Introduction

Toxicological studies have shown that acrylamide is neurotoxicant, reproductive toxicant, genotoxic and carcinogen in animals.¹ In fact, it has been categorized as “probable carcinogen to human” (category 2A) by International Agency for Research on Cancer (IARC) in 1994. Although the toxicity of acrylamide has been acknowledged long ago, it was not until the finding of its presence in food by Tareke in 2002 that it gets the attention as it has nowadays.² Studies have revealed that acrylamide mostly present in carbohydrate-rich material which is processed by high temperature (> 120°C). It is not found in protein-rich material such as meat and fish. Method of food processing is also an important factor because acrylamide is found in baked, fried, roasted products, but not found in boiled products.³ Acrylamide formation in food has been associated with Maillard reaction and with the presence of asparagine and reducing sugar in the food materials.⁴

Our previous work has demonstrated that acrylamide is a significant issue in deep fried fritter,⁵ hereafter referred simply as fritter, products in Indonesia. The fritter products which have high carbohydrate content are typically fried at 150-180°C for about 5-10 mins.⁶ Among several types of fritter, banana fritter and sweet potato fritter have become subject for acrylamide studies. For instance, Tandi et al.,⁷ have found considerable acrylamide content in banana fritter samples collected from Manado market. This finding is in agreement with studies conducted by Daniali et al.,⁸ in Malaysia, and Komthong et al.,⁹ in Thailand for the similar banana based fritter products. Meanwhile, acrylamide in sweet potato fritter samples that collected in Manado has been analyzed by Sengke et al.,¹⁰ The findings from monitoring studies were also consistent with banana and sweet potato fritter which were produced in laboratory set up.^8,11

Occurrence rate of cancer in Indonesia has been estimated to be at least 170-190 new cases per 100.000 people annually.¹² Several factors, including carcinogen exposure, might contribute to this rate. In order to assess if a certain carcinogenic substance (for instance, acrylamide) contributes a risk to population, risk assessment can be performed.

In the case of acrylamide in fritter products in Indonesia, there have been several publications covering the acrylamide level as aforementioned. However, to the best knowledge of author, none has discussed about its exposure and risk assessment. Risk assessment for acrylamide has been subject for several studies such as in Spanish potato crisps,¹³ total diets in Hong Kong and Dutch population^14,15 and even total diets at international level.¹⁶ However, Indonesia has typical practices in street food production (e.g. time/temperature conditions during frying and typical carbohydrate-rich foods) which implies different acrylamide concentrations, and obviously different consumption patterns. Therefore, the objective of this paper was to perform a probabilistic exposure assessment and risk characterization of acrylamide in Indonesian fritter products and possible mitigation strategies were also discussed.

Methodology

Model Design

The exposure assessment was carried out with probabilistic approach. Simulation was conducted using Excel spreadsheet (Microsoft Corporation) with @Risk add-in program (version 7, Palisade, Newfield, NY). The software program was widely used in other risk assessment studies(Vinci et al.,¹⁷; Yogendrarajah et al.,¹⁸; Lachenmeier and Rehm¹⁹; Van de Perre et al.,²⁰; Imathiu²¹). Acrylamide exposure was modeled as the result of multiplication between concentration and consumption. The exposure model with detailed cell addresses and formula used was tabulated in Table 1. For modeling, first order Monte Carlo simulations using Mersenne Twister generator with 10,000 iterations were applied and run in triplicate to test the stability of the outcomes.

Table 1: Exposure model: cell addresses and formula used in the exposure assessment of acrylamide in Indonesian fritter products.

Concentration data were taken from several publications including both monitoring and laboratory set up studies. Only two fritters, namely banana fritter and sweet potato fritter were used in the model due to lack of data for other fritter types, including tempeh fritter which was the most preferred by Indonesian.⁵ The values were modeled in PERT or Uniform distribution, depending on the type of data available. Minimum, most likely and maximum values were used to define the PERT distribution while only minimum and maximum values were used for Uniform distribution. Prevalence of acrylamide in the fritters was not put into model as 100% of analyzed samples in cited publications were positive with acrylamide. Meanwhile, the primary consumption data was based on Pratama⁵ which covered 263 respondents with the age ranged of 20-40 years old (adult).

Value/function input setting for acrylamide in banana fritter

Four studies were selected to define the distribution of acrylamide content in banana fritter (Table 2). The study by Daniali et al.⁸ was the only laboratory set up study used, where fritter’s acrylamide content was measured as the effect of banana maturity. Information on the total of sample analyzed and the analysis results were used to produce the minimum, most likely and maximum value of each study. The values were then used to make PERT distribution of each study. The total acrylamide distribution of banana fritter (cell C12) was calculated with Discrete distribution from all 4 PERT distributions by considering the proportion of analyzed samples in each cited studies.

Table 2: Acrylamide content of banana fritter from previous studies.

* Total of samples that are included in this table

Value/function input setting foracrylamide in sweet potato fritter

Only a few studies related to acrylamide content in sweet potato fritter were found from literature. Lim et al.,¹¹ studied the effect of frying oil types and the number of frying sequences toward the acrylamide content. The acrylamide value used in this study was the value taken from the first frying sequence with palm olein frying oil which ranged from 251.1 – 340.9 µg/kg (from 4 analyses). It is in accordance with the typical preparation of fritter in Indonesia which uses palm olein and fried only once. The second study was a monitoring study from 7 samples collected from Manado, Indonesia market, which acrylamide content ranged from 118.54 – 866.75 µg/kg.¹⁰ Due to the type of information available, Uniform distribution was applied for each cited study. Subsequently, a Discrete distribution similar to banana fritter was used to estimate the total sweet potato fritter’s acrylamide content (cell C13).

Fritter Consumption and Exposure

The frequency of Indonesian consumer on consuming the fritters product (consumption per week) based on our previous study⁵ are shown in Table 3. Answer in a range of certain periods (e.g. 2-3 times/week) was converted into the average value. Amount (in pieces) of fritters consumed in one chance is presented in Table 4 where consumption more than 5 pieces was collectively converted to 6 pieces. Frequency and amount of consumption were modeled using Discrete distribution by taking response proportion into account (Table 1). In order to estimate the consumption per week body weight or Estimated Weekly Intake (EWI), it was assumed that average body weight is 60kg and one piece of fritter weights 100g. It was found that 55.5% of population like banana fritter whereas only 17.1% like sweet potato fritter.⁵ This information was incorporated to calculate the exposure of acrylamide from fritters product (cell C19 Table 1).This data was the original exposure or “Scenario 1”. Meanwhile, to illustrate the effect of a possible mitigation, an if scenario or “Scenario 2” of 70% reduction was simulated based on study by Jung et al.,²² which achieved 73.1% reduction of acrylamide by immersion in citric acid 1% for 1 hour.

Table 3: Frequency of fritters consumption⁵

Table 4: Amount of fritter consumed in one occasion⁵

Risk Characterization

Risk characterization is the last part of a risk assessment. It combines the information from the previous steps, such as hazard characterization and exposure assessment into advice relevant for decision making.²³ Margin of Exposure (MoE) approach was used in this study. The approach has been preferred for carcinogenic and genotoxic compound such as acrylamide.²⁴ The MoE is defined as ratio between toxicological threshold (benchmark dose) and estimated human intake, and the value is often used to compare risk among toxic compounds.¹⁹ Lower MoE value indicates a higher risk and according to EFSA,²⁴ MoE value lower than 10,000 shows a potential public health concern. MoE was calculated as below equation:

MoE = Bench Mark Dose Lower limit (BMDL₁₀)/ estimated dietary exposure

Estimated acrylamide BMDL₁₀ for mammary tumor of 0.16 mg/kg-bw/day or 1120 µg/kg-bw/week¹⁶ was used in this study. Meanwhile, dietary exposure was taken from mean and percentiles of simulated probabilistic model.As a second type of risk characterization, the estimated weekly intake was also compared with Tolerable Weekly Intake (TWI) of acrylamide according to Tardiff et al.,²⁵ which was 18.2 µg/kw-bw/week.

Results and Discussion

Estimated Acrylamide Intake

The Monte Carlo simulation of two scenarios (original exposure and 70% reduction) resulted in probabilistic distribution of acrylamide intake shown in Figure 1 and summarized in Table 5. It was estimated that 90% of acrylamide intake from fritter consumption was between 0 – 76.1 µg/kw-bw/week (Scenario 1), while the simulated scenario of 70% acrylamide reduction resulted in intake of 0 – 23.4 µg/kg-bw/week (Scenario 2). The median (P50) of the Scenario 1 was 4.10 µg/kw-bw/week (P95 = 76.06 µg/kg-bw/week). These results showed that 50% (P50) of the Indonesian population had an acrylamide intake at or below 4.10 µg/kw-bw/week from fritter products, while 95% had an acrylamide intake at or below 76.06 µg/kw-bw/week.

Figure 1: Probabilistic density of estimated acrylamide intake from fritter by Indonesian population. Cut off by TWI = 18.2 µg/kw-bw/week, X-axis value was truncated to give better visualization Click here to View figure

The result was higher than the similar study which actually covered a range of products (including bakery, cereals, potato, etc) in Dutch population which had P50 of 3.5 µg/kw-bw/week and P95 of 8.4 µg/kw-bw/week.¹⁵ Therefore, if the total diet is concerned, the acrylamide intake in Indonesian population would be even higher compared to that of Dutch population.Furthermore, the simulated average intake in this study was 14.85 µg/kg-bw/week which could be considered as high level due to a single food group, while the estimated international mean dietary exposures for total dietranged between 7.7 – 33.6 µg/kw-bw/week.²⁶ Comparing the exposure with TWI, it was estimated that 17.4% of population exceed the tolerable intake value.

Table 5: Estimated acrylamide intake (µg/kg-bw/week) and MoE at mean and percentiles from fritter products by Indonesian population in original situation and after presumed 70% reduction in acrylamide concentration as potential mitigation strategy.

The second scenario was to illustrate the impact of possible mitigation to decrease the acrylamide content in the fritters product. The usage of citric acid 1% immersion method was chosen due to its simplicity, and the citric acid is also commercially available with affordable price in Indonesia. Considering the fritters are mainly sold as street food⁵, cost and simplicity of method would play an important factor for mitigation success. Furthermore, the nature of banana as raw material has sour taste, so that the method was expected to not greatly impact the sensory of the product. However, this issue needs to be confirmed, especially for the second product (sweet potato fritter) in this study. With the addition of acid, product had a lower pH which led to the conversion of nonprotonated amine (–NH₂) into protonated amine (–NH₃⁺) of free asparagine (acrylamide precursor). This phenomenon was effective in blocking its reaction with carbonyl group from reducing sugar and eventually reduced the acrylamide formation in the product.²²

With the 70% reduction of acrylamide concentration (Scenario 2), the estimated average intake was reduced to 4.51 µg/kw-bw/week. P90 was 12.29 µg/kw-bw/week, which was lower than the estimated TWI 18.2 µg/kw-bw/week. It showed that 90% of the population had an exposure less than tolerable limit if the scenariowas carried out successfully. Thus, the remaining population who still exceed the TWI was estimated to be 6.9%.

Risk Characterization

The MoE estimation based on simulated dietary exposure of acrylamide is shown in Table 5. The MoE derived from average exposure estimate for Scenario 1 was 75, while high level estimate from 90% percentile (P90) was 28. Bolger et al.,¹⁶ estimated the MoE using international mean intakes (total diet) of acrylamide and several toxicological baselines as point of departure (POD). When the same POD was used (i.e. mammary tumor’s BMDL₁₀), the current study had lower values than  the study by Bolger et al.,¹⁶ which estimated MoE were 160 and 40 for mean and P90, respectively.Lower MOE indicates higher risk posed by the substance and by using the same POD means that estimated exposure in current study was higher compared to the study byBolger et al.,¹⁶ Consequently, by lowering the exposure intake in Scenario 2, estimated MoE increased (lower risk) to 248 and 48 for mean and P90, respectively.

Based on EFSA,²⁴ MoE of 10,000 or higher, if it is calculated using BMDL₁₀, would be a low concern from public health point of view, and thus could be low prioritized in a risk management action. Further categorization was used by Lachenmeier and Rehm,¹⁹ where MoE < 100 fall into ‘risk’ category, and ‘high risk’ category is given when MoE < 10. According to the mentioned classification, acrylamide exposure from fritter consumption possess concern for Indonesian public health and hence needs a risk management.

Using the average MoE of 75, acrylamide in Indonesian fritter falls into ‘risk’ category and would be excluded from the category if the mitigation scenario was in place (MoE 248). However the value was still below 10,000 which indicated public health concern persists, most likely due to the high consumption level. Therefore, the reduction on acrylamide level should be followed by the moderation on consumption to successfully lower the exposure. Nevertheless, using the MoE estimate will help risk managers (e.g. food safety authorities) to prioritize actions, especially when dealing with multiple toxic substances.

Conclusion

The dietary exposure model in this study has demonstrated the probabilistic acrylamide intake from fritter products in Indonesian population. Margin of exposure was calculated from the estimated intake and showed a potential food safety concern among Indonesian population. Therefore, this quantitative risk analysis could be used as an advice for risk managers (Food Safety Authorities) to mitigate the problem.Potential impact of mitigation act in reducing the risk level associated with acrylamide intake from fritter has been shown by the scenario study. As with any risk assessment studies, the quality of the data and assumptions used are the defining factors of the estimate accuracy. Thus, uncertainties for this study arise from lack of concentration data and large scale consumption survey. Addressing the issues is expected to improve the accuracy of the risk assessment study in the future.

Acknowledgement

The research was made as part of International Training Program in Food Safety, Quality Assurance and Risk Assessment organized by Department of Food Safety and Food Quality, Faculty of Bioscience Engineering, Ghent University.

Funding

Authors are grateful for funding from VLIR-UOS scholarship.

Conflict of Interest

The author(s) do not have any conflict of interest.

References

1. Exon JH. A review of the toxicology of acrylamide. J Toxicol Environ Health B Crit Rev. 2006;9(5):397-412. doi:10.1080/10937400600681430
   CrossRef
2. Krishna kumar T, Visvanathan R. Acrylamide in Food Products: A Review. Food Process Technol. 2014;5(7):344. doi:10.4172/2157-7110.1000344
   CrossRef
3. Arvanitoyannis IS, Dionisopoulou N. Acrylamide: formation, occurrence in food products, detection methods, and legislation. Crit Rev Food Sci Nutr. 2014;54(6):708-733. doi:10.1080/10408398.2011.606378
   CrossRef
4. Das AB, Srivastav PP. Acrylamide in snack foods. Toxicol Mech Methods. 2012;22(3):163-169. doi:10.3109/15376516.2011.623329
   CrossRef
5. Pratama Y. Potential food safety hazard of acrylamide in deep fried fritter sold as street food in Indonesia. J Appl Food Technol. 2016;3(2):1-6. doi:http://dx.doi.org/10.17728/jaft.2
   CrossRef
6. Ilmi IMB, Khomsan A, Marliyati SA. Frying oil and fritter’s quality in Indonesian household. J Apl Teknol Pangan. 2015;04(02):61-65 [In Bahasa Indonesia]. doi:10.17728/jatp.2015.12
   CrossRef
7. Tandi D, Fatimawali F, Wehantouw F. Acrylamide Analysis of Banana Fritter Sold in Manado City Using HPLC. Pharmacon. 2012;1(02):79-85. [In Bahasa Indonesia].
8. Daniali G, Jinap S, Hanifah N, Hajeb P. The effect of maturity stages of banana on the formation of acrylamide in banana fritters. Food Control. 2013;32(2013):386-391.
   CrossRef
9. Komthong P, Suriyaphan O, Charoenpanich J. Determination of acrylamide in Thai-conventional snacks from Nong Mon market, Chonburi using GC-MS technique. Food Addit Contam Part B. 2012;5(1):20-28. doi:10.1080/19393210.2012.656145
   CrossRef
10. Sengke C, Citraningtyas G, Wehantouw F. Acrylamide Analysis of Sweet Potato Fritter Sold in Manado City Using HPLC. Pharmacon. 2013;2(03):91-95. [In Bahasa Indonesia].
11. Lim PK, Jinap S, Sanny M, Tan CP, Khatib A. The Influence of Deep Frying Using Various Vegetable Oils on Acrylamide Formation in Sweet Potato (Ipomoea batatas L . Lam) Chips. J Food Sci. 2014;79(1):115-121. doi:10.1111/1750-3841.12250
    CrossRef
12. Tjindarbumi D, Mangunkusumo R. Cancer in Indonesia, present and future. Jpn J Clin Oncol. 2002;32(Supplement 1):S17-21. doi:10.1093/jjco/hye123
    CrossRef
13. Arribas-Lorenzo G, Morales FJ. Dietary exposure to acrylamide from potato crisps to the Spanish population. Food Addit Contam Part A Chem Anal Control Expo Risk Assess. 2009;26(3):289-297. doi:10.1080/02652030802477954
    CrossRef
14. Wong WWK, Chung SWC, Lam C, Ho YY, Xiao Y. Dietary exposure of Hong Kong adults to acrylamide: results of the first Hong Kong Total Diet Study. Food Addit Contam Part A Chem Anal Control Expo Risk Assess. 2014;31(5):799-805. doi:10.1080/19440049.2014.898189
    CrossRef
15. Boon PE, De Mul A, Van Der Voet H, Van Donkersgoed G, Brette M, Van Klaveren JD. Calculations of dietary exposure to acrylamide. Mutat Res – Genet Toxicol Environ Mutagen. 2005;580(1-2):143-155. doi:10.1016/j.mrgentox.2004.10.014
    CrossRef
16. Bolger PM, Leblanc JC, Setzer RW. Application of the margin of exposure (MoE) approach to substances in food that are genotoxic and carcinogenic EXAMPLE: Acrylamide (CAS No. 79-06-1). Food Chem Toxicol. 2010;48(SUPPL. 1):S25-S33. doi:10.1016/j.fct.2009.10.040
    CrossRef
17. Vinci RM, Jacxsens L, Van Loco J, et al. Assessment of human exposure to benzene through foods from the Belgian market. Chemosphere. 2012;88(8):1001-1007. doi:10.1016/j.chemosphere.2012.03.044
    CrossRef
18. Yogendrarajah P, Jacxsens L, Lachat C, et al. Public health risk associated with the co-occurrence of mycotoxins in spices consumed in Sri Lanka. Food Chem Toxicol. 2014;74:240-248. doi:10.1016/j.fct.2014.10.007
    CrossRef
19. Lachenmeier DW, Rehm J. Comparative risk assessment of alcohol, tobacco, cannabis and other illicit drugs using the margin of exposure approach. Sci Rep. 2015;5:8126. doi:10.1038/srep08126
    CrossRef
20. Van de Perre E, Jacxsens L, Lachat C, El Tahan F, De Meulenaer B. Impact of maximum levels in European legislation on exposure of mycotoxins in dried products: Case of aflatoxin B1 and ochratoxin A in nuts and dried fruits. Food Chem Toxicol. 2015;75:112-117. doi:10.1016/j.fct.2014.10.021
    CrossRef
21. Imathiu S. Quantitative Microbiological Risk Assessment of Two Street Foods Sold in a Kenyan Town with Regard to Salmonella Contamination. Curr Res Nutr Food Sci J. 2018;6(1):41-50. doi:10.12944/crnfsj.6.1.05
    CrossRef
22. Jung MY, Choi DS, Ju JW. A Novel Technique for Limitation of Acrylamide Formation in Fried and Baked Corn Chips and in French Fries. J Food Sci. 2003;68:1287-1290. doi:10.1111/j.1365-2621.2003.tb09641.x
    CrossRef
23. Renwick AG, Barlow SM, Hertz-Picciotto I, et al. Risk characterisation of chemicals in food and diet. Food Chem Toxicol. 2003;41(9):1211-1271. doi:10.1016/S0278-6915(03)00064-4
    CrossRef
24. EFSA [European Food Safety Authority]. Opinion of the Scientific Committee on a request from EFSA related to A Harmonised Approach for Risk Assessment of Substances Which are both Genotoxic and Carcinogenic. EFSA J. 2005;282:1-31. doi:10.1021/bk-2010-1048
    CrossRef
25. Tardiff RG, Gargas ML, Kirman CR, Leigh Carson M, Sweeney LM. Estimation of safe dietary intake levels of acrylamide for humans. Food Chem Toxicol. 2010;48(2):658-667. doi:10.1016/j.fct.2009.11.048
    CrossRef
26. JECFA. Evaluation of certain contaminants in food. World Heal Organ Tech Rep Ser. 2011;(959):1-105, back cover. http://apps.who.int/iris/bitstream/10665/44514/1/WHO_TRS_959_eng.pdf.