EFSA - Opinion of the Scientific Committee/Scientific Panel: Safety of advantame as a food additive

EFSA Journal 2013;11(7):3301 [68 pp.].doi:10.2903/j.efsa.2013.3301

Type:Opinion of the Scientific Committee/Scientific PanelOn request from:European CommissionQuestion number:EFSA-Q-2010-00943Adopted:03 July 2013Published:31 July 2013Affiliation: European Food Safety Authority (EFSA), Parma, Italy

The Panel on Food Additives and Nutrient Sources added to Food (ANS) provides a scientific opinion on the safety of advantame as a sweetener for use in the food categories specified in the dossier. Advantame is stable under normal storage conditions. The Panel noted that there is an indication of advantame instability in acidic beverages and thermally treated foods. Metabolism and toxicokinetics of advantame and its main metabolite, ANS9801-acid, have been studied in mice, rats, rabbits, dogs and humans. Advantame is rapidly but poorly absorbed and the main excretion route is via faeces. The Panel concluded that advantame does not raised concern with regards to genotoxicity and carcinogenicity. The critical effect observed in animal studies was maternal toxicity (gastrointestinal disturbances) in the prenatal developmental toxicity study in rabbits. The NOAEL for this effect was 500 mg advantame/kg bw/day. Advantame was well tolerated in single or repeated doses up to 0.5 mg/kg bw/day by normo-glycemic or diabetic subjects. The Panel established an ADI of 5 mg/kg bw/day based on the application of a 100-fold uncertainty factor to the NOAEL of 500 mg/kg bw/day for maternal toxicity from the prenatal developmental toxicity study in the rabbit. Conservative estimates of advantame exposure for high level adults and children consumers were below the ADI for the proposed use levels.

© European Food Safety Authority, 2013

Following a request from the European Commission, the Panel on Food Additives and Nutrient Sources added to Food (ANS) was asked to provide a scientific opinion on the safety of advantame as a high-intensity sweetener.

Advantame (ANS9801) is described by the applicant as a derivative of aspartame and it is reported to be approximately 37 000 times sweeter than sucrose.

The final advantame product is described as having a minimum purity of 97 % (on an anhydrous basis). The Panel noted that in several non-consecutive batches of the final product, the levels of platinum and palladium, residues from the catalysts used in the catalytic hydrogenation process, could amount to 1.7 and 5.3 mg/kg, respectively, and therefore, the Panel considered that a maximum limit for palladium and for platinum should be included in the specifications.

The advantame bulk material has been demonstrated to be stable following storage for up to five years under normal conditions and for up to six months under accelerated conditions. The stability of advantame has also been demonstrated under intended conditions of use in foods and beverages. In heat-treated beverages, it is reported that about 50 % of advantame is degraded. A similar result has been reported in yellow cake, where approximately 39 % of advantame is degraded during batter preparation and baking, and production of N-[N-[3-(3-hydroxy-4-methoxyphenyl)propyl]-α-aspartyl]-L-phenylalanine (ANS9801-acid) occurs. The Panel noted that there was an indication of advantame instability in acidic beverages and thermally treated foods.

The metabolism and toxicokinetics studies show that orally administered advantame is rapidly converted in the gastrointestinal tract to ANS9801-acid which can be rapidly absorbed. However, the bioavailability of the ANS9801-acid is limited. Following i.v. administration, advantame is rapidly converted to the ANS9801-acid. Following oral dosing of radiolabelled advantame, radioactivity is excreted mainly in the faeces, with urinary excretion representing a minor route. The ANS9801-acid metabolite has been detected in rat and human plasma. In contrast, in dogs there is little free ANS9801-acid in plasma. There was no evidence of accumulation of ANS9801-acid in healthy and type II diabetes subjects after 4 or 12 weeks of consumption. This is consistent with the short estimated half-lives of radioactivity of ANS9801-acid from the preclinical and clinical studies. ANS9801-acid can be hydrolysed to phenylalanine and an aspartic acid derivative, N-(3-(3-hydroxy-4-methoxyphenyl))propyl-L-aspartic acid (HF-1). HF-1 is a minor urinary metabolite in rats, dogs and humans. HF-1 can be further metabolised to 3-(3-hydroxy-4-methoxyphenyl)-1-propylamine (HU-1), which is also postulated to be formed directly from ANS9801-acid. HU-1 is a minor urinary metabolite in rats, dogs and humans, which has not been isolated from the faeces of any species. In human and dogs, HF-1 was the major faecal metabolite, however in rats the demethylated ANS9801-acid was the major faecal metabolite. Other minor urinary and faecal metabolites have been detected but not characterised. Studies were only carried out with advantame radiolabelled on the 3-hydroxy-4-methoxy-phenyl moiety of the molecule, and whilst the metabolic fate of the remainder of the molecule could not be investigated, it could be predicted. The applicant concluded that the rat is a suitable metabolic model for systemic metabolism of advantame in human. They further concluded that effects due to differences in faecal metabolites between rats and human would have been addressed by studies in dogs where faecal metabolites are similar to those in human. The Panel agreed that the experimental metabolic data supported the use of the available toxicological studies in rat and dogs for the safety assessment of advantame in humans.

In 13-week dietary studies in mice and rats, a chronic toxicity and carcinogenicity study in rats and a subchronic toxicity study in dogs, the rate and extent of systemic exposure of animals to advantame and ANS-9801 acid were highly variable and appeared to be characterised by non-linear kinetics and with no clear underlying mechanism over the dietary concentration range used in these experiments. By considering the changes in AUC24 data in the course of these subchronic or chronic studies, the Panel considered that advantame or ANS9801-acid do not accumulate even in case of high dietary doses given to these animal species for periods as long as 52 or 104 weeks.

Toxicokinetic data derived from a prenatal developmental toxicity study in rabbits demonstrated that when advantame administered by gavage at 2000 mg/kg bw/day, a dose similar to that achieved at advantame dietary concentration of 50 000 mg/kg diet in subchronic toxicity study in dogs and chronic toxicity and carcinogenicity study in rats, the systemic exposure to both advantame and ANS9801-acid is 5 to 25 times higher than in dietary administered animals. This would represent a worst case situation for a maximal systemic exposure of animals to advantame and ANS9801-acid. The Panel also noted that in this study, advantame and ANS-9801-acid were described to accumulate at the higher doses since increasing the dose of advantame resulted in a disproportionately higher systemic exposure to advantame than would be predicted from a linear relationship.

Overall, the Panel considered that the high variability and the lack of clear underlying mechanism did not permit such a clear conclusion on accumulation or non-linearity in kinetics

The safety of advantame has been examined in in vitro studies and in sub-acute, subchronic and long-term studies in mice, rats and dogs and in reproductive and developmental studies in rats and rabbits.

Since the LD50 for advantame in rats was shown to be higher than 5000 mg/kg bw in an oral rat study there is no concern with respect to the acute toxicity of advantame.

The Panel considered these changes as treatment related as food consumption in treated groups was comparable to controls but decreases in food efficiency were recorded and several haematological, blood chemistry and urinalysis parameters were also statistically significantly altered at high dietary concentrations of advantame. At lower dose levels the recorded changes

In all of the subchronic toxicity studies apart from the 13-week study in rats and 1-year study in dog decreases in body weight/body weight gain were seen at a dietary concentration of 50 000 mg advantame/kg diet. The Panel considered these changes as treatment related as food consumption in treated groups was comparable to controls but decreases in food efficiency were recorded and several haematological, blood chemistry and urinalysis parameters were also statistically significantly altered at high dietary concentrations of advantame. At lower dose levels the recorded changes, although statistically significant, were often slight and within the physiological ranges for the species, a dose-response was not apparent, they were not accompanied by morphological changes in any organs, and there was no consistency between the sexes in the recorded changes, or between the studies in the same species or between the species.

The Panel noted the changes in several parameters, indicative of an altered function of the immune system in the 13-week study in rats. The Panel noted that the repeated observation of effects on thymus and lymphocytes throughout the different subchronic studies in rats and dogs may be indicative of an effect on the immune system following exposure of laboratory animals to high dietary doses of advantame.

An additional study (TNO Triskelion report, 2013), investigating the potential immunotoxic effects of dietary administered advantame in young male and female Wistar rats, was submitted to EFSA. There were no treatment-related differences in total or differential white blood cell counts or in total protein concentration, albumin concentration and albumin/globulin ratio. There were no treatment-related differences in lymphocyte subsets. No treatment-related effects on proliferative response of splenocytes to mitogen stimulation by Con A were observed. There were no treatment-related differences in any cytokine production after mitogen stimulation. The weights of the adrenals, spleen and thymus at the end of treatment and the recovery period were not affected by treatment. Macroscopic examination and microscopic examination of lymphoid organs at the end of the treatment did not reveal any treatment related findings. The Panel considered that this study did not reveal any immunotoxic effects of advantame on the endpoints that were examined in the rat. The Panel noted that the data provided evidence for an absence of immunotoxic effects of advantame as were suggested by the previous rat 13-week study.

Advantame was found to be not genotoxic in the bacterial reverse mutation assay, the in vitro mouse lymphoma TK assay and the in vivo mouse micronucleus assay.

In the carcinogenicity study in mice, the malignant tumours with statistically significantly increased incidences were bronchio-alveolar adenocarcinoma and histiocytic sarcoma of the haematopoietic system. The pulmonary tumours are known to have a high background incidence in mice. For both tumours there was no dose-response relationship in either sex and their incidences were within historical control ranges for CD-1 mice, thus the Panel considered that these tumours were not treatment-related. Advantame was not considered by the Panel to be carcinogenic to mice at doses up to 50 000 mg/kg diet, the highest concentration tested, and equal to 5693 and 7351 mg/kg bw/day in males and females, respectively. The Panel noted that at the high dose the body weight gain of females was significantly lower than controls and a trend toward lower body weight was also observed for male mice, however the difference did not achieve statistical significance. With regard to the lower body weight gain in the high-dose group the no observed adverse effect level (NOAEL) was 10 000 mg advantame/kg diet, equal to 1057 and 1343 mg advantame/kg bw/day in males and females, respectively.

In the carcinogenicity study in rats a higher incidence of pancreatic islet-cell carcinoma or of pancreatic islet-cell carcinoma and adenoma combined was seen in high-dose males but the incidence was not statistically significantly different from that in controls and these pancreatic tumour incidences were within the historical control values. The incidence of mammary gland adenoma in high-dose females was statistically significantly higher than that of controls but it was within the range of the historical control values. In contrast, the incidence of mammary adenocarcinoma was not statistically significantly increased in any of the treated groups. Combining the incidence of mammary gland adenomas and adenocarcinomas no pairwise comparisons achieved statistical significance. When the authors of the study combined the incidence of mammary gland adenomas, adenocarcinomas and fibroadenomas a positive trend was noted with dose but no pairwise comparisons achieved statistical significance. The trend test was no longer statistically significant upon exclusion of the high-dose group. Similar results were described by the authors for combined incidences of adenomas and fibroadenomas. Since the incidence rates of all mammary tumour types and combinations were within the historical control range, and as mammary tumours are part of the background pathology of the ageing female rat, the Panel considered the increased incidences of mammary tumours in this study to be unrelated to treatment. Consequently, the Panel considered that the neoplastic findings in the carcinogenicity studies in rats do not provide evidence of carcinogenicity of advantame tested at dietary concentrations of 50 000 mg/kg diet, the highest dose tested. This dietary concentration was equal to 2621 or 3454 mg/kg bw/day in male and female rats respectively. The Panel noted that at the high dose body weight gains in the males and females treated for 104 weeks were reduced, attaining statistical significance in males only.

Reproductive and developmental toxicity studies on advantame included a 2-generation reproduction toxicity study in rats, and prenatal developmental studies in rats and rabbits. The NOAEL for reproductive toxicity in 2-generation reproduction toxicity study in rats was 50 000 mg advantame/kg diet, the highest dose tested. The NOAEL in rats for maternal toxicity from the prenatal developmental study was 15 000 mg advantame/kg diet (equal to 1419 mg/kg bw/day) and for prenatal developmental effects was 50 000 mg advantame/kg diet/day (equal to 4828 mg/kg bw/day) the highest dose tested. In rabbits the NOAEL for maternal toxicity was 500 mg advantame/kg bw/day based on disturbances of gastrointestinal tract causing morbidity that required killing the animals for welfare reasons at the next highest dose (1000 mg/kg bw/day) and the NOAEL for developmental toxicity was 1000 mg advantame/kg bw/day.

In other special studies conducted in rats, no behavioural or physiological alterations were observed and no mortality resulted from the administration of advantame. In addition, the locomotor activity of rats was unaffected by the administration of advantame compared with vehicle-treated animals. Although no significant effects were reported on the gastrointestinal motility of rats treated with 10 or 100 mg advantame/kg bw, a dose of 1000 mg/kg bw resulted in a moderate statistically significant decrease in gastrointestinal motility compared to rats administered the vehicle. Studies in dogs showed that the duodenal administration of advantame resulted in no significant effects on general respiratory status, resistance of the peripheral vasculature, and or on electrical status of the myocardium compared with the vehicle control group. Intra-duodenal administration of advantame produced no biologically meaningful effects on cardio-respiratory parameters.

The human tolerability of advantame has been tested in three clinical studies. Advantame was well tolerated in single doses up to 0.5 mg/kg bw or repeated doses up to 0.5 mg/kg bw/day, in healthy volunteers and in both normo-glycemic individuals and diabetics. Advantame did not affect the levels of various biochemical, haematological, or urinalysis endpoints and did not affect plasma levels of glucose or insulin in normo-glycemic individuals, or alter glucose tolerance or insulin resistance in the diabetic subjects provided advantame.

The degradation products of advantame β-ANS9801, β-ANS9801-acid, ANS9801-imide, and HF-1 have not been identified as impurities of the final advantame product. The safety of these degradation products was assessed individually in a series of in vitro genotoxicity assays and in an in vivo mouse micronucleus assay. Based on these data the Panel concluded that there is no concern with respect to genotoxicity of the degradation products of advantame. The Panel has previously reviewed data on phenylalanine and methanol concluding that they would not be of safety concern at levels greater than could be released from advantame.

Since advantame is a secondary amine, its potential to participate in the formation of N-nitroso compounds in foods and beverages and in the stomach was evaluated. No nitrosation of advantame was identified by the applicant and, therefore, the Panel considered that the potential formation of N-nitroso advantame does not raise a safety concern

After consideration of all the data on advantame and its metabolites and major and minor degradation products available for this evaluation, the Panel concluded that there were sufficient data with which to establish an acceptable daily intake (ADI).

Following specific studies the prior concern about possible immunological effects (as discussed above) other critical studies considered in setting the ADI were those indicating body weight effects in laboratory animals and those on developmental toxicity with regard to maternal toxicity and gastrointestinal symptoms in the rabbit.

The Panel considered the findings in the prematurely sacrificed rabbits from the prenatal developmental study to be related to advantame exposure. The Panel noted that although only one animal was affected in the mid-dose group (1000 mg advantame/kg bw/day), the effects observed in this animal were consistent with those found in sacrificed animals from the higher dose group (2000 mg advantame/kg bw/day). The Panel considered that the single incidence of adverse effects in the mid-dose group may be indicative of the proximity of the boundary dose for adverse effects of advantame, since incidence increased in the higher dose group. Taking into account that effects indicative of maternal toxicity in form of decreased feed intake were also observed in the developmental toxicity study in rats upon administration of advantame through the diet, the Panel considered that effects observed in the rabbits cannot be disregarded. Therefore the Panel identified a NOAEL of 500 mg advantame/kg bw/day for maternal toxicity in the prenatal developmental toxicity study in NZW rabbits by oral gavage.

This NOAEL was further supported by the difference in the toxicokinetics of high doses of advantame (around 2000 mg/kg bw/day) in the subchronic and chronic dietary administration studies in mice, rats and dogs and the oral gavage administration in reproductive and prenatal developmental toxicity studies in the rabbit. The internal doses of advantame and ANS9801-acid as estimated by AUC24 were markedly higher following oral gavage administration in the rabbit compared to dietary administration in rat, mouse and dog. There were no high dose kinetic studies by gavage available for rat, mouse and dog nor were high dose dietary kinetic data available for the rabbit. Therefore the Panel was unable to ascertain whether the observed differences in internal dose were due to the route of administration or were species specific. However given the similarity of kinetics at low bolus doses in all species studied, including humans, the Panel considered that it would not be appropriate, on the available information, to disregard the oral gavage study because of the apparent increase in internal dose. Therefore, the Panel considered that the NOAEL of 500 mg advantame/kg bw/day for maternal toxicity in the rabbit developmental toxicity study should be used as the point of departure for the derivation of the ADI.

The Panel noted that either a default uncertainty factor of 100 or lower uncertainty factor based on inter-species allometric scaling and inter-human variability could be applied to account for inter-species and inter-human variability in maternal toxicity. However, the Panel considered that the more conservative of these uncertainty factors would be appropriate in this case, due to the marked variability in the internal doses arising following dietary or bolus administration which was seen at both high and low dose levels, the absence of complete kinetic data set (high and low dose with both methods of administration) in any species studied and the corresponding difficulty in defining a clear kinetic model.

Anticipated exposures to advantame from its proposed use as a food additive and its metabolites (methanol and phenylalanine) have been calculated (mean and 95th percentile of consumers only) using the food consumption data at the individual level (e.g. raw data on food consumption by the individual consumer).

High-level consumption was only calculated for those foods and population groups where the sample size was sufficiently large to allow calculation of the 95th percentile. The Panel estimated chronic exposure for the following population groups: toddlers, children, adolescents, adults and the elderly. Calculations were performed using individual body weights.

Thus, for the present assessment, food consumption data were available from 26 different dietary surveys carried out in 17 different European countries.

When considering the proposed maximum use levels, the mean dietary exposure to advantame in European children-adolescents (aged 1-17 years) ranged from 0.02 to 0.33 mg/kg bw/day, and from 0.05 to 0.74 mg/kg bw/day for high level consumers (95th percentile). Exposure estimates for the European adult population give a mean dietary exposure to advantame ranged from 0.01-0.12 mg/kg bw/day, and from 0.03-0.28 mg/kg bw/day for high level consumers (95th percentile).

The Panel noted that its estimates could be considered as being conservative as it was assumed that all processed foods and beverages contained the sweetener advantame added at the maximum proposed use levels. It has to be noted that the anticipated exposure to advantame did not include possible applications from its use as an excipient in pharmaceutical products, an application which might occur if advantame were an approved food additive.

The exposure to methanol, which may result from ingestion of advantame-containing foods and beverages, was considered negligible compared to that from other dietary sources and as such of no concern from the safety point of view.

The exposure to phenylalanine expected from ingestion of advantame as a general purpose sweetener was considered of no safety concern for healthy consumers (adults and children). For a phenylketonuric child, the additional phenylalanine intake expected from ingestion of advantame-containing foods and beverages would represent a relatively small increment in the exposure to phenylalanine.

After considering all the data available, the Panel concluded that advantame does not raise concern with regards to genotoxicity and carcinogenicity. The critical effect observed in animal studies was maternal toxicity (gastrointestinal disturbances) in the prenatal developmental toxicity study in rabbits. The NOAEL for this effect was 500 mg advantame/kg bw/day.

The Panel established an ADI of 5 mg/kg bw/day based on the application of a 100-fold uncertainty factor to the NOAEL of 500 mg advantame/kg bw/day for maternal toxicity from the prenatal developmental toxicity study in the rabbit.

Conservative estimates of advantame exposure for high level adults and children consumers were below the ADI for the proposed use levels.

After considering all the data on stability, degradation products, toxicology and exposure, the Panel concluded that advantame would not be of safety concern at the proposed uses and use levels as a sweetener.

The Panel considered that a maximum limit for palladium and for platinum should be included in the specifications.

http://www.efsa.europa.eu/en/efsajournal/pub/3301.htm