), including at least in part higher CVD risks(Reference Oteng and Kersten1,Reference Lichtenstein4–Reference Bjoernsbo, Joensen and Joergensen7) . This is because of the harmful effect of trans-fat on cholesterol metabolism – simultaneously increasing the level of LDL cholesterol and decreasing the level of HDL cholesterol(Reference Yanai, Katsuyama and Hamasaki8,Reference Takeuchi and Sugano9) . Trans-fat intake has also been shown to be positively associated with all-cause mortality, risks of cognitive disorder and inflammation(Reference Oteng and Kersten1,Reference de Souza, Mente and Maroleanu10,Reference Barnard, Bunner and Agarwal11) . As a result, government and public health agencies around the world have introduced plans and legislations to reduce the population’s trans-fat intake. For example, in May 2018, the WHO proposed the ‘REPLACE trans-fat’ action plan, with a goal to remove industrial trans-fat ingredients from the global food supply chain by 2023(Reference Ghebreyesus and Frieden12). The USA has completely banned the use of PHVO in pre-packaged foods in 2018(Reference Bloks13). In 2019, the European Commission imposed a trans-fat limit (for industrially produced trans-fat only) of ≤ 2 % of total fat (%total fat) to in pre-packaged food, with full enforcement starting in 2021(Reference Kušar, Hribar and Lavriša14,Reference Junker15) . Many countries around the world, however, still do not have plans in place to eliminate industrial trans-fat from the food supply – particularly many Asian and African countries(16).
While food retailers, importers and manufacturers in Hong Kong need to comply with the Nutritional Label Scheme which mandated the labelling of total trans-fat (i.e. both industrially produced and ruminant trans-fat) on the nutrition information panel(17), the accuracy of such labeling is seldom studied. In May 2021, 68·4 % of cheeses sampled in Hong Kong were found to have inaccurate trans-fat labelling(18). Together with ambiguous names used in the ingredients list, such as ‘vegetable fat’ without specifying whether the fat was hydrogenated; or ‘hydrogenated vegetable oil’ without specifying the degree of hydrogenation, it may be difficult for the general public to correctly identify foods containing trans-fat by reading food labels. In fact, two recent Brazilian studies suggested that ingredients with ambiguous names, such as ‘vegetable fat’, were found in foods claimed to have no trans-fat(Reference Silveira, Gonzalez-Chica and da Costa Proenca19,Reference Ricardo, Peroseni and Mais20) . Given Hong Kong imports most of its pre-packaged food supply, the same issue may also be present in Hong Kong.
Therefore, this cross-sectional study aimed to describe the proportion of pre-packaged foods: (1) exceeding the EU industrially produced trans-fat limit of ≤ 2 %total fat (Reference Junker15); (2) labelled to contain specific and non-specific ingredient(s) indicative of the possible presence of trans-fat and (3) labelled as trans-fat free but contains specific ingredient(s) indicative of the presence of trans-fat. We also examined differences between food categories and region of origin.
Methods
Data source and collection
FoodSwitch Hong Kong is a project that collected data of pre-packaged foods available for sale in Hong Kong annually between 2017 and 2019(Reference Wong, Coyle and Wu21). The 2019 version of the FoodSwitch Hong Kong database(Reference O, Coyle and Dunford22) was used in this study. Details of the data collection protocol were described elsewhere(Reference Wong, Coyle and Wu21). In brief, data were collected from one megastore each, all from affluent areas, of City!Super (selling mainly imported products), Marks and Spencer (selling a wide range of its home brand products originating from the UK), AEON (selling mainly products imported from Japan), as well as Park’nShop and Wellcome (both selling numerous imported and typical local branded products). Altogether these supermarket chains account for more than 70 % of the market share of pre-packaged foods(Reference Li23). The data were collected by trained research assistants who visited the stores, who took photographs of each product including the nutrition information panel, front-of-package and barcode, using a bespoke smartphone application(Reference Dunford and Neal24).
Data entry and pre-processing
For each food and beverage, data recorded in the 2019 FoodSwitch Hong Kong database include the brand name, product name, barcode, as well as content of total energy, protein, total fat, saturated fat, trans-fat, total carbohydrate, total sugar and sodium listed. If the trans-fat content per serving or package was displayed on the label, it was converted to the trans-fat concentration per 100 g or per 100 ml of the food. It should be noted that the nutrition labelling of Hong Kong does not require the separate reporting of industrially produced v. ruminant trans-fat. The countries of origin were identified using the 3-digit prefix of the Global Trade Item Number standard barcode(25), which identifies the issuance country, as a proxy if that corresponds to a single country, e.g. ‘489’ for Hong Kong. For items which have Global Trade Item Number prefixes that correspond to more than one issuance country, or items with non-Global Trade Item Number-standard barcodes, the on-pack declaration of country of origin was used. The countries of origin were then grouped into four regions, namely Asia, Europe, North America and others (including Australia, New Zealand, South America, the Middle East and Africa)(Reference O, Coyle and Dunford22).
Trans-fat ingredient ‘indicators’
Trans-fat ingredient ‘indicators’ were ingredient terms used to determine whether food items labelled as containing 0 g trans-fat may potentially contain industrially produced trans-fat. Specific trans-fat ingredient indicators referred to ingredients known to contain industrially produced trans-fat, presence of which indicate ‘false negatives’ for products that were labelled as having zero trans-fat on the nutritional label. Non-specific trans-fat ingredient indicators referred to ingredients that may or may not contain industrially produced trans-fat. A full list of specific (n 31) and non-specific (n 113) trans-fat ingredient indicators is provided as Supplemental Table 1. Of note is that ‘vegetable oils/fats’ without specification of hydrogenation level were considered as non-specific trans-fat ingredient indicators in foods that are likely to have industrially produced trans-fat, including biscuits, cake and pastries, salad dressings, snack foods and chocolates(26).
Product categorisation and exclusion criteria
Categorisation of food products was based on the food classification system developed by the George Institute for Global Health(Reference Dunford and Neal24), which categorises products into eighteen major food groups. The food categories ‘alcohols’ (n 6), ‘eggs’ (n 67), ‘special foods’ (n 335), ‘vitamins and supplements’ (n 5) and ‘unable to be classified’ (n 23) were excluded as they are unlikely to contain trans-fat or not considered a major source of trans-fat in the local diet, leaving thirteen major food groups for final analysis. If the same item is sold in multiple stores (as identified by the same GTIN barcode), only one entry of that item was included in the database. For the same item in different package sizes (e.g. cola soft drink in 600 ml bottle v. 6 × 330 ml cans), they were manually identified by a researcher (CCWC), based on the brand name, product name and nutrition information, and only one entry was included in the data set for statistical analysis. To better align with the EU trans-fat limit which refers only to industrially produced trans-fat, the remaining items which reported to contain trans-fat but have no trans-fat ingredient indicator in the ingredients list (n 1463), suggesting the trans-fat is likely of ruminant origin, were excluded from the relevant analyses.
Statistical analysis
Data analyses were conducted using SPSS (version 26; IBM Corp.). The median and interquartile range of trans-fat content per 100 g of food and as %total fat, stratified by food categories were computed. The proportion of trans-fat containing items which has an industrially produced trans-fat content above the EU limit of ≤2 %total fat (Reference Junker15) was also calculated. The number of specific, non-specific and total trans-fat ingredients indicators for each item was counted. Results were stratified by major food categories, and region of origin was applicable. Pearson’s χ 2 was used to test for the differences between the proportion of items with different number of trans fat indicator in the ingredients list from different regions. The differences in the mean ± sd trans-fat content (%total fat) based on the information in nutrition label of items from different regions were tested using one-way ANOVA with Tukey’s correction for multiple comparison. A two-tailed P < 0·05 was considered statistically significant.
Results
Of the 21 122 records in the 2019 FoodSwitch Hong Kong database, 1656 products were excluded to have incomplete nutritional information panels without trans-fat values, and 439 were excluded due to not being a main source of trans-fat in the diet of the Hong Kong population (Fig. 1). Among the remaining 19 027 items, 2192 (11·5 %) reported a trans-fat content greater than 0, and 729 of these also contained at least 1 trans-fat ingredient indicator in the ingredients list, suggesting the trans-fat contained in them was likely to be industrially produced (Table 1). The median trans-fat content of these 729 items was 0·3 g/100 g or 100 ml (interquartile range (IQR): 0·1–0·6) and 1·2 %totalfat (IQR: 0·6–2·9). Of these, items in ‘edible oils and oil emulsions’ (median = 1·2, IQR = 0·6–1·4), ‘sauces, dressings, spreads and dips’ (median = 0·6, IQR = 0·3–1·0) and ‘non-alcoholic beverages’ (median = 0·5, IQR = 0·3–0·6) had the highest trans-fat concentration (g per 100 g or 100 ml), while ‘dairy’ (median = 3·3, IQR = 2·1–4·4), ‘non-alcoholic beverages’ (median = 3·1, IQR = 1·6–4·7) and ‘meat and meat products’ (median = 2·9, IQR: 2·3–3·6) had the highest proportion of fat as trans-fat (%totalfat). In contrast, ‘cereal and grain products’ had the lowest median trans-fat concentration (median = 0·1, IQR = 0·1–0·3), while ‘fruit and vegetables’ (median = 0·7, IQR = 0·5–1·3) and ‘snack foods’ (median = 0·6, IQR = 0·4–1·3) had the lowest proportion of fat as trans-fat. ‘Bread and bakery products’ had highest proportion of items with industrially produced trans-fat (18·9 %). ‘Snack foods’ had 9·7 % and ‘confectionery’ 6·6 %. Overall, the EU trans-fat limit of ≤ 2 %totalfat was exceeded by 1·3 % of the included products, and ‘bread and bakery products’ (6·8 %) and ‘snack foods’ (1·7 %) had the highest proportion of items exceeding the EU limit.
Fig. 1 Flow of data preparation and analysis. Black boxes represent exclusion, and grey boxes represent statistical analyses
Table 1 Trans-fat content reported on nutrition labels of the sampled pre-packaged foods
* Trans-fat that was likely to be industrially produced, based on the presence of trans-fat ingredient indicators in the ingredients list.
† Only for items reported to have > 0 g/100 g or 100 ml trans-fat with 1 or more trans-fat ingredient indicators in the ingredients list.
Specific and non-specific trans-fat ingredient indicators were found in 234 (1·2 %) and 3298 (17·3 %) of the 19 027 included items, respectively. For items labelled as having 0 g trans-fat (n 16 835), 14 173 (84·2 %) were found to have no specific or non-specific trans-fat ingredients indicators on the ingredient list, suggesting the trans-fat labelling was correct. Of the remaining 2657 items, 200 (7·5 %) were found to have at least 1 specific trans-fat indicator in the ingredients list (Fig. 2). ‘Non-alcoholic beverages’ was found to have the highest proportion of products of ‘false negatives’ (59·3 %), followed by ‘dairy’ (23·4 %) and ‘meat and meat products’ (16·7 %).
Fig. 2 Number of specific, non-specific and total trans-fat ingredients indicators in the ingredients list of pre-packaged items labelled as having 0 g trans-fat and contain at least 1 trans-fat ingredients indicator (n 2657), stratified by category
In terms of products with non-specific trans-fat ingredient indicators, the majority (97·2 %) of the 2657 items had at least one of these indicators in their ingredients list, indicating they were ‘possible false negatives’. All items in ‘confectionery’, ‘edible oils and oil emulsions’, ‘snack foods’ and ‘sugars, honey and related products’ contained at least one non-specific trans-fat ingredient indicators. ‘Bread and bakery products’ (17·7 %), ‘cereal and cereal grain products’ (14·8 %) and ‘sugar, honey and related products’ had the highest proportion of items with three or more total trans-fat ingredients indicator.
Most products with trans-fat ingredient indicators originated from Asia (70 %; Fig. 3(a)), which also tended to have more trans-fat ingredients indicators in their ingredients list compared with items from Europe (P = 0·036) or North America (P = 0·003), and no statistically significant difference was observed between other region pairs (Fig. 3(b)). When comparing the trans-fat content (as %totalfat) between items labelled as having > 0 g trans-fat from different regions (Fig. 4), no significant differences between regions were observed (P ANOVA = 0·247).
Fig. 3 (a) The proportion of pre-packaged items labelled as having 0 g trans-fat and contain at least 1 trans-fat ingredients indicator from different regions of origin; (b) the number of total trans-fat ingredients indicators in the ingredients list of pre-packaged items labelled as having 0 g trans-fat and contain at least 1 trans-fat ingredients indicator (n 2657), stratified by region of origin. Differences were statistically significant for the following pairs: Asia v. Europe, P = 0·036; Asia v. North America, P = 0·003
Fig. 4 Boxplot of the trans-fat content (as % of total fat) of pre-packaged items reporting to have > 0 g trans-fat and have at least one trans-fat ingredient indicator in the ingredients list (total n 729), stratified by region of origin. For better layout of the figure, nine outliers for Asia (y = 53·9, 53·7, 45·7, 41·3, 28·6, 28·2, 26·4, 25·0, 21·0) and one outlier for North America (y = 33·0) were not displayed. No significant differences between regions were observed (P ANOVA = 0·247)
Discussion
In our audit, we found that 3·8 % of the pre-packaged foods in Hong Kong likely contained industrially produced trans-fats according to the nutrition labels, and around 7·5 % of the sampled items labelled as having zero trans-fat had likely mislabelled their true trans-fat content. We also found that 1·3 % of the sampled items exceeded the EU industrially produced trans-fat limit(Reference Junker15). The use of non-specific trans-fat ingredient indictors also appeared widespread, which could potentially mislead the general public to make a wrong decision in food purchasing selection.
Our findings suggest that there were two major issues with the trans-fat labelling requirements in Hong Kong. First, we found that among the 16 835 items labelled as zero trans-fat, 200 indeed contained specific trans-fat ingredients indicators (suggesting that they are ‘false-negatives’), and 2581 items contained non-specific trans-fat ingredients indictors in their ingredients lists. This is likely due to the lenient tolerance limit of the ‘trans-fat free’ claim in Hong Kong, which is set at < 0·3 g/100 g or 100 ml of food. Consumers may be misled by this lenient limit to believe items possibly containing industrially produced trans-fat as ‘trans-fat free’, which is counter-intuitive to the general health advice of lowering industrially produced trans-fat intake. A more stringent ‘trans-fat free’ limit of industrially produced trans-fatty acids (e.g. < 0·1 g/100 g or 100 ml) should be imposed to minimise the potential to mislead consumers. Second, we found that the use of non-specific trans-fat ingredient indicators in the ingredients lists is widespread. For example, fat or oil ingredients were often labelled ambiguously regarding their hydrogenation level, making it difficult for consumers to identify potential sources of industrially produced trans-fat by reading the ingredients lists, a strategy often suggested by health professionals in educating consumers(Reference Hess, Yanes and Jourdan27,Reference Kamel and Al Otaibi28) . Similar issues regarding ambiguity of trans-fat ingredients labelling were reported by researchers from Brazil(Reference Silveira, Gonzalez-Chica and da Costa Proenca19,Reference Ricardo, Peroseni and Mais20) , Australia(Reference Huang, Federico and Jones29) and Saudi Arabia(Reference Kamel and Al Otaibi28).
An interesting observation in our study is that non-alcoholic beverages have the highest level of industrially produced trans-fat. This is likely a result of the food categorisation system used in the current study, where beverage mixes such as 3-in-1 coffees (containing non-dairy creamer) and coffee/tea creamers are included under the food group ‘non-alcoholic beverages’. With these items (n 32) excluded, all remaining items were found to either be labelled as zero trans-fat or have no trans-fat ingredient indicator in the ingredients list. Similar issues were observed for ‘dairy’, which included ‘dairy alternatives’ (e.g. soya cheese and imitation cream) which may contain industrially produced trans-fat.
We also found that most items found to contain trans-fat ingredients indicator were from Asian countries. This is likely the nature of the market in Hong Kong, where people in general prefer products imported from other Asian countries for similar taste preferences, as well as lower costs. Products imported from other parts of the world are generally more often found in high-end, niche retail outlets, thereby limiting their presence in the market.
Among items labelled as having >0 g trans-fat, those from the ‘Other’ region were found to have the highest trans-fat content (as %totalfat). Nonetheless, the data used in this study were collected in 2019, before the enforcement of the new EU limit of industrially produced trans-fat of ≤ 2 %total fat in April 2021(Reference Kušar, Hribar and Lavriša14), which should be taken account when considering this finding. There is currently no such limit in Hong Kong(30).
Food reformulation is often considered an effective way to remove trans-fat from the food supply(Reference Buttriss31), which has been quite successful in the USA(Reference Stender, Astrup and Dyerberg32,Reference Mozaffarian, Jacobson and Greenstein33) . To reduce trans-fat in pre-packaged foods, ideally manufacturers should either replace them with cis-unsaturated fatty acids which can maximise health benefits by simultaneously reducing trans-fat and increasing unsaturated fat intake(Reference Mozaffarian, Jacobson and Greenstein33–Reference Downs, Bloem and Zheng35) or replace PHVO with vegetable oil solidified using interesterification(Reference Berry, Bruce and Steenson36), instead of replacing trans-fat with saturated fats(Reference Mozaffarian, Jacobson and Greenstein33). However, since Hong Kong imports most of its pre-packaged food supply(37), food reformulation is not something that the Hong Kong government has direct regulatory control.
Given the intake of trans-fat has been consistently shown to be associated with the risk of CVD and dyslipidaemia(Reference Collison, Makhoul and Inglis38,Reference Ganguly and Pierce39) , our findings are concerning as they suggest Hong Kong is quite far from achieving the goal to remove industrial trans-fat in the food supply set out in the WHO ‘REPLACE trans-fat’ action plan(Reference Ghebreyesus and Frieden12). To facilitate the achievement of this goal, the Hong Kong government has recently passed an amendment to the relevant regulation that prohibits from 1 December 2023 the import and sales of any foods, fats and oils that contain partially hydrogenated oils(Reference Lin40), which is a great step forward. Meanwhile, as discussed above, the Hong Kong government should also mandate the use of standardised terminology for oils and fats in food labels, as well as impose a more stringent ‘trans-fat free’ limit(Reference Ricardo, Peroseni and Mais20). The government should continue its effort in regularly examining the trans-fat content of the local food supply, with a focus on items more likely to contain trans-fat such as bread and bakery products and edible oils.
There are several strengths of our study. First, we utilised a pre-packaged food database that represents ∼70 % market share of grocery, meaning we should have covered the major dietary sources of trans-fat of the local Hong Kong population. Second, we have examined not only the nutrition information panel for trans-fat concentration but also the terms for specific and non-specific trans-fat ingredients indicators to identify potential ‘false negative’ similar to the work of Ricardo et al.(Reference Ricardo, Peroseni and Mais20)
However, we caution the readers to some limitations to our study. First, selection biases may have been introduced as products available only in supermarkets not sampled or stores such as convenience stores were not covered, although as explained above we do not believe this is critical when considering the main dietary sources of trans-fat of the local population. Second, although attempts have been made to differentiate ruminant v. industrially produced trans-fat based on the presence/absence of trans-fat ingredient indicators in most products, we were unable to do so for products containing a mix of the two types of trans-fat (e.g. cakes that use both dairy cream and palm oil). Since the EU trans-fat limit refer to industrially produced trans-fat only(Reference Junker15), we may have over-estimated the proportion of products exceeding the limit. Third, our audit found several items having > 60 % of total fat as trans-fat, which appears abnormal. However, upon further investigation, these were items reported to contain ∼0·5 g trans-fat/100 g with a low total fat content (e.g. 0·9 g/100 g), which may be due to errors in labelling, or that the manufacturers/importers chose to report a value that is at or above the cut-off of trans-fat free definition (< 0·3 g/100 g) in a bid to stay within the tolerance limit and avoid breaching the food labelling regulations. Last, the trans-fat concentration was obtained from the nutrition information panel instead of chemical analysis. However, conducting chemical analysis on all sampled items is logistically and financially unfeasible. Instead, future studies should focus on assessing the trans-fat concentration in products with ‘false negative’ trans-fat labelling using gas chromatography-mass spectrometry (GC-MS).
Conclusions
Around 4 % of pre-packaged food and beverages sold in Hong Kong in 2019 still likely contained industrially produced trans-fat, and about a third of these foods exceeded the EU limit of industrially produced trans-fat. The current trans-fat labelling requirements in Hong Kong are ambiguous and may not be effective in assisting consumers in identifying products free from trans-fat. Surveillance of trans-fat concentration in pre-packaged food and beverages likely to contain industrially produced trans-fat should continue until the removal of trans-fat from the food supply is mandated in Hong Kong.
Acknowledgements
None. Financial support: This study was funded by HKU internal funding. No external funding was received. Authorship: C.C.W.C. coded the data set, performed the statistical analyses, interpreted the data, wrote the first draft of the manuscript and critically reviewed the subsequent drafts of the manuscript. J.H.Y.W. provided the essential data set, interpreted the data, contributed substantially to the writing of the discussion and critically reviewed the subsequent drafts of the manuscript. J.C.Y.L. conceived, designed and supervised the study, interpreted the data and critically reviewed the subsequent drafts of the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work. Ethics of human subject participation: Not applicable.
Conflicts of interest:
The authors have no conflicts of interest relevant to this article to disclose.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980022002464
Open access
