Evaluating Hazards Posed by Additives in Food: A Review of Studies Adopting a Risk Assessment Approach

Processing and packaging has increased the use of food additives in the food industry. Some of these additives have associated health risks. This review looks at studies on risk assessment of food additives published between 2000-2015. These studies have majorly focused on synthetic food colors and preservatives like benzoate, sorbate, nitrite and nitrate. Most of the studies have shown that the intake was below the acceptable daily intake (ADI) for average consumers. For extreme consumers (95th percentile), intake was found to be above the ADI or approaching ADI for additives like sunset yellow FCF, erythrosine, tartrazine, sulphite, benzoate and nitrite. It is advisable to look at multiple scenarios of dietary exposure while evaluating risk. A surveillance system which documents adverse effects to food additives as well as monitors risk on a regular basis is important for every country to have. Such data would be beneficial to regulatory authorities as well as the industry in fixing usage levels of the additive in an effort to minimize health risk.

The extent to which a food additive can pose a health risk depends upon its toxicity and the dietary exposure.Joint FAO/WHO Expert Committee on Food Additives (JECFA) and Scientific Committee on Foods (SCF) establishes Acceptable daily intake (ADI) values for food additives.ADI values are calculating using a safety factor which ensures that if the additive is consumed daily at that level for the rest of one's life, there would be no "appreciable health risk" 2,3 .Food additives require to undergo a vigorous risk assessment procedure before their approval and entry into the market.Components of the process of risk assessment are hazard identification, hazard characterization, exposure assessment and risk characterization 4 .JECFA evaluates the safety of food additives and contaminants in food. 2 .All food additives that are used in a food-stuff are declared on the package using E-number given by European Union (EU) 5 or by an International Numbering System (INS-No.) signifying approval by JECFA 6 .
Member countries may request FAO / WHO or the Codex Committee on Food Additives and Contaminants (CCFAC) to initiate the process for risk assessment to be carried out by JECFA 7 .The scientific process of risk assessment not only estimates human risk associated with consumption of food additives but also assists in arriving at and establishing the ADI values for food additives 8 .

Risk assessment process
The risk assessment of food additives is purely a scientific process that requires the nutritionist and the toxicologist to work together 9 .The four steps of risk assessment are being discussed here.

Step 1: Hazard identification
The step of hazard identification identifies intrinsic properties of a food additive which can potentially affect health 10 .This is done by using the "weight-of-evidence approach".For this, review of appropriate published scientific databases is done for searching human and animal toxicological studies related to the food additive of concern.Epidemiological studies are given preference over lab based research data 4 .

Step 2: Hazard characterization
The safety of food additives is assessed by evaluating toxicity data.Hazard characterization involves both "dose-response extrapolation" and "dose scaling".In dose-response extrapolation, the toxicity levels estimated for animals need to be extrapolated both qualitatively and quantitatively to much lower doses for comparison with human exposure levels.In dose-scaling, JECFA uses mg / kg body weight for "inter-species" scaling as equivalence in toxic doses is difficult to establish 4 .Hazard needs to be characterized using multiple methods and using different approaches.Mathematical modelling can be used to characterize dose-response relationships 11 .
Step 3: Exposure assessment Estimating consumption of food additives is not simple and requires the assistance of experts in the field of nutrition.Usually the 24-hour dietary recall / record or the food frequency questionnaire (FFQ) is the tool of choice for estimating the intake of foods likely to contain additives.Other direct and indirect methods used for collecting data on food consumption are reviewed elsewhere 1 .Concentration of the additive in different foods is chemically estimated to ultimately calculate the dietary exposure to the additive.Some studies have also used the maximum permissible levels (MPLs) in foods or used industrial usage data to determine exposure.

Step 4: Risk characterization
In this step, the probability of occurrence of adverse toxic effects in humans as a result of exposure to food additive is assessed.This is usually done by comparing ADI values of the additive with exposure levels among humans 4 .Risk can be characterized using different exposure scenarios as depicted in Table 1.Scenarios 1 and 2 give the best estimate of the population where the exposure is assessed among the part of the population which only consumes foods containing the selected food additive and the level of added additive is average.Scenarios 3 and 4 covers the part of the population that always consume the same brand (brand loyal customers), and it is assumed that they are loyal to the brand with the highest reported level of food additive 12 .The worst case intake scenario is where content of additive in permitted food products is the maximum permissible levels specified by regulatory authorities and it is assumed that all consumed food products, in which addition of additive is permissible, contain that additive.This approach has been used in a number of studies [13][14][15][16][17][18] for assessing risk associated with consumption of certain food additives.
Hazard Index (HI) has also been used for characterizing risk and is calculated by the average daily dose (ADD) for an additive from the diet expressed as a percentage of ADI.If the HI is less than 100 % then there is no harm from exposure to that additive 19 .Characterization of risk from food additives not only focuses on the data generated for general population but also focuses on the at risk population who may be extreme consumers of the additive for instance, sub-groups like diabetics in case of artificial sweeteners.According to JECFA, occasionally exceeding the ADI of a food additive for individuals is not usually a big concern.However, monitoring of such groups of individuals would be wise in order to estimate the frequency of such exposures as well as the dose of the additive to which these individuals are regularly exposed.

Review of studies done from 2000 onwards on risk assessment of food additives
Studies from various countries between the period of 2000-2015 on assessment of risk of food additives have been reported here.A total of twentyfour studies on risk assessment of food additives were obtained after careful search of databases which used the keywords "risk assessment" in their title or while giving their results or discussion.The summary of these studies have been presented in Table 2 covering artificial sweeteners, artificial colours and chemical preservatives.

Artificial sweeteners
High doses (> 1500 mg / kg body weight / day) of artificial sweetener like saccharin have shown toxic effects like swelling of renal glomeruli, growth depression and carcinoma in rats 36 .Acute, sub-acute and chronic toxicity studies done for aspartame have shown no adverse effects 37 .However a study has reported that intake at half the ADI for aspartame may lead to neurobehavioral effects like irritable moods, depression, lower performance among human adults 38 .The emerging data for sucralose based on human and rodent studies have shown that high sucralose intake alters glucose, insulin and glucagon-like peptide-1 levels.It has been found to be mutagenic in nature in some studies [39][40] .
Only three studies on artificial sweeteners could be retrieved.The artificial sweeteners studied were acesulfame, aspartame, cyclamic acid, saccharin and sucralose in beverages.Exposure to selected sweeteners in one study 12 was assessed using all the four scenarios of exposure assessment.In two studies, individuals from all age-groups were studied 12,21 and in the third study dietary exposure of school children was studied 20 .Either chemical analysis using high performance liquid chromatography (HPLC) was used 20 or usage levels provided by manufacturers were used 21 for estimating sweetener content in foods.The data on food consumption was collected using a semi-quantitative FFQ.In the studies, the content of sweeteners in food items was below the MPLs.Hazard index for both average and high consumers in the studies was lower than 100% for all age-groups.Thereby, the risk was considered to be miniscule but more such studies for susceptible sub-groups like diabetics need to be carried out where the dietary exposure is likely to be higher.

Artificial colour
In most of the studies, colours studied were carmoisine, erythrosine, indigo carmine, ponceau 4R, sunset yellow FCF and tartrazine.Other colours studied were brilliant blue FCF, fast green FCF, allura red, azorubine, brilliant black, brown HT, green S, quinolline yellow, annatto, amaranth and riboflavin.Presence of non-permissible colours like auramine, rhodamine, orange II, blue VRS and malachite green in foods were detected in developing countries like India, where use of these colours as adulterants has been documented [24][25][26]28 . In al studies, levels of colours in food-stuffs manufactured by the unorganized sector were found to exceed their MPL's.
High doses (40 mg / kg body weight / day) of erythrosine administered to rats have shown toxic effects like thyroid follicular cell adenomas 41 .Doses between 1500 -2250 mg / kg body weight / day of Sunset Yellow FCF when administered to rats have shown toxic effects like body weight reduction and diarrhea 4 and Ponceau 4R (500 -1000 mg / kg body weight / day) has shown toxic effects like hepatic cirrhosis and renal problems in rats 42 .Although humans are not likely to be exposed to such high doses as used in animal studies, exceeding the ADI is not an unlikely scenario here.Usually in most of the studies risk due to additive exposure was assessed using scenario 1 which is the best estimate of actual consumption of colour by the targeted population.Only one study 22 assessed the risk using scenario 1 and 3.The exposure was studied in individuals of all age groups.In some studies, exposure to colours was studied only for children probably because foods likely to be coloured are consumed mostly by children.Also children are at a greater risk of exceeding ADIs because of lower body weights.Chemical analysis has been done to analyze colour content in most of the studies except one 30 where maximum permissible levels for colours (worst-case scenario evaluation) in food products was used.The method used for collecting data on food consumption varied from semi-quantitative FFQ to one-day 24-hour recall.In one study 23 a 7-day food diary was also used.
In some Indian studies [25][26]28 seasonal variation was also taken into account by conducting the study twice in a year. Thiswas done as consumption of coloured food products like beverages and ice creams tends to be higher in summer months.Adulteration with non-permissible colours like Rhodamine and Orange II were reported in studies [25][26] .Children appeared to have a higher risk of being an extreme consumer and sunset yellow FCF, tartrazine and erythrosine were the colours for which the HI was exceeding 100% in high consumers.Major contributors to the intake of artificial colours were desserts and beverages.In studies in Thailand, fish balls were found to be the major contributor to colour intake.

Synthetic preservatives
High doses (280 mg / kg body weight / day) of sulphite in rats have shown several toxic effects like irritation of stomach and intestinal lining, vomiting reflexes and hemorrhages while in humans (at doses of 1-50 mg /day), adverse reactions like urticaria, angioedema and aggravation of asthmatic symptoms have been observed 43 .High dosage (75mg) of benzoates administered to human participants in a study showed toxic effects like urticaria and angioedema 44 .Anaphylactic reactions were observed in humans when they were administered an oral dosage of 25 mg of sodium nitrite 45 .
The chemical preservatives for which risk assessment studies have been reported in literature are benzoic acid, sulphite, sorbic acid, nitrate and nitrites.In certain studies exposure to preservative was studied for one specific food item i.e. beverages 21 ,sauces 27 ,red curry paste 34 and meat products 33 assuming that exposure to the additive was majorly from one food item.In most of the studies risk was assessed using scenario 1 of exposure.In the Norway study 12 exposure to the selected preservative was assessed using all the four scenarios of exposure assessment.
In case of two studies 19,35 total diet study was conducted where consumption of all food items was studied.Exposure was assessed for all individuals above 1 year of age in almost all studies.Only in one study 21 usage levels provided by industry was used for defining preservative levels in foods.The data on food consumption was collected using either a 24-hour recall or a semi-quantitative FFQ.
Studies have shown that average intakes of preservatives studied were lower than the ADI for all the age-groups probably because only one major contributor to the preservative intake was assessed.HI approached 100% for benzoate, nitrate / nitrite in case of children 21,33,35 and exceeded 100% for children in some studies 21,[32][33] .The reason for intakes exceeding the ADI was the use of preservatives at levels above the maximum permissible in food stuffs which is again a problem of unregulated small scale manufacturing in developing countries.Most of the risk assessment studies have looked at exposure only according to scenario 1 which uses data on actual consumption of food containing the additive at average concentration levels.Changes in food habits, industry usage levels, brands used etc. can alter intake levels over a period of time.Hence risk should be examined at any point using multiple exposure scenarios in order to cover all possibilities.The process of risk assessment has to be carried out on a regular basis at periodic intervals to capture changes in exposure environment.
Although a big safety margin is incorporated while setting ADI levels, and occasionally approaching or exceeding ADI may not pose appreciable risk, adverse effects occurring in individuals at intake levels below ADI have also been documented.Food additives are valuable in the food processing industry.Responsible usage and monitoring are the keys to ensuring safety of the consumer.Risk assessment studies are hence important food safety tools.

COnCluSiOn
The risk assessment process is used to evaluate risk associated with intake of food additives.Various studies have been published between 2000-2015 on risk assessment of food additives.The studies have majorly focused on synthetic food colors, preservatives like benzoate, sorbate, sulphite, nitrite and nitrate and artificial sweeteners.No new dose-response data have been reported in any of the studies reviewed.In most of the studies the concentration of the additive in the food has been determined chemically, whereas in a few studies MPLs and usage levels provided by manufacturers have been used.The dietary intake data has been collected mostly using either 24-hour dietary recall or semi-quantitative food frequency questionnaire.The intake data has been collected either for individuals of all the age groups or children who are considered to be a vulnerable age group by virtue of choosing foods which are processed and packaged and because of the fact that they have lower body weight.
Most of the studies have shown that the intake was less than the ADI for average consumers.For high consumers (95 th percentile), intake was found to be above the ADI or approaching ADI for additives like sunset yellow FCF, erythrosine, tartrazine, benzoate and nitrite.Risk in most of the studies has been assessed using scenario 1.
A monitoring and surveillance system especially in developing countries is needed to document episodes of adverse health effects related to additive intake in the population which may escape detection because of symptoms being non-specific.This epidemiological data will be vital to evaluate the actual risk to consumers.It will also help to reevaluate usage levels of the additives by the industry as well as ADI levels.

ACknOWlEDGEMEntS
The first author is grateful to UGC (India) for granting her the senior research fellowship.No conflict of interest was declared by the authors.

table 2 : Summary of risk assessment studies on food additives Food Country Study Risk Characterization Additive (Reference) Participants Mean Additive High Consumer Major Food Age Group at Comments intake vs. ADi intake vs. ADi Contributor to Maximum Risk Additive intake from Additive
*ADI-Acceptable Daily Intake; MPL -Maximum Permissible Levels.