EFSA Journal 219

The EFSA Journal (2005) 219, 1-36
Opinion of the Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food on a request from the Commission related to Semicarbazide in food Question number EFSA-2003-235
Adopted on 21 June 2005 by written procedure
SUMMARY The Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC) has been asked to advise the European Commission on the occurrence of semicarbazide (SEM) in food. The European Food Safety Authority (EFSA) issued preliminary advice on SEM in 2003, when the occurrence of SEM in food, derived from food packaging, was first discovered. The Panel was asked on this occasion to gather data on the occurrence of SEM in all types of food, to explain the conditions under which SEM may be formed in food and to evaluate the analytical methods used. In the light of this information, the Panel was asked to assess the risks posed by semicarbazide in all types of food. The approach taken by the Panel on this question was to search the scientific literature and to take account of information from the Commission, national authorities and trade associations. SEM has been found to occur in different types of foods and the source of SEM varies. SEM is a metabolite of the veterinary medicine nitrofurazone, but since the use of nitrofurazone is illegal in the EU, SEM from this source should not be detectable in foods. SEM can be present in foods as a result of migration from sealing gaskets used in the metal lids of jars and bottles. In this case, the origin of the SEM is thermal breakdown of azodicarbonamide, a blowing agent used to foam the plastic gaskets. SEM has been found in food products made using flour in which azodicarbonamide has been added as a dough-improver a practice that is not permitted in the EU. Other sources have been suggested but are less well documented. SEM is reportedly formed as a reaction product of the action of hypochlorite on food additives such as carrageenan and on foods such as egg white powder. Finally, SEM may be present at background levels naturally, may be formed at low levels when some foods are dried, and may also derive from as yet unidentified sources. The method of analysis used to test foods for SEM involves acid hydrolysis and a derivatisation step with 2-nitrobenzaldehyde. The derivative is then determined using liquid chromatography coupled to tandem mass spectrometry with a detection limit in the region of 0.2 g/kg. The acid hydrolysis step liberates bound residues for analysis and the analysis method therefore measures total SEM (free and bound) in the sample. The acid hydrolysis conditions used in the analytical method are not dissimilar to normal gastric conditions. Since the current state of knowledge on the bioavailability of any bound residues is incomplete, in this evaluation there is no distinction made between SEM that is present as such in a food
http://www.efsa.eu.int/science/afc/afc_opinions/catindex_en.html
Semicarbazide in food
The EFSA Journal (2005) 219, p. 2 of 36
sample and any SEM that may have been formed from precursors in the food under acidic conditions used in the analysis. It is concluded that the method of testing for SEM provides concentration data that are suitable for this risk evaluation. On the basis of the information available, migration of SEM from the breakdown of azodicarbonamide (ADC) in sealing gaskets is by far the largest source of exposure known. The concentration data available from analyses of food undertaken by different countries were similar. The highest potential intakes of SEM are in infants consuming ready-to-feed infant milk and baby food, attributable to the larger gasket areas involved in the packaging and their small body weight. Reasonable worst case estimates of intake for infants fed on products packaged in glass jars and bottles range from 0.35 to 1.4 g/kg bw/day. Adult exposures to SEM from this source are likely to be much lower than infant exposures, due to the lower contribution of foods packaged in bottles and jars to the total diet of adults, the lower contamination levels derived from the smaller gasket areas involved for that packaging, and the higher adult body weight. A reasonable worst case estimate of intake for an adult would be 0.02 g/kg bw/day. Commission Directive 2004/1/EC prohibits the use of azodicarbonamide in food contact materials from 2nd August 2005. Once existing stocks of packaged foods are used up, exposure of consumers by this route will have been eliminated. Other possible sources of SEM in foods contribute far less to exposure than that estimated above for packaging. Bread made using flour treated with ADC can contain SEM. In laboratory tests the SEM concentration in bread was 28 g/kg. ADC is not permitted as a flour treatment agent in the EU and the importation of bread and bakery ware is likely to be very low. There is the potential for exposure from breaded animal products imported into the EU (e.g. frozen breaded chicken or fish products). Taking an upper figure of 5 g/kg of product this would give an intake of SEM of 1 g/person from a consumption of 200g of product. For a high consumer of egg products that may be contaminated by 50 g/kg SEM as a result of using hypochlorite as a sanitising solution on production equipment, a reasonable worst-case estimate of exposure is 0.008 g/kg bw/day. For the food additive carrageenan, that may become contaminated with SEM at a mean concentration of 65 g/kg from use of hypochlorite in the production process, if consumption was up to the Acceptable Daily Intake (ADI) for carrageenan then the intake of SEM from this source could be up to 0.005 g/kg bw/day. SEM has been shown to be carcinogenic in mice, but not rats. Literature data on genotoxicity together with the results of recent studies indicate that SEM is mutagenic but not clastogenic in some test systems in vitro, notably in the absence of an exogenous metabolising system. In vivo, negative results were reported in studies on DNA damage in liver and lung of mice, and in the micronucleus assay in the mouse. Based on the overall weight of evidence provided by the studies performed, which included a study using a highly sensitive methodology, the Panel concluded that the weak genotoxicity exerted by SEM in vitro is not expressed in vivo. The new data allaying the concern on genotoxicity in vivo, and the likely reductions in exposure following replacement of the most significant, currently known source of SEM in the diet (gaskets on glass jars and bottles), offer further support to the preliminary advice given by EFSA in 2003 that the risk, if any, from consumption of foods containing SEM is judged to be
very small, not only for adult consumers but also for infants. In this respect the Panel noted that SEM is a weak non-genotoxic carcinogen for which a threshold mechanism can be assumed. A large margin of at least 5 orders of magnitude exists between the dose causing tumours in experimental animals and human exposure, including that of infants. The Panel therefore concluded that the issue of carcinogenicity is not of concern for human health at the concentrations of SEM encountered in food. KEYWORDS Semicarbazide, CAS No 57-56-7, SEM, azodicarbonamide, CAS No 123-77-3 nitrofurazone, CAS No 59-87-0, gaskets, hypochlorite.
TABLE OF CONTENTS SUMMARY........................................................................................................................1 KEYWORDS .....................................................................................................................3 BACKGROUND................................................................................................................6 TERMS OF REFERENCE...............................................................................................7 INTRODUCTION .............................................................................................................7 ANALYTICAL ASPECTS ...............................................................................................7 SOURCES AND POSSIBLE MECHANISMS OF FORMATION OF SEM ..............8 SEM from nitrofurazone use in veterinary medicine............................................8 SEM from azodicarbonamide (ADC) foamed plastic gaskets ..............................8 Regulation of ADC used in food contact plastics ..............................................8 Detection of SEM in food originating from packaging materials .....................9 SEM from hypochlorite-treated foods and food additives ...................................9 SEM in hypochlorite treated foods ....................................................................9 SEM in egg powder ...........................................................................................10 SEM in food additives........................................................................................11 SEM from azodicarbonamide-treated flour...........................................................11 SEM present at background levels naturally or from as yet unidentified sources .......................................................................................................................11 OCCURRENCE AND EXPOSURE ................................................................................12 SEM from nitrofurazone use in veterinary medicine............................................12 Migration of SEM from lid gaskets ........................................................................12 Concentration data from different countries .....................................................12 Summary of levels reported ...............................................................................14 Food intake ........................................................................................................14 Exposure estimates ............................................................................................15 SEM from hypochlorite-treated foods and food additives ...................................16 SEM from the use of chlorinated wash water....................................................16 SEM from egg powder .......................................................................................16 SEM from carrageenan .....................................................................................17 SEM from azodicarbonamide-treated flour...........................................................18 TOXICITY OF SEM.........................................................................................................18 Metabolism and toxicokinetics. ...............................................................................18 Acute toxicity ............................................................................................................18 Developmental toxicity .............................................................................................18 Oral studies. ......................................................................................................18 Intraperitoneal studies.......................................................................................19 In vitro study......................................................................................................21 Carcinogenicity .........................................................................................................21 Genotoxicity ..............................................................................................................23 Bacterial mutagenicity assays ...........................................................................23 Mammalian cell gene mutation assay ...............................................................23 In vitro chromosomal aberration assays ...........................................................24 In vivo cytogenetics ...........................................................................................24 DNA damage in vitro/in vivo .............................................................................25
Effects on DNA in acellular systems .................................................................25 Summary and discussion of the toxicity studies ....................................................25 CONCLUSIONS AND RECOMMENDATIONS ..........................................................27 The occurrence and the precursors of SEM in foods ............................................27 The method of analysis used for SEM in foods......................................................27 Estimates of exposure...............................................................................................28 Toxicity ......................................................................................................................29 INFORMATION PROVIDED TO EFSA .......................................................................30 REFERENCES ..................................................................................................................30 Annex 1. The EFSA advice on baby foods from October 2003.........................................35
BACKGROUND In early 2003 semicarbazide (SEM) was reported to have been found in a number of food products from different manufacturers that are packaged in glass jars and bottles with metal lids sealed with plastic PVC gaskets. Examples of the foods involved include some types of baby food, fruit juices, jams and conserves, honey, pickles and sterilized vegetables, mayonnaise, mustard, sauces and ketchups. The presence of SEM was linked by industry to the permitted use of azodicarbonamide (ADC) as a blowing agent to make a foamed plastic which is suitable for gaskets. The amounts that were detected in foods were variable but low, ranging from not detectable up to 20 microgrammes per kg of food (20 g/kg). Levels of 1 - 7 mg SEM per kg of gasket material have been detected in extracts of the gaskets themselves. Preliminary advice on the possible occurrence of SEM in certain packaged foods was issued by EFSA in July 2003 and is available at (http://www.efsa.eu.int/science/afc/afc_documents/365/p_afc_doc_0111.pdf). In October 2003, an ad hoc group of experts was convened, including members of EFSAs Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC Panel), its Panel on Contaminants in the Food Chain and its Panel on Dietetic Products, Nutrition and Allergies. It heard a presentation from the food and packaging industries updating their findings since July 2003. This new information covered industry progress on development of alternative technologies and strategies to reduce or eliminate SEM in foods, further information on the range of concentrations of SEM found so far in foods packaged in glass jars and bottles, and exposure estimates for infants. The ad hoc expert group was specifically asked by EFSA to advise further on possible risks to infants, given that this is the consumer group for which potential exposure to SEM on a body weight basis is likely to be the highest in Europe. The advice of the ad hoc group can be found at http://www.efsa.eu.int/science/afc/afc_documents/364/p_afc_doc_03_en11.pdf and included discussion of microbiological and nutritional aspects of baby foods (Annex 1), which need to be considered alongside the toxicological aspects. Member States informed the Commission through the Rapid Alert System that SEM had also been found in a wide variety of food products which could not be attributed to food contact materials. SEM was found in egg products and its presence was attributed to the production process for these egg products rather than the illegal administration of nitrofurazone to laying hens. Two further sources of possible SEM contamination of food were also recognised during 2003; breakdown of ADC used as a flour treatment agent and SEM formed by hypochlorite action on food or food additives. ADC is not permitted as a flour treatment agent in the EU. In addition SEM is a metabolite of nitrofurazone and had for a long time been used as a marker residue to detect use of this banned veterinary medicine in animal products. The presence of residues of nitrofurans and/or their breakdown products in foods is illegal in the EU. In view of the potential impact of SEM on public health it is necessary for the Commission to have a complete picture on occurrence in all types of food and feed; and especially whether the presence of SEM is due to the intentional use of an avoidable component or processing step or whether it results from contamination.
TERMS OF REFERENCE The Commission requests EFSA to advise on the scientific basis for the occurrence of SEM in food. The EFSA is asked especially to: Gather and collate data of occurrence of SEM in all types of food (e.g. food of animal origin, animal feed, processed food, raw materials, etc.). The analytical method used and the performance criteria would be needed in order to compare the results. Explain the conditions under which SEM may be formed in food (e.g. during food processing, during the analytical procedure, etc.) Assess the risk of semicarbazide levels currently reported in food products, as well as to examine new information which might become available, especially through the data collection process. Complete the full assessment of the risks posed by semicarbazide in all types of food. INTRODUCTION The approach taken by the Panel on this question was to search the scientific literature and the Panel also received information from the Commission, national authorities and trade associations. ANALYTICAL ASPECTS Semicarbazide (H2N-NH-CO-NH2) is a small molecule that belongs to the hydrazine group of chemicals. The current methods of analysis used to detect SEM in food involve acid hydrolysis and a derivatisation step with 2-nitrobenzaldehyde (2NBA). These steps are to help extract and measure SEM bound to protein. The SEM-NBA derivative is then determined using LC-MS/MS (liquid chromatography coupled to tandem mass spectrometry). The methods were developed for the testing of meat for SEM as a marker for nitrofurazone, which is rapidly metabolised to SEM. These methods may not be needed for the detection of SEM in free solution and may in theory lead to artefacts, for example the production of SEM as a result of hydrolysis of ADC or some of its impurities, during the analytical work-up. The 2NBA method measures total SEM (free and bound) in the sample and cannot distinguish between the two forms. The detection limit of the method is in the region of 0.2 g/kg. There are currently no methods for measuring free SEM with g/kg sensitivity in food matrices. For the purposes of this evaluation, no distinction is made between free and bound SEM in estimating intakes - the prudent assumption is made that the all SEM measured in foods is freely available. Tissue-bound residues of SEM are reported to be bioavailable in the digestive systems of mammalian species (Hoogenboom et al. 1992; Gottschall & Wang 1995; McCracken et al. 1997). A number of efforts are on-going to develop new analytical methods. For example, DG-JRC (IRMM-Geel) has the following research objectives: a) identification of an alternative marker residue for nitrofurazone use in animal derived food products; b) development and validation of analytical methods for SEM in various food products; c) baby food analysis (method validation, monitoring and investigation of source of contamination); d) organisation and coordination of proficiency test exercises for the determination of SEM in food products.
SOURCES AND POSSIBLE MECHANISMS OF FORMATION OF SEM There are presently five chemical sources to which the presence of SEM detected in foods has been attributed. These are, in approximate chronological order of discovery: 1. As a metabolite of the veterinary medicine nitrofurazone. 2. As a thermal breakdown product arising from the use of azodicarbonamide to make foamed plastic sealing gaskets for metal lids on glass food jars. 3. As a reaction product formed by the action of hypochlorite on food additives and/or foods. 4. As a decomposition product of azodicarbonamide used as a flour treatment agent. 5. Present at background levels naturally or from as yet unidentified sources. SEM from nitrofurazone veterinary medicine use The foods in which SEM has now been found include foods not of animal origin and its presence in such foods can not be explained by veterinary use of nitrofurazone. The limits of detection for SEM are in the range of 0.2 to 1 g/kg. Minimum required performance limits have been set up by Commission Decision 2002/657/EC for the metabolites of nitrofurans in products of animal origin. Consequently, if the ban on nitrofurazone use is enforced effectively then there should be no detectable levels of SEM from this source in the food of European consumers. SEM from azodicarbonamide (ADC) foamed plastic gaskets ADC is structurally related to SEM as shown below. ADC. Azodicarbonamide SEM. Semicarbazide H2N-CO-N=N-CO-NH2 H2N-NH-CO-NH2
ADC is authorized for use in the EU as a blowing agent for plastics in contact with food. These plastics are used in sealing gaskets for metal lids on glass bottles and jars. Blowing agents are added to polymers during processing to form minute gas cells throughout the plastic. During high temperature processing, ADC decomposes to form gases, primarily nitrogen and carbon monoxide together with some carbon dioxide and ammonia and non-volatile residues such as biurea. Normally, the residues of non-volatiles are only small (biurea about 2% of added ADC). However under certain conditions the residues of biurea could be as high as 34%. Other non-volatile products can be urazole, cyanuric acid, and cyamelide. Regulation of ADC used in food contact plastics ADC is listed in Commission Directive 2002/72/EC relating to plastics materials and articles intended to come into contact with foodstuffs. The directive permits ADC as an additive with the condition that it is used only as a blowing agent. Commission Directive 2004/1/EC of 6th January 2004 has been issued and this amends Directive 2002/72/EC with the effect that, as regards the additive ADC, its use in food contact materials is prohibited as from 2nd August 2005. Foods packed before that date may still be sold after that date.
ADC was re-evaluated by the EC Scientific Committee on Food in 2003 when it was confirmed as acceptable for use in food contact materials, but only as a blowing agent and not for materials in contact with alcoholic beverages. It was classified in List 3 (substances for which an ADI or a TDI could not be established, but where the present use could be accepted). SEM is not specifically regulated by EU food packaging directives but if it were present in food packaging materials, for instance as an impurity or a reaction or degradation product, its presence in food would be covered by the Framework Regulation EC 1935/2004. Under Article 3 of this Regulation it could be present in food contact materials provided it did not transfer into foodstuffs in quantities that could endanger human health. Detection of SEM in food originating from packaging materials In 2003, a laboratory conducting routine monitoring for the presence of residues of nitrofuran antimicrobials discovered traces of SEM in a purely vegetable-based product. Further investigation revealed that SEM (tested as 2NBA-SEM) occurs in different food products that are packaged in glass jars and bottles with metal lids and furnished with a foamed PVC seal. The use of metal closures with a foamed gasket include baby food jars. These typically are Press-on Twist-off closures which include a tamper-evident safety button. A feature of these closures is that a greater surface area of the PVC seal is in contact with the food. The gaskets are used in the packaging for a wide range of foods in jars and bottles including fruit juices, jams and conserves, honey, baby food, pickles and sterilized vegetables, mayonnaise, mustard and sauces. SEM, ADC and biurea have similar chemical structures and it was necessary to prove that SEM was not an artefact of the analytical method used. In November 2003 an independent test method for free SEM was developed and validated in a two-laboratory exercise. This work reported the finding of free SEM in the chemicals ADC, biurea and urazole. Subsequently, it has been demonstrated that SEM is a minor thermal decomposition product of azodicarbonamide used in the gaskets of certain food jar and bottles and that it migrates from the gaskets to the packaged food. SEM from hypochlorite-treated foods and food additives SEM in hypochlorite treated foods Gatermann et al. (2003) reported the results of laboratory experiments on the formation of SEM in food subjected to hypochlorite treatment. Food and food products were treated with a hypochlorite solution containing 12% active chlorine, levels far higher than would normally be used, in order to investigate the source and mechanism of formation of SEM. They found that SEM was formed at 10 65 g/kg in shrimps, venison, soybean flakes, carrots, and honey. Levels of SEM were 350 450 g/kg in egg powder and gelatine after hypochlorite treatment. These findings do not appear to have been confirmed yet by a second independent laboratory. These same workers followed-up their findings by using lower hypochlorite levels (Hoenicke et al. 2004). After incubation with 1% active chlorine at room temperature overnight, chicken, milk, egg white powder, soybean flakes, red seaweed, carrageenan, locust bean gum, gelatine,
and starch all showed a small formation of SEM. In all cases the formation was less than 1 g/kg except for carrageenan (7.7 g/kg), gelatine (4.5 g/kg) and starch (1 g/kg). When the same foods/ingredients were treated with 0.05% active chlorine overnight, only egg white powder, carrageenan and starch showed any SEM formation, in the range 0.1 to 0.3 g/kg. At 0.015% active chlorine any increase was barely detectable, with only carrageenan and starch showing any formation, at 0.1 and 0.3 g/kg respectively, after treatment overnight. Chlorinated water may be used as a processing aid to wash foods, for example fruits and vegetables, provided that it meets the definition of a processing aid i.e. does not perform a function such as preservation in the final product and leaves no harmful residues [Directive 89/107/EEC). It is a reasonably widespread practice to wash certain ready-to-eat foods using water with a chlorine content up to 100 parts per million (0.0001%). Similarly, the Codex committee on fish and fishery products stated that chlorinated water at a recommended level of 10 parts per million of free chlorine is used on fish and fishery products. Given that these concentrations are 100-fold and 1000-fold lower respectively than the concentration of 0.015% active chlorine that gave barely any detectable SEM formation in the tests described above, given that the chlorine wash will be for a far shorter period that the overnight conditions used in the laboratory tests, and given that the processes also generally incorporate a final rinse with chilled water with just 2 to 4 parts per million free chlorine, it is considered that the use of chlorinated water as a processing aid is highly unlikely to give any detectable residues of SEM in the washed food. Disinfection of equipment and surfaces with disinfecting agents such as sodium hypochlorite is common practice in industrial food production. This should be rinsed off before the cleaned surface is used again into contact with food. With effective rinsing, no subsequent formation of SEM is to be expected. SEM in egg powder The presence of SEM in egg products was detected in Belgium in 2003 by the routine monitoring of egg white powder for residues of the nitrofuran class of antibiotics. Low but persistent contamination (1-150 g/kg) was found in many batches tested. The Belgian authority took action since this finding was considered as a non conformity in regards to Council regulation EEC/2377/90 which forbids the use of nitrofurazone (with SEM as marker) in animal entering the food chain. The lots affected were sequestered, information was posted using the Rapid Alert System for Food and Feed (RASFF), and action was started to trace the origin of SEM. These investigations showed that contamination with SEM occurred during the extraction of lysozyme from egg albumin because bleach solution was used to sanitize the carrageenan column used for the extraction. Up to that time, in 2003, the possibility that hypochlorite bleach could react with organic substances to form SEM (Gatermann et al., 2003) was unknown and unexpected. Through the RASFF, the Belgian authorities announced that batches under temporary seizure were released if they contained less than 50 g/kg of SEM (RASFF, 2003). Another possible cause of SEM formation in egg white powder could be any heat treatments applied. It has been reported that heating at 70-80 C for 3 days yields up to 30 g/kg of SEM if heated in an open system (Gatermann et al., 2004). In a closed system, in contrast, the formation of SEM was very limited, no more than 1-2 g/kg. Other products were reported to show similar results, although less pronounced, and the authors suggested a role for oxygen in the heat-formation of SEM from compounds in foods.
SEM in food additives As described above, reports of SEM in processed egg products were attributed to the carrageenan columns used during processing, which were sanitised using chlorine bleach. Carrageenan is also used as a food additive (E 407) as a thickening, gelling and suspending agent for example in ice cream, pudding, yoghurt, fruity jellies, chocolate milk, and different meat products. It is a complex mixture of different polysaccharides obtained from red seaweed (Rhodophyceae) by several processing steps including extraction, washing, precipitation, drying and grinding. One type of carrageenan is bleached during production. This is semirefined carrageenan or Processed Euchema Seaweed (PES, E407a) which is bleached to remove colour since the production process does not involve a precipitation step. Bleaching with calcium hypochlorite is a standard practice in making PES from seaweed. The industry association for producers of agar, alginates, carrageenan and processed eucheuma seaweed (Marinalg) reported that they were initiating trials to replace calcium hypochlorite with an alternative bleaching agent for the production of PES. Information was received from the Norwegian Food Safety Authority on their analysis of 8 samples of food grade carrageenan from three different companies. Seven out of the eight samples were below the detection limit of 5 g/kg and one sample contained SEM at 750 g/kg. The geographic location of the supplier of this sample was unknown. SEM from azodicarbonamide-treated flour ADC is not permitted as a flour treatment agent in the EU. Where permitted, for example in the USA and Brazil, the use of ADC up to 45 mg/kg of flour is allowed. In a recently published paper, the formation of SEM from ADC used to treat flour was studied and reported (Pereira et al., 2004). The authors found that when ADC was added to flour it formed SEM at about 0.1% yield. For example, they estimated that chicken products with added coatings made from flour could contain 0.2-5 g/kg SEM in the whole product. They observed that this level would be sufficient to be detected, with the result that the sample could fail EU import controls using SEM as the metabolic marker for nitrofurazone use. In a second publication, bread was made using commercial flour that had ADC listed as an additive (Becalski et al., 2004). The level of treatment of the flour was considered by the authors to have been about 20 mg/kg as is normal industry practice in Canada. The treated flour contained SEM at 3 g/kg. Bread made using the flour contained 28 g/kg with most of the SEM concentrated in the crust. SEM present at background levels naturally or from as yet unidentified sources Other sources of SEM in foods that had been suggested, based on structural similarities, included the herbicides triazophos, diflufenzopyr and roxarsone, but these possibilities have since been discounted (van Rhijn et al., 2004). Saari & Peltonen (2004) have reported finding SEM in all 18 samples of wild crayfish samples tested. The crayfish were trapped in various locations in Finland in Autumn 2003 and the
authors ruled-out all possibility that the crayfish had been treated with nitrofurazone or that they had become cross-contaminated during handling and transport. The crayfish were boiled in tap water and the authors noted that Helsinki uses ozone and not chlorine for tap water supplies. The meat of all 18 crayfish samples tested contained SEM, at 0.4 to 18 g/kg. Chicken and shrimp samples were cooked and tested as blank controls and no SEM was detected in them. OCCURRENCE AND EXPOSURE SEM from nitrofurazone veterinary medicine use If the ban on nitrofurazone use is enforced effectively then there should be no detectable levels of SEM from this source in the food of European consumers. The detection limits of the surveillance methods used are in the range 0.2 to 1 g/kg. Taking the UK as an illustrative example, the average consumption of all animal products (meat, meat products, fish, poultry) excluding milk, is 103 gram per person per day. For high adult consumers of animal products (average x 3), consumption of 300g of food and making the unlikely assumption that it was all contaminated with SEM at the highest detection limit of 1 g/kg, the intake would be 0.005 g/kg bw/day for a 60 kg bw person. Migration of SEM from lid gaskets Concentration data from different countries Analytical surveys of the concentration of SEM in foods packaged in glass jars and bottles with metal lids have been reported recently for several European countries (Information provided to EFSA see list page 30). The levels reported are reasonably consistent between different surveys. The individual surveys are described below. For the calculation of the average concentrations, the upper-bound assumption was made, that samples reported as not detected contained SEM at the detection limit stated in the survey. The Netherlands. In a survey of 40 samples of baby foods, covering 7 brands, purchased in The Netherlands in October 2003, all samples contained detectable SEM. The samples of baby food were fruit, vegetables, meat, pasta, and combinations thereof. The range of concentrations found was 3 to 26 g/kg and the average was 13 g/kg. The authors of the report noted that there appeared to be no relation between the level of SEM, type of food, content (size of jar), or brand. They also measured high levels (ppm) of SEM in gaskets and so concluded that migration from lids was the most probable origin of the SEM in foods. The same workers also conducted a smaller survey of general foods (i.e. not specifically baby foods) including fruit and vegetables packaged in glass jars and bottles. The 22 samples were purchased in October 2003. SEM was found in a range up to 2 g/kg. 10 of the 22 samples had no detectable SEM (<0.03 g/kg). The average concentration for the 22 samples was 0.5 g/kg (http://www.vwa.nl/download/rapporten/Voedselveiligheid/20040209_rapport_sem_in_baby_f ood.pdf)
Industry data. The European industry has supplied results of analyses to EFSA. Levels of SEM in 63 samples of baby foods, of all types and covering several brands and countries of production, were in the range 0.1 to 27 g/kg. The average for all samples was 11 g/kg. Results for 7 samples of ready-to-feed infant milk were also provided. The range of results was 5 to 14 g/kg and the average SEM concentration in the 7 samples was 9 g/kg. Results for 45 samples of general foods packed in glass jars and bottles with metal lids were also provided. The foods were sauces, jams, pickles etc. The range of concentrations of SEM was 0.3 to 3.3 g/kg and the average of all 45 results was 1.2 g/kg. 24 of the 45 samples contained no detectable SEM, with the limit of detection generally being 1 g/kg or lower. Spain. From surveys covering several regions of Spain, results for 88 samples of baby foods in glass jars and bottles have been reported. A small number of other results were reported too, but it was noted by the researchers that these samples may not have been made adequately homogeneous before analysis and so these results have not been included here. The baby food samples included fish, fruit, vegetables, meat, etc. The range of SEM concentrations was 1 to 87 g/kg and the average for the 88 samples was 16 g/kg. The same laboratories also conducted smaller surveys of general foods (i.e. not specifically baby foods) including vegetables, jams, sauces and fish packaged in glass jars and bottles. The samples were 30 in number. SEM was in the range 0.5 to 6 g/kg. The average of the 30 samples was 0.9 g/kg. France. In a small laboratory study, 4 different types of fruit and vegetable baby foods were studied for SEM concentration. More than 10 samples were tested and the range of SEM concentrations was 6 to 15 g/kg. The method of showed good repeatability when different specimens for the same jar were analysed (c.v. ca. 10%) and the variability when different jars of the same product were analysed was not great (c.v. ca. 15%). The authors noted that for some samples there was a concentration gradient down the jar consistent with the proposal that migration from the gasket was the source of SEM. The authors also noted that further warming of the foods in the jars, to 50C as if in preparation for feeding, did not increase SEM levels further. Germany. Results have been reported for 133 samples of all types of baby foods purchased in Germany. The range of concentrations of SEM was <0.5 to 140 g/kg. The average was 13 g/kg. The authors noted that 15 of the 133 samples were above 25 g/kg. 26 of 133 samples had no detectable SEM (<0.5 g/kg). A small number of other foods, not specifically intended for babies, was also tested. For the 12 samples, the range of values was <0.5 to 6 g/kg and the average was 1.5 g/kg. 8 of the 12 samples contained no detectable SEM (<0.5 g/kg). Finland. In a small survey of 15 samples of baby foods in jars, covering both domestic production and imported products in 2003, SEM was detected in 12 of the 15 samples. The range of values was <1 to 28 g/kg and the average was 11 g/kg. The measurement uncertainty was reported to be 0.9 g/kg when measuring in the range 1 to 10 g/kg and was 1.2 g/kg when measuring samples containing SEM above 10 g/kg. Five samples of pickles and sauces were also tested. Four of these samples contained 2 g/kg or less of SEM and one sample, preserved cucumber, contained 10 g/kg. The average for the 5 samples was 3.1 g/kg. Ireland. In a survey of 50 samples of baby foods covering 6 different brands, with 12 samples purchased in 2003 and 38 samples purchased in 2004, the majority of samples (37/50) contained detectable SEM. The range of concentrations found was <1 to 42 g/kg and the
average was 7 g/kg. Seven other miscellaneous food products were tested and none contained detectable SEM, <3 g/kg. Slovenia. Twenty samples of babyfoods taken from the Slovenian market were analysed for SEM in 2004. The range of values reported was 1 to 17 g/kg. The individual data and the average were not reported (Perharic et al., 2004). Summary of levels reported Baby foods are reported to have higher levels of SEM than other packaged foods. This may be because they have a higher ratio of gasket area to food mass for these small pack sizes and because of the heat processing conditions used. The levels found in different countries are similar. The average levels were 13 g/kg (NL, n=40 samples reported); 11 (Industry, n=63); 16 (ES, n=88), 13 (D, n=133); 11 (FIN, n=11) and 7 g/kg (IRL, n=50). The similarity in the range of concentrations and the average concentrations reported, suggests that there are not any great differences in the production and/or sales in the different countries. Neither is there any evidence for a large methodological bias (e.g. measuring error) in any of the reported surveys. The average for all 385 samples of infant foods tested was 13 g/kg. Far fewer data were provided for ready-to-feed infant milk. The range of results was 5 to 14 g/kg and the average SEM concentration in the 7 samples was 9 g/kg. For other miscellaneous foods, not specifically intended for infants, there was again general agreement in the levels found in different countries. The samples included, for example, fruit, vegetables, jams, pickles, sauces and fish. The average levels were 0.5 g/kg (NL, n=22 samples reported); 1.2 (Industry, n=45); 0.9 (ES, n=30), 1.5 (D, n=12); 3.1 (FIN, n=5) and <3 g/kg (IRL; n=7). The average for all 121 miscellaneous food samples tested was 1.0 g/kg. Food intake Human exposure is likely to occur over a lifetime, resulting from the consumption of a wide range of food products in jars and bottles, including baby foods, fruit juices, sauces, mayonnaise, ketchup and preserves. A significant variation in exposure is expected according to life stage, with the highest intakes on a body weight basis likely to occur within the first year of life, due to the progressive introduction of dietary solids from 4 months of age on including commercial baby foods. In the earlier EFSA assessment of SEM intake (EFSA, 10/2003), a conservative value of 700 g/day of processed baby foods was assumed for a 6-month baby weighing 7.5 kg. In this evaluation, intake data for relevant manufactured baby food products was reviewed from data available in the literature. Data from Germany (Kersting et al., 1998), France (Even et al., 2001) and UK (Mills and Tyler, 1992) were considered. These intake studies were compared in the recent EFSA opinion on epoxidised soybean oil (EFSA, 2004). The food consumption data from the DONALD study in Germany (Kersting et al., 1998) were slightly higher than for the other 2 studies and the DONALD data are used here.
Table. Intakes (average and 95th percentile, consumers only) of commercial baby foods and drinks known to be potential sources of SEM exposure in infants (Kersting et al., 1998) Age in months 3 6 9 12 Average bodyweight in kg 5.8 7.8 8.8 9.8 Average intake in gram per infant per day 67 +/- 65 195 +/- 114 234 +/- 127 208 +/- 128 P95 intake in gram per infant per day 192 407 464 424
Beyond the first year, studies in France and UK have indicated that the contribution of baby foods in jars to the total diet decreases significantly. Regarding adults, the quantity of available data on target foods is very limited, but it is expected that the contribution of food in jars to the total diet of adults is likely to be much smaller than in the diet of infants. Exposure estimates Exposure assessment was targeted to highly exposed subjects i.e. to subjects who systematically consume foods which are likely to contain semicarbazide. Available analytical data suggest that variability in semicarbazide concentration in products packaged in jars and bottles with metal lids and PVC gaskets is generally low and is of the same order of magnitude between and within the different brands of processed foods. Therefore, it was considered appropriate to estimate chronic exposure to semicarbazide on the basis of the average observed semicarbazide concentration and that such exposure assessment would be valid even for subjects who are loyal to specific brands of processed food. To estimate the exposure of infants to SEM from commercial baby foods and juices, an average contamination value of 13 g/kg was combined with the average and the 95th percentile of food consumption. The average contamination value of 13 g/kg was derived from the survey data summarised above. The survey data are dominated by baby foods in jars with relatively few samples of juices for drinking tested. However, the UK (Mills & Tyler, 1992) and the more recent French studies (Even et al., 2001) indicated that a significant proportion (probably up to 40%) of the baby foods consumed are vegetable or fruit products and/or fruit juices. Since lower levels of SEM are generally found in fruit juices compared to baby foods, using an average concentration level of 13 g/kg is likely to be conservative. With regards to taking the average and the 95th percentile of food consumption, a conservative assumption is that all baby foods consumed are packaged in glass jars and bottles with metal seals, rather than in other packaging materials such as metal cans or plastic pouches. Such an assumption is not unrealistic.
Table. Exposure (average and 95th percentile, consumers only) to SEM taking an average concentration of 13 g/kg in all commercial baby foods and drinks known to be potential sources of SEM exposure in infants Age in months 3 6 9 12 Average bodyweight in kg 5.8 7.8 8.8 9.8 Average exposure in g/kg bw/day 0.15 0.33 0.35 0.28 P95 exposure in g/kg bw/day 0.43 0.68 0.69 0.56
To estimate the exposure of infants to SEM from ready-to-feed infant milk pre-packaged in glass bottles with metal lids, the average SEM concentration of 9 g/kg (from 7 samples) can be combined with a consumption of 700 mL each day by a 4.5 kg bodyweight infant. (SCF, 2002) The resulting exposure would be 1.4 g/kg bw/day. Adult exposures to SEM are likely to be much lower than infant exposures, due to the lower contribution of foods in bottles and jars to the total diet of adults, the lower contamination levels derived from the smaller gasket areas involved for that packaging, and the higher adult body weight. Taking an assumption that 1 kg of food contaminated with SEM at an average concentration of 1.0 g/kg is consumed each day, the exposure for a 60 kg bodyweight adult would be 0.02 g/kg bw/day. SEM from hypochlorite-treated foods and food additives SEM from the use of chlorinated wash water Based on the foregoing discussions, it is considered that the use of chlorinated water as a processing aid is highly unlikely to give any detectable residues of SEM in washed food. SEM from egg powder In the analysis of production from one manufacturing facility in Belgium that sourced eggs from a number of countries, 78% (478 of 613 samples) contained no detectable SEM, <2 g/kg. 126 samples (21%) contained SEM less than 50 g/kg and 9 samples (1.5%) contained SEM in the range 50-100 g/kg (EFSA/AFC/SEM/16). Dried egg white powder may be used to make foods such as cakes and mayonnaise. According to the producer, the egg white powder is normally used at a final concentration of about 1% in most food products that use this powder (Bernard, 2003). That is, the powder is reconstituted by a 10-fold dilution with water and then the reconstituted egg product used at about 10% in the food product. In the worst-case, using the reconstituted product at 100% and with no allowance for any water loss from the finished food product, the dilution factor of egg powder to finished product would be 10%.
Using this figure of 10% dilution, taking 50 g/kg as a high but not unrealistic level, the final food product could contain 5 g/kg of SEM. At a 1% usage dilution the concentration in the final foods could be 0.5 g/kg. The average UK consumption of eggs and egg products is 14 gram per person per day. For a high consumer of egg products, consuming 1 kg/day of food made with 1% dry egg powder or 100g/day of food made using the maximum likely usage of 10% egg powder, the intake of SEM would 0.008 g/kg bw/day for a 60 kg bw person. SEM from carrageenan The industry association for producers of agar, alginates, carrageenan and processed eucheuma seaweed (Marinalg) provided a description of two industrial process to obtain carrageenan and one process to obtain PES. They also provided analytical results from testing production batches from across the industry for SEM. For 34 batches of carrageenan made by Process A the range of results was 0 to 4 g/kg SEM in the food additive with a mean of 1 g/kg. For carrageenan prepared using Process B the average for 13 batches was 6 g/kg and the range was 0 to 46 g/kg. (Hoenicke et al. (2004)) have reported that the finding of SEM as a natural constituent in red seaweed which can be detected after drying, suggests that there may be other natural sources of SEM. If seaweed or seaweed products that contained such background levels of SEM were to be used in animal feed, it is not expected that any detectable SEM would occur in products for human consumption that were derived from those production animals. For example, when broiler chickens were fed a diet containing 200 mg/kg SEM the level of tissuebound SEM in the muscle meat was 60 g/kg a transmission factor of just 0.03% (Kennedy & van Rhijn, 2003). For PES prepared using a bleaching step, for the 25 batches reported by Marinalg the range for SEM was 9 to 380 g/kg with a mean of 65 g/kg. Industry stated that carrageenan and PES are used interchangeably in many food additive applications, depending on customer requirements. Using the average reported value for SEM in PES and using the available information on maximum usage levels of this food additive, the SEM levels which could in principle be found in various foods have been calculated and are shown in the table. Table. Potential concentrations of SEM that may result from the use of additive E407a at the maximum usage levels and assuming an average contamination of 65 g/kg of the food additive Food Additive usage levels Beverages e.g. milk-based 0.02 to 0.04% additive used drinks Foods In the range 0.02 (ice cream) to 1% (jams, marmalade) Confectionery Up to 2% (e.g. gums) SEM concentration up to 0.026 g/kg up to 0.65 g/kg up to 1.3 g/kg
The ADI for carrageenan is 75 mg/kg bw/day (SCF, 2003b). If consumption was up to the full ADI and if all of this food additive contained SEM at 65 g/kg then the intake of SEM from this source could be up to 0.005 g/kg bw/day. JECFA (2002) summarised data on intake from Europe, Canada and the USA, noting that the resulting mean intake estimates were consistent, falling within a range of 30-50
mg/person/day from the use of carrageenan and PES as food additives. If all of this food additive was PES and contained SEM at 65 g/kg, then the intake of SEM from this source could be up to 0.003 g/person/day. SEM from azodicarbonamide-treated flour In Canadian tests, bread made using the ADC-treated flour contained 28 g/kg SEM. ADC is not permitted as a flour treatment agent in the EU and the importation of bread and bakery ware is likely to be very low. There is the potential for exposure from breaded animal products imported into the EU (e.g. frozen breaded chicken or fish products). Taking the upper figure of 5 g/kg of product from the study by Pereira et al (2004) this would give an intake of SEM of 1 g/person from a consumption of 200g of product. TOXICITY OF SEM Metabolism and toxicokinetics Most biochemical data on SEM concern its uses as an enzymatic (amine oxidase) inhibitor and as an aldehyde trapping agent. No information on its toxicokinetics or in vivo metabolism is available. In vitro, SEM was shown to be metabolised by rat liver homogenate to formaldehyde and N2 at a rate of 0.028 and 2.5 mol/g of liver per hr, respectively (Kroeger-Koepke et al., 1981). Acute toxicity Doses causing acute toxicity in mice (LD50s by oral, intraperitoneal (i.p.), subcutaneous and intravenous routes) ranged from 123-176 mg/kg bw (IARC, 1974). At high doses it caused convulsions (Jenney and Pfeiffer, 1958). A dose of 10 mg/kg bw was reported not to cause death in the rat (NTP Chemical Repository, 1991). The i.p. LD50 for rats was reported to be 212 mg/kg bw (de la Fuente del Rey 1986). Developmental toxicity Some studies have been reported on the effects of SEM hydrochloride on development in chick embryos (Shephard, 1989), however these are not considered relevant for human risk assessment. Oral studies. SEM was included in a study on lathyrogenic agents (Steffek et al., 1972). Lathyrism is a disease of collagen cross-linking caused by copper sequestration by nitriles. Groups of 3-11 female Sprague-Dawley rats were given SEM hydrochloride by aqueous, oral gavage at doses of 5, 10, 25, 50 or 100 mg/day on days 10-16 or 12-15 of gestation. The rats were killed one day before term and were examined for the presence of cleft palate. At the highest dose, 3 out
of 9 pregnant rats died. The incidence of resorptions (embryo-foetal deaths) was 3%, 0%, 3%, 38% and 56% and the incidence of cleft palate in surviving foetuses was 0%, 0%, 43%, 95% and 100% in 5, 10, 25, 50 or 100 mg/day dose groups respectively. Other lathyrogenic agents also caused cleft palate. From this study a no-effect level of 10 mg/day (equivalent to around 30-40 mg/kg bw/day) for cleft palate and intrauterine death was apparent. It should be noted that this was not a comprehensive study of developmental toxicity; dosing did not cover the whole of embryogenesis and the foetuses were not examined for internal or skeletal abnormalities. In a study on hamsters (Wiley and Joncja, 1978), groups of 5 or 6 pregnant animals were given 100, 150 or 200 mg/kg bw of SEM hydrochloride orally by gavage on gestation day 7. The animals were killed on gestation day 14 and the contents of the uterus examined. The two higher doses caused death of all the mothers, usually within 48h of dosing and all implants showed signs of resorptions. At 100 mg/kg bw, 1.3% of implants were dead or resorbed. Of the live fetuses, mean fetal weight was not significantly reduced, but 16.5% of fetuses showed growth retardation (defined as weighing <60% of untreated control mean body weight. There were no skeletal of visceral abnormalities, but 5.1% of fetuses were reported to have malposition of the fore-and hindlimbs. The significance of this observation in the absence of true malformations is difficult to interpret Other lathyrogenic treatments examined in the same study produced greater embryotoxicity and malformations including neural tube defects. Intraperitoneal studies It should be noted that the following studies used a route or administration (i.p.) which is regarded as inappropriate for risk assessment of oral exposure to chemicals because the uterus and its contents are directly exposed to high concentrations of chemical. It should also be noted that there was no information on maternal toxicity other than mortality, with no information given on the effect of the treatments on maternal weight gain. The studies are however described below for completeness. An early research study in rats (Karlson and Rosengren, 1959) used SEM to inhibit the enzyme histidine decarboxylase and thereby reduce the formation of histamine in the embryo and fetus. Pregnant rats were injected with SEM (presumably i.p. but not stated) at 50 mg/kg bw twice daily on gestation day 7 followed 75 mg/kg bw twice daily from gestation days 8-15. When this treatment was combined with a pyridoxine (vitamin B6) deficient maternal diet (pyridoxine being the coenzyme for histidine decarboxylase) then histamine production was reduced by 85% and all the fetuses died and were resorbed in late gestation. In controls given SEM alone there was no adverse effect on fetal survival. Another early study in rats (Kreybig, 1967) it was reported that SEM, given by i.p. injection on gestation days 10-13 at a dose of 140 mg/kg bw/day, had no adverse effect on embryonic development. A series of more comprehensive studies was conducted by de la Fuente and colleagues. Groups of Wistar rats were given a single i.p. injection of a dose of SEM hydrochloride, either 50, 75, 100 or 150 mg/kg bw in saline, on one of days 7, 10 or 13 of gestation (de la Fuente et al., 1983a). Some from each group (8-19 dams) was killed on gestation day 21 and the contents of the uterus examined. Fetuses were weighed and examined for gross abnormalities, skeletal and soft tissue abnormalities. The remaining animals were allowed to give birth and
rear their litters to 21 days of age. The numbers of live fetuses and fetal weights were significantly reduced in a dose-related manner compared to saline controls following all treatments except 50 mg/kg on day 13. Increased incidences of gross, skeletal and soft tissue abnormalities were found in all treated groups. These included anophthalmia, haemorrhages in the brain, liver and intestine, hydronephrosis, absent testes, incomplete ossification, absent sternum and rib abnormalities. Postnatally, there were significant, increases in offspring mortality in all treated groups, ranging from 25-63% compared with <5% in controls and peaking with treatment on gestation day 10. The female parents and 21-day-old offspring showed various significant reductions in DNA, RNA and/or protein levels in lung and liver, compared with controls. In a related study, de la Fuente et al. (1983b) showed similar effects on DNA, RNA and protein in 21-day old fetuses and 1, 7 and 30-day-old offspring following i.p. administration 100 mg SEM/kg bw on gestation day 10 to each of 3 successive generations of rats. The findings from both these studies are difficult to interpret in the absence of postnatal body weight data. In another study (de la Fuente del Rey, 1986), groups of 13-15 Wistar rats were given a single i.p. injection of a dose of SEM hydrochloride, either 50, 75, 100 or 150 mg/kg bw in saline, on one of days 5, 7, 10, 13 or 15 of gestation. Another group was injected throughout gestation with a dose of 17 mg/kg bw/day. Groups of 10 control rats received saline i.p. on the corresponding day or days of pregnancy. Approximately two-thirds of the treated rats and half the controls were killed on day 21 of gestation, the contents of the uterus examined, the fetuses removed, counted, weighed and examined for external and internal abnormalities and skeletal abnormalities. The remaining treated and control rats were allowed to deliver their litters and rear them for one month. There was a high incidence of maternal deaths at 150 mg/kg bw (21/64), with a few deaths at 100 mg/kg bw (5/50) and at 50 mg/kg bw (1/46). In those killed on day 21 of gestation, single injections of SEM caused significant reductions in numbers of live fetuses and in mean fetal weight at all doses tested and on nearly all days of gestation on which SEM was given. Injections given on day 7 or 10 were more embryotoxic than injections on days 5, 13 or 15. The effects were generally dose-related. In those treated throughout gestation with 17 mg/kg bw/day, there was a significant reduction in the mean number of implantations and in the mean number of live fetuses, but no effect on fetal weight. In those given single injections of SEM, the occurrence of anomalies and malformations was described according to the time of treatment but the information from all 4 dose levels was combined. Abnormalities seen in treated animals but not in controls were anophthalmia, cleft palate, severe liver haemorrhage, absent testes, incomplete ossification of the skull and limbs, absent sternum and rib malformation. The incidences of brain damage (hydrocephaly, exencephaly and meningocoele), severe intestinal haemorrhage, hydronephrosis and incomplete ossification of the sternum were increased in treated groups compared with controls. A similar pattern of abnormalities was seen in those injected with 17 mg/kg bw/day, with similar or lower incidences than those for all the other single-injection dose groups combined. In those animals allowed to litter out, postnatal mortality was significantly increased in all treated groups (between 10 and 60%) compared to controls (4-5%). The effect was dose-related. In the group treated throughout gestation with 17 mg/kg bw/day, the postnatal mortality of pups in the first month of life was over 50%. The author hypothesised that the high pup mortality in this study could be due either to congenital abnormalities or to a previously demonstrated effect of SEM in reducing the levels of phosphatidylcholine in lung surfactant (de la Fuente et al., 1983).
In vitro study In a study reported in abstract only (Simpson and Freeman, 1986), 9.5 day rat embryos in culture were exposed to SEM at concentrations of 10, 50, 100 or 250 g/ml for 48 h. There were no effects on growth or development at 10 g/ml, but at higher exposures there were concentration-related deficits in growth and differentiation, with haemorrhage in or absence of yolk sac development. These studies suggest a direct toxic and teratogenic effect of SEM on the embryo perhaps mediated via a disturbed yolk sac circulation. Carcinogenicity In a study by Mori et al. (1960) primarily aimed at evaluating the induction of pulmonary tumours in female dd mice with isonicotinic acid hydrazide a group of mice treated with SEM hydrochloride was included for comparison. Thirteen 6-8-week old female dd mice were fed a basal diet containing 0.1% SEM hydrochloride for 7 months (equivalent to approximately 150 mg/k bw.) It is stated that the animals gained weight on the diet. After the 7 months on the diet 6 of 8 survivors (75%) had developed 8 lung tumours compared to 1 tumour in 1 of 20 (5%) control animals. Groups of 50 male and 50 female Swiss albino mice, 6 weeks old (47 days) at the beginning of the experiment, were administered SEM hydrochloride at a concentration of 0.0625% (625 mg/L) in the drinking water for the rest of their life span. Based on the recordings of the average daily intake of drinking water the authors estimated the average daily intake of SEM hydrochloride to be 3.3 mg for females and 4.8 mg for males (130 mg/kg bw/day for females weighing 25 gram and 160 mg/kg bw/day for males weighing 30 gram). Groups of 100 male and 100 female mice, observed from weaning (5 weeks of age), served as controls. The lifespan was reduced in the treated animals, especially in the males. Complete necropsies were performed on all animals. All organs were examined macroscopically and histological examination was carried out on the liver, spleen, kidney, bladder, thyroid, hearth, pancreas, testis, brain, nasal turbinale, and lungs as well as on organs showing gross pathological changes. The occurrence of lung tumours and tumours of vascular origin were increased in treated females compared to controls. A total of 37 lung tumours occurred in 25 treated female mice (50%) compared to 31 lung tumours in 21 of the 100 control animals (21%). In the treated females 20 had 31 adenomas and 5 had 6 adenocarcinomas. In the treated males, 15 (30%) developed 21 lung tumours whereas 23 (23%) of control males developed 35 lung tumours. In the treated males 11 had 12 adenomas, 3 had 5 adenocarcinomas, and 1 had an adenoma and three adenocarcinomas. Tumours of vascular origin were observed in 9 treated females (18%) compared to 5 (5%) in the controls. The 9 vascular tumours in the treated females were 3 angiosarcomas and 2 angiomas in the liver, 2 angiosarcomas and 1 angioma in the ovaries, and 1 angioma in the lymph nodes. In the treated males, 3 animals (6%) developed vascular tumours compared to 6 (6%) of the control males. The incidences of a number of other tumours did not differ between treated and control mice. The authors also reported that induction of lung and blood vessel tumours in mice was a common finding during the testing of a number of substituted hydrazines in their laboratory (Toth et al. 1975). Toth and Shimizu (1974) has given a detailed description of the occurrence and morphology of the lung tumours in the control animals. In a meeting abstract, Ulland et al. (1973) reported that SEM was non-carcinogenic in Charles River CD rats administered the maximal tolerated dose and half that level for 18 months
followed by an observation period of 6 months. However, no further information were given except that SEM at the highest dose level had a lathyrogenic effect and that the animals suffered severe skeletal deformities. Weisburger et al. (1981) reported more details on the study of SEM hydrochloride in Charles River CD rats apparently mentioned by Ulland et al. (1973) in an abstract. SEM hydrochloride was among 14 different environmental and industrial chemicals that were tested for carcinogenicity in Charles River CD rats. Groups of 26 male and 26 female Charles River CD rats, 5-6 weeks old, were given diets containing SEM hydrochloride at concentrations of 500 or 1000 mg/kg (equivalent to 25 or 50 mg/kg bw/day, respectively). The authors considered the highest dose to be the maximal tolerated dose (MTD). A total of 184 animals per sex served as controls. The animals receiving the diet containing 500 mg SEM hydrochloride/kg were treated for 78 weeks. They thereafter received control diet and were observed for a further 26 weeks (a total of 104 weeks on test). In each of the groups receiving the diet containing 1000 mg SEM hydrochloride/kg the number of early deaths was so large that the treatment was discontinued at week 32. They therefore received control diet for the remainder of the study period of 104 weeks. The authors state that the survival was adequate in all the groups (at least 20 animals) at 78 weeks except for the high-dose males of which only 13 survived to 52 weeks. The mean weights of the treated animals were less than those of the matched controls. All animals were necropsied and an extensive tissue histopathology examination was carried out. The authors reported that SEM hydrochloride produced negative results as regards potential carcinogenic activity. However, no details were provided on the tumour incidences after treatment with SEM hydrochloride as well as the other compounds reported to be negative in the study, although detailed information was provided in this respect for the control animals and the chemicals reported to be carcinogenic. Osteolathyrism was diagnosed grossly commencing with the 10th week in all treated groups. Signs of the disorder were rough coat, protrusion of the sternum and bowing of the legs, and stiffness of the joints with bony growths. Histologically, osteoporosis was noted in the long bones. In conclusion, the authors stated that SEM hydrochloride had been tested at sufficiently high dose levels to cause toxic effects in both sexes and except for the high-dose males, survival was adequate in the treated groups. However, there was no indication of tumour induction from the treatment. A number of hydrazine and related substances were evaluated by IARC in 1974 whereas carbazides, including SEM was evaluated by IARC in 1976 (IARC 1974; 1976). In its update of IARC Monographs Volumes 1 to 42 IARC evaluated that the evidence for carcinogenicity is inadequate in humans and inadequate or limited in experimental animals and placed SEM hydrochloride in Group 3 (IARC 1987). Gold et al. (2004) in their Carcinogenic Potency Database have calculated TD50 values for the three statistically significant carcinogenic effects of SEM hydrochloride observed in the male and female mice in the study by Toth et al. (1975). The TD50 is defined as follows: For any particular sex, strain, species and set of experimental conditions, the TD50 is the dose rate (in mg/kg bw/day) that, if administered chronically for a standard period the standard lifespan of the species will halve the mortality-corrected estimate of the probability of remaining tumourless throughout that period. A TD50 of 223 mg/kg bw/day was calculated for lung tumours combined in females and of 833 mg/kg bw/day for males, whereas a TD50 of 395 mg/kg bw/day was estimated for blood vessel tumours combined in females. Cheeseman and co-workers (1999) have grouped 709 positive animal carcinogens in the Carcinogenic Potency Database according to structural class and used the TD50s to compare their potencies. In the structural group of hydrazines/triazenes/azides/azoxy compounds more than 30 substituted
hydrazines are included with TD50s ranging from 0.102 mg/kg bw/day for 1,2dimethylhydrazine hydrochloride to 561 mg/kg bw/day for 1,2-diformylhydrazine. Among the hydrazines tested, SEM hydrochloride was the fourth-least potent (TD50 of 223 mg/kg bw/day). Genotoxicity Bacterial mutagenicity assays A sample of commercial semicarbazide (carbamyl hydrazine, ICN-K&K Lab., purity and solvent unspecified) was tested in the Salmonella typhimurium reversion test using the plate incorporation method, with and without S9 from Aroclor induced rat liver and from Aroclor induced mouse liver and lung. In experiments without S9, SEM was weakly positive toward strain TA1535 (0.0009 induced revertants/nmol); the mutagenic response was partially or totally depressed in the presence of liver and lung S9, respectively. Negative results were obtained with strains TA1537, TA1538, TA98 and TA100, either with or without S9 (De Flora. (1981), De Flora et al. (1984). Negative results were obtained in a study using a modified test protocol with S.typhimurium strains G46, C3076, D3052, TA1535, TA1537, TA1538, TA98, TA100 and the E.coli strains WP2 and WP2 uvrA, with and without metabolic activation by Aroclor induced rat liver S9 (no experimental details given) (McMahon et al. 1979). A sample of pure SEM hydrochloride (99.7%, Sigma-Aldrich) was dissolved in water and tested in the plate incorporation test with S.typhimurium strains TA1535, TA1537, TA98, TA100 and E.coli WP2 uvrA at doses ranging from 62 to 5000 ug/plate, with and without exogenous metabolic activation by Aroclor induced rat liver S9. In the absence of S9, a clearcut mutagenic response was observed in strain TA1535, with a 16-fold increase in revertant colonies over control value at the highest dose applied. A two-fold increase in revertants colonies was also observed in strain TA1535 in the presence of S9. Borderline mutagenicity was also observed also in strain TA100, only in the absence of exogenous metabolism. No mutagenicity was observed in the other bacterial strains (TNO 2004a). A sample of pure (99.7%, Fluka) SEM hydrochloride and tested in the plate incorporation test with S.typhimurium strains TA1535, TA1537, TA98, TA100 and TA102 at doses ranging from 50 to 5000 ug/plate, with and without exogenous metabolic activation by Aroclor induced rat liver S9. A weak mutagenic response was observed toward the strain TA1535, only in the absence of S9, with two-fold increases in revertant colonies at 3000 ug/plate and above. No mutagenicity was observed in the other bacterial strains (Herbold, 2003). Mammalian cell gene mutation assay A sample of pure SEM hydrochloride (99.7%, Sigma-Aldrich) was dissolved in culture medium and tested in the forward mutation assay at the tk locus in mouse lymphoma cells L5178Y. The substance was applied at nine dose levels, from 0.21 to 10.0 mM (maximum recommended dose level), with and without metabolic activation. Treatments in the absence of S9 produced mild toxicity, with 50 % inhibition of growth at the highest dose. No toxicity was observed without S9. A dose related increase in mutant colonies, which exceeded twice the
spontaneous frequency at 10 mM, was observed in the experiment without S9. A borderline increase in mutant colonies was also observed at the highest dose in the presence of S9 (TNO 2004b). In vitro chromosomal aberration assays SEM hydrochloride (99.7 %, Bayer) was tested in a chromosomal aberration assay in Chinese hamster V79 cells. SEM was applied for 4 hours at 1120 ug/ml (10 mM, maximum recommended dose level) and at two lower doses (500 and 250 ug/ml), with and without S9. Cells were harvested after 18 hours and 30 hours (top dose only). In addition, a continuous treatment of 18 hours was applied without S9. Chromosomal aberrations were scored in 200 metaphases per dose (100 for each duplicated culture). Mild toxicity (about 20 % decrease of mitotic index and survival) was observed at the highest dose, only in the absence of S9. No increase in the frequency of aberrant metaphases and polyploidy was observed in any of the cultures treated with SEM compared with those treated with the vehicle only (deionized water) (Herbold, 2004). In another study, a sample of pure SEM hydrochloride (> 99.0 %, Sigma-Aldrich), dissolved in culture medium, was tested in a chromosomal aberration assay in CHO cells. A range finding experiment showed that the maximum recommended dose of 10 mM (1115 ug/ml) was exceedingly toxic. On this basis, in the first experiment the substance was applied at the doses of 150, 300 and 600 ug/ml for 4 hours, with and without S9, and cells were harvested 14 hours later; in the absence of S9 a continuous treatment of 18 hours was also applied. In the second test, 700, 800 or 900 ug SEM/ml were applied with continuous treatments of 18 or 32 hours without S9; with S9 cells were treated for 4 hours and harvested 14 or 28 hours later. Treatments produced mild toxicity, with 20-50 % reductions of mitotic index, and no significant increase in the incidence of cells with structural chromosomal aberrations. A significant increase of endoreduplicated cells was observed only after treatments in the presence of S9, in cells harvested at the early sampling time (TNO 2004c). The relevance of endoreduplication with respect to genotoxicity hazard identification is questionable, as this end-point is mainly consequence of alterations in cell cycle control rather that DNA damage. In vivo cytogenetics Semicarbazide hydrochloride (Sigma-Aldrich, Sweden) was tested in a flow cytometry-based micronucleus assay in vivo. Male CBA and Balb/C mice were injected intraperitoneally with 0, 40, 80 and 120 mg/kg bw SEM dissolved in phosphate buffer solution (maximum tolerated dose, determined in a toxicity pretest). Three or four animal per dose were used in the first experiment, with Balb/C mice, and five animal per dose (only 80 and 120 mg/kg bw) in the second experiment, performed with CBA mice. Blood samples were collected from each animal 42 hours after treatment. The incidence of micronuclei was determined in about 200,000 polychromatic erythrocytes per animal using a FACSVantage SE flow cytometer. Neither in CBA or Balb/C mice the treatment with SEM resulted in a frequency of micronucleated polychromatic erythrocytes (MnPCE) different from control values, nor did cell proliferation (% of PCE) show any supression. A distinct increase in MnPCE was induced by the positive control substance colchicine (Abramsson-Zetterberg and Svensson, 2005).
In a study in grasshopper germ cells, SEM hydrochloride was given by injection (5 ul/animal of 1 M solution in water). Increased incidences of chromatid and chromosome aberrations (breaks, dicentrics, etc) were observed in spermatocytes (overall incidence 0.3% in treated vs < 1 % in controls) (Bhattacharya 1976). The relevance of this study, based on an unvalidated test method in insects, for the assessment of the clastogenic/aneugenic potential of SEM in mammalian cells is limited. DNA damage in vitro/in vivo Negative results in a DNA repair test with SEM, using rat primary hepatocytes, are reported with no details in an abstract (Sugie et al. 1987). The DNA damaging activity of a series of hydrazine derivatives, including semicarbazide, was evaluated in vivo in mouse liver and lung using the alkaline elution technique. SEM (95-99% pure) was given according to the following treatment schedules to groups of 6 11 male Swiss mice: i) single i.p. injection of 3.28 nmol/kg b.w. (LD50) 6 h before sacrifice; ii) five daily i.p. injections of 0.55 nmol/kg b.w. Treatments did not induce any significant increase of DNA elution rate over controls, and SEM was evaluated as inactive (Parodi et al. 1981). Pure (>99 %, Fluka) semicarbazide hydrochloride was also tested in the in vivo/in vitro liver UDS assay. SEM was given to female CD-1 mice as a single oral dose of 100 or 200 mg/kg b.w. (MTD, based on clinical signs). Animals were sacrificed after 2 and 16 hours from treatment, and UDS was evaluated by autoradiography. No significant increase in net nuclear grains or in cells in repair was observed in hepatocytes isolated from animals treated with SEM (Central Toxicology Laboratory 2004). Effects on DNA in acellular systems SEM has been shown to specifically modify cytosine, replacing the C4 amino group with a semicarbazide residue. The reaction occurs at low pH (4.2), and with lower efficiency in double stranded DNA (Hayatsu et al. 1966; Hayatsu and Ukita 1966). Treatment of naked DNA with SEM (10- 100 M) and Cu (II) (20 M CuCl2), but not SEM alone, led to DNA strand breakage and base oxidation (8-oxodG) due to the formation by autooxidation of the carbamoyl radical CONH2 and ROS production (Hirakawa et al. 2003). Whether these reactions also can occur at physiologically free copper levels (< 10-6 uM), is not known. Summary and discussion of the toxicity studies None of the developmental toxicity studies available so far are comprehensive or conducted to current guidelines. In the hamster, a single oral dose of SEM hydrochloride of 100 mg/kg bw on gestation day 7 was not teratogenic but did cause minor growth retardation. Studies in rats showed that high doses of around 50-300 mg/kg bw/day in the rat, given either orally over several days or as single i.p. doses during gestation, cause maternal deaths, embryo-fetal deaths
and congenital abnormalities. Lower doses of 50 and 75 mg/kg bw given i.p. on single days of gestation also increased congenital abnormalities as did 17 mg/kg bw/day given i.p. throughout gestation. The relationship of these lower doses, below 100 mg/kg bw, to any maternal toxicity was not reported. Doses of 50 mg/kg bw and above given i.p. on single days during gestation also caused high postnatal mortality in the rat; a NOAEL for this effect has not been established. The no-effect level observed for induction of cleft palate in the rat (30-40 mg/kg bw/day orally) may not be applicable for other abnormalities. It can be noted however that the doses causing adverse effects are around three or more orders of magnitude above likely exposures of humans. None of the carcinogenicity studies available were conducted and reported to current guidelines. They show that SEM hydrochloride at a dose of approximately 130 160 mg/kg bw/day in the drinking water increased the incidence of lung tumours (adenomas + adenocarcinomas) in female and male mice and of blood vessel tumours in females. Induction of lung and blood vessel tumours is a common finding in mice following oral administration of substituted hydrazines. Comparison of the TD50s for a number of substituted hydrazines to be found in the Carcinogenic Potency Database reveals that SEM hydrochloride is among the least potent hydrazines tested for carcinogenicity so far. In rats, no carcinogenic effects were observed in either sex fed a diet containing 500 mg SEM hydrochloride/kg for 72 weeks (corresponding to 25 mg/kg bw/day; considered to be half the maximal tolerated dose) and observed for an additional 26 weeks. In the study, 14 different environmental and industrial chemicals were tested for carcinogenicity in rats and no detailed findings were reported for SEM hydrochloride and the other chemicals showing negative effects, whereas adequate information is given for the compounds found to be carcinogenic. At a higher concentration of 1000 mg/kg diet SEM hydrochloride was toxic to the rats and the treatment was discontinued after 32 weeks. Also in this group no carcinogenic effect was observed after 104 weeks. Literature data, and the results of recent studies performed to current standards with pure SEM hydrochloride, indicate that SEM is mutagenic in some test systems in vitro. Direct mutagenicity was observed, especially in the absence of an exogenous metabolising system, in bacterial strains detecting base pair substitutions and in a forward mutation test in mouse lymphoma cells. No clastogenic activity was observed in cytogenetic tests in vitro. A treatment related increase of endoreduplicated cells was observed in one study, but not confirmed in another investigation using a different cell line. The relevance of endoreduplication for genotoxic hazard identification is questionable. In vivo, the intraperitoneal administration of SEM to mice did not induce micronuclei in peripheral blood in a sensitive flow-cytometer-based micronucleus assay, nor DNA damage, as measured by alkaline elution, in liver and lung, whereas positive results were obtained with known genotoxic hydrazine derivatives. Negative results were also obtained in the liver UDS assay after oral administration of SEM to in female CD-1 mice, the mouse strain where long term administration of SEM resulted in increased incidence of lung and vascular tumours (IARC, 1976).
CONCLUSIONS The occurrence and the precursors of SEM in foods SEM has been found to occur in different types of foods depending on the source of this substance. SEM is a metabolite of the veterinary medicine nitrofurazone but since the use of nitrofurazone is illegal in the EU, SEM from this source should not be detectable in foods. SEM arises in foods, and especially baby foods, as a result of migration from sealing gaskets used in the metal lids of jars and bottles. In this case, the origin of the SEM is thermal breakdown of azodicarbonamide, the blowing agent used to foam the plastic gaskets. These first 2 potential sources, nitrofurazone use and azodicarbonamide-blown gaskets, are the best documented with independent verification. Other sources have been suggested too. SEM has been found in food products made using flour in which azodicarbonamide has been added as a dough-improver; a practice that is not permitted in the EU. SEM is also reportedly formed as a reaction product of the action of hypochlorite on food additives such as carrageenan and on foods such as egg white powder. Finally, SEM may be present at background levels naturally, made be formed at low levels when some foods are dried, and may also derive from as yet unidentified sources. The method of analysis used for SEM in foods The method of analysis used to test foods for SEM involves acid hydrolysis and a derivatisation step with 2-nitrobenzaldehyde (2-NBA). The SEM-NBA derivative is then determined using LC-MS/MS (liquid chromatography coupled to tandem mass spectrometry). The detection limit of the method is in the region of 0.2 g/kg. This approach is used worldwide and it was developed for the testing of animal products using SEM as a marker for illegal nitrofurazone use. The performance of the method meets the requirements of Commission Decision 2002/657/EC, Article 4, on minimum performance criteria for quantitative analytical methods for toxicants in foods. In animals treated with nitrofurazone, the veterinary drug is metabolised rapidly to SEM which becomes covalently tissue-bound. The acid hydrolysis step is used to liberate the bound residue for analysis. The analysis method therefore measures total SEM (free and bound) in the sample. There is evidence that when SEM originates from the other sources described above, a fraction may also become bound to the matrix. However, the acid hydrolysis conditions used in the analytical method are not dissimilar to normal gastric conditions. Tissue-bound residues of SEM are reported to be bioavailable in the digestive systems of mammalian species (Hoogenboom et al. 1992; Gottschall & Wang 1995; McCracken et al. 1997). Therefore, in this evaluation there is no distinction made between free or bound SEM and the total result from the 2-NBA-LC-MS/MS method is used. Similarly, given that a number of chemical source of SEM have been identified and that there may still be as-yet unidentified sources and mechanisms of formation, it is possible that the acid hydrolysis step used in the test method may form low levels of SEM from labile precursors. Again, while the current state of knowledge is incomplete, in this evaluation there is no distinction made between SEM that is present as such in a food sample and any SEM that may have been formed from precursors in the food under acidic conditions used in the analysis. It is concluded that the method of testing for SEM provides concentration data that are suitable for this risk evaluation.
Estimates of exposure On the basis of the information available, migration of SEM from gaskets is by far the largest source of exposure known. The concentration data supplied by for different countries were similar. Taking a conservative scenario for a 9-month old infant with a body weight of 8.8 kg and eating exclusively food and drink from glass jars and bottles contaminated with SEM at an average concentration of 13 g/kg, the intake of SEM would be 0.35 g/kg bw/day for an average (consumers only) consumption of 234 gram per day and the intake would be 0.69 g/kg bw/day for consumption at the 95th percentile of 464 gram per day. For an infant with a body weight of 4.5 kg and fed exclusively each day with 700 mL of infant milk pre-packaged in glass bottles with metal lids, containing SEM at an average concentration of 9 g/kg, the intake would be 1.4 g/kg bw/day. Adult exposures to SEM are likely to be much lower than infant exposures, due to the lower contribution of foods packaged in bottles and jars to the total diet of adults, the lower contamination levels derived from the smaller gasket areas involved for that packaging, and the higher adult body weight. Taking an assumption that 1 kg of food contaminated with SEM at an average concentration of 1.0 g/kg is consumed each day, the exposure for a 60 kg bodyweight adult would be 0.02 g/kg bw/day. Commission Directive 2004/1/EC prohibits the use of ADC in food contact materials from 2nd August 2005 and so, once existing stocks of packaged foods are used up, exposure of consumers by this route will have been eliminated. The Panel was informed that industry is making significant progress on the development of new seal technology and expects to be able to meet the date of August 2005 for the ban on the use of azodicarbonamide in food contact materials. Other possible sources of SEM in foods contribute far less to exposure than estimated above for packaging. Bread made using flour treated with azodicarbonamide can contain SEM. In laboratory tests the SEM concentration in bread was 28 g/kg. ADC is not permitted as a flour treatment agent in the EU and the importation of bread and bakery ware is likely to be very low. There is the potential for exposure from breaded animal products imported into the EU (e.g. frozen breaded chicken or fish products). Taking the upper figure of 5 g/kg of product from the study by Pereira et al (2004) this would give an intake of SEM of 1 g/person from a consumption of 200g of product. For a high consumer of egg products that may be contaminated by 50 g/kg SEM as a result of using hypochlorite as a sanitising solution on production equipment, a reasonable worst-case estimate of exposure is 0.008 g/kg bw/day. For carrageenan food additive that may likewise become contaminated with SEM, if consumption was up to the full ADI and if all of this food additive contained SEM at 65 g/kg then the intake of SEM from this source could be up to 0.005 g/kg bw/day. The Panel notes that SEM may be formed at low levels when some foods are dried. SEM, may also derive from as yet unidentified sources, and SEM may be present at very low background levels naturally.
Toxicity The only new toxicological information that has become available since EFSA issued its preliminary advice on SEM in July and October 2003 concerns genotoxicity. There are no new data on carcinogenicity, reproductive and developmental toxicity, or other aspects of repeatdose toxicity and the available data on all these aspects are not as comprehensive as the Panel would wish. While these data deficiencies add some uncertainties to the risk assessment, given the generally low exposures to SEM via food, it is the aspects of genotoxicity and carcinogenicity which remain the focus of this assessment. SEM has been shown to be carcinogenic in mice, but not rats. Literature data on genotoxicity together with the results of recent studies with pure SEM hydrochloride, indicate that SEM is mutagenic but not clastogenic in some test systems in vitro, notably in the absence of an exogenous metabolising system. In vivo, negative results were reported in adequately performed studies on DNA damage in liver and lung of mice, and in the micronucleus assay in the mouse. Based on the overall weight of evidence provided by the studies performed, which included a study using a highly sensitive methodology, it is concluded that the weak genotoxicity exerted by SEM in vitro is not expressed in vivo. The new data allaying the concern on genotoxicity in vivo, and the likely reductions in exposure following replacement of the most significant, currently known source of SEM in the diet (gaskets on glass jars and bottles), offer further support to the preliminary advice given by EFSA in 2003 that the risk, if any, from consumption of foods containing SEM is judged to be very small, not only for adult consumers but also for infants. In this respect the Panel noted that SEM is a weak non-genotoxic carcinogen for which a threshold mechanism can be assumed. A large margin of at least 5 orders of magnitude exists between the dose causing tumours in experimental animals and human exposure, including that of infants. The Panel therefore concludes that the issue of carcinogenicity is not of concern for human health at the concentrations of SEM encountered in food.
INFORMATION PROVIDED TO EFSA From Industry Communication Possible contaminant in gaskets of metal lids from industry to the European Food Safety Authority, dated 25 June 2003 and. Communicated to EFSA on 8 July 2003. Semicarbazide in food, submitted November 2003. In vivo mouse liver unscheduled DNA synthesis assay, Danone In vitro chromosome aberration test with Chinese Hamster V79 cell, submitted by Industry. Formation of semicarbazide in seaweeds used as food additives. Submitted by Marinalg. From Member States: Belgium: Semicarbazide in powdered eggs Finland: Semicarbazide in baby food and some other items France: Semicarbazide in baby food Germany: Semicarbazide in baby food Ireland: Semicarbazide in baby food Netherlands Semicarbazide in baby food Norway: Semicarbazide in carrageenan Spain: Semicarbazide in baby food REFERENCES Abramsson-Zetterberg,L., and Svensson, K. (2005) Semicarbazide is not genotoxic in the flow cytometry-based micronucleus assay in vivo. Toxicol. Lett. 155: 211-217 Bhattacharya, A.K. (1976) Chromosome damage induced by semicarbazide in spermatocytes of a grasshopper. Mutat. Res. 40: 237-242 Becalski, A., Lau, B. P-Y., Lewis, D. & Seaman, S. W. (2004) Semicarbazide formation in azodicarbonamide-treated flour: a model study. J. Ag. Fd. Chem., 52, 6730-5734. Bernard, A. (2003). Nitrofurazone residues in Belova egg products. Evaluation of the health risks. Statement of April 5 2003. Central Toxicology Laboratory (2004) Semicarbazide: in vivo mouse liver unscheduled DNA synthesis assay. CTL study n. SM1207 Cheeseman, M.A., Machuga, E.J. and Bailey, A.B. (1999). A tiered approach to threshold of Regulation. Fd Chem. Toxicol., 37, 387-412. De Flora S. (1981) Study of 106 organic and inorganic compounds in the Salmonella/microsome test. Carcinogenesis 2: 283-298 De Flora S., Zanacchi P., Camoirano A., Bennicelli C., Badolati G. (1984) Genotoxic activity and potency of 135 compounds in the Ames reversion test and in a bacterial DNA-repair test. Mutat. Res. 133: 161-198 De la Fuente, M., Hernanz, A., Alia, M. (1983a). Effect of semicarbazide on the perinatal development of the rat : changes in DNA, RNA and protein content. Methods and Findings in Exptl.Clin.Pharmacol. 5, 287-297. De la Fuente, M., Hernanz, A., Alia, M. (1983b). Effect of semicarbazide on DNA, RNA and protein hepatic levels of four subsequent generations of Wistar rats. Comparative Biochemistry and Physiology 74C, 447-450.
De la Fuente, M., Hernanz, A., Alia, M (1983c). Effect of semicarbazide on the surfactant phospholipid percentages during fetal and postnatal lung development in rat. Comparative Biochemistry and Physiology 74, 115-118. De la Fuente del Rey, M. (1986). Teratogenic effect of semicarbazide in Wistar rats. Biology of the Neonate 49, 150-157. EC (1990) Article 5 of the Council Regulation 2377/90 laying down a Community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin and article 6 of the Directive 2001/82/EC on the Community code relating to veterinary medicinal products. EFSA/AFC/SEM/16. Semicarbazide in powdered egg. Belgium analytical data.
EFSA (2004). Opinion of the Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC) on a request from the Commission related to the use of Epoxidised soybean oil in food contact materials. The EFSA Journal, 64, 1-17.
Even, I, Berta, J.L., and Volatier, J.L. (2002). Evaluation de lexposition thorique des nourrissons et enfants en bas ge aux rsidus de pesticides apports par les aliments courants et infantiles, AFFSA, January 2002. Gold, T.S., Slone, T.H. and Manley, N.B. (2004). Condensed Carcinogenic Potency Database available on the Internet at http://potency.berkeley.edu/. Gatermann, R., Hach, P., Hartig, L., Hoenicke, K., Mandix, M., and Otte, S. (2003) Formation of semicarbazide in food by hypochlorite treatment. Poster Presentation at the 1st International Symposium on Recent Advances in Food Analysis, Prague, Czech Republic. Gatermann, R., Hoenicke, K. & Mandix, M. (2004) Formation of semicarbazide in natural compounds in foods by heat treatment. Czech J. Fd. Sci., 22, 353-354. Gottschall, D. & Wang, R. (1995) Depletion and bioavailability of 14C-furazolidone residues in swine tissues. J. Ag. Fd. Chem., 43, 2520-2525. Hayatsu U., Takeishi K.I., Ukita T. (1966) The modification of nucleosides and nucleotides. 3. A selective modification of cytidine with semicarbazide. Biochim. Biophys. Acta 123: 445457 Hayatsu U., Ukita T. (1966) The modification of nucleosides and nucleotides. 4. The reaction of semicarbazide with nucleic acids. Biochim Biophys Acta 123: 458-470 Herbold B. (2003) Semicarbazide hydrochloride Salmonella/microsome test plate incorporation method. Bayer HealthCare Study n. T 5072934, September 8th, 2003 Herbold B. (2004) Semicarbazide hydrochloride. In vitro chromosome aberration test with Chinese hamster V79 cells. Bayer HealthCare Study n. AT01020, February 24th, 2004 Hirakawa K., Midorikawa K., Oikawaa S., Kawanishi S. (2003) Carcinogenic semicarbazide induces sequence-specific DNA damage through the generation of reactive oxygen species and the derived organic radicals. Mutat. Res. 536: 91-101 Hoenicke. K., Gatermann, R., Hartig, L., Mandix, M. & Otte, S. (2004) Formation of semicarbazide (SEM) in food by hypochlorite treatment: is SEM a specific marker for nitrofurazone abuse? Fd. Add. & Contamin., 21, 526537. Hoogenboom, L., Van Kammen, M., Berghmans, M., Keoman, J. & Kuiper, H. (1991) The use of pig hepatocytes to study the nature of protein-bound metabolites of furazolidone: a new analytical method for their detection. Fd. Chem. Toxic., 29, 321-328.
IARC (1974). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man: Some Aromatic Amines, Hydrazine and Related Substances, N-Nitroso Compounds and Miscellaneous Alkylating Agents. Volume 4. International Agency for Research on Cancer, Lyon. IARC (1976). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man: Some Carbamates, Thiocarbamates and Carbazides. Volume 12. International Agency for Research on Cancer, Lyon, France, pp 209 - 214. IARC (1987). IARC Monographs on the Evaluation of the Carcinogenic Risk to Humans: Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. Supplement 7. International Agency for Research on Cancer, Lyon, France, pp 71. JECFA (2002). Carrageenan and processed Eucheuma seaweed (addendum). Safety evaluation of certain food additives and contaminants. WHO Food Additives Series: 48, pp 91-101. Prepared by the 57th Meeting of the Joint FAO/WHO Expert Committee on Food Additives. World Health Organization, Geneva. Jenney, E.H. and Pfeiffer, C.C. (1958). The convulsant effect of hydrazides and the antidotal effect of anticonvulsants and metabolites. Journal of Pharmacology and Experimental Therapeutics 122, 110-123. Karlson, G., Rosengren, E. (1959). Prevention of foetal development by enzyme inhibition. Nature 184, 1238-1239. Kennedy, D. G. & van Rhijn, J. A. (2003) http\\www.rikilt.wur.nl, July 2003. Nitrofurazone and semicarbazide,
Kersting M., Alexy U., Sichert-Hellert W., Manz F. and Schch G. (1998). Measured consumption of commercial infant food products in German infants: Results from the DONALD study. J. of Pediatric Gastroenterology and Nutrition 27: 547-552. Kreybig, Th. (1967). Chemische Konstitution und teratogene Wirkung in einigen Verbindungsgruppen. Naunyn-Scmiedebergs Arch.Pharmakol.Exp.Pathol. 257, 296-298. Kroeger-Koepke, M.B., Koepke S.R., McClusky, G.A., Magee, P.N., Michejda, C.J. (1981) Hydroxylation pathway in the in vitro metabolism of carcinogenic nitrosamines: NNitrosodimethylamine and N-nitroso-N-methylaniline. Proc. Natl. Acad. Sci. U.S.A. 78, 6489-6493. Mills, A., Tyler, H. (1992) Food and nutrient intakes of British Infants aged 6-12 months. The Stationery Office, London. McCracken, R. & Kennedy, D. (1997) The bioavailability of residues of the furazolidone metabolite 3-amino-2-oxazolidone in porcine tissues and the effect of cooking on residue concentrations. Fd. Add. Contam., 14, 507-513. McMahon R.E., Cline J.C., Thompson C.Z. (1979) Assay of 855 test chemicals in ten tester strains using a new modification of the Ames test for bacterial mutagens. Cancer Res. 39: 682-693 Mori, K., Yasuno, A. and Matsumoto, K. (1960). Induction of pulmonary tumors in mice with isonicotinic acid hydrazid. Gann, 51, 83-89. New Jersey DHSS (1999). Semicarbazide hydrochloride. Hazardous Substance Fact Sheet. New Jersey Department of Health and Senior Services, USA. December 1999.
NTP Chemical Repository (1991). Semicarbazide hydrochloride (Radian Corporation, August 29, 1991). Available at: http://ntp-db.niehs.nih.gov/NTP_Reports/NTP_Chem_H&S/NTP_Chem5/Radian563-41-7.txt Parodi S., De Flora S., Cavanna M., Pino A., Robbiano L., Bennicelli C., Brambilla G. (1981) DNA-damaging activity in vivo and bacterial mutagenicity of sixteen hydrazine derivatives as related quantitatively to their carcinogenicity. Cancer Res. 41: 1469-1482 Pereira, A. S., Donato, J. L. & De Nucci, G. (2004) Implication for the use of semicarbazide as a metabolic target of the nitrofurazone contamination in coated products. Fd. Add. Contamin., 21, 63-69. Perharic, L., Zorie, A., Golja, V. and Dekleva, N. (2004). Semicarbazide in baby foods in Slovenia. Poster presentation at Life Science 2004, Nova Govica, Solvenia, 18-22 September 2004. RASFF (2003). Additional Information - Rapid alert November 2003/086. Saari, L. & Peltonen, K. (2004) Novel source of semicarbazide: levels of semicarbazide in cooked crayfish samples determined by LC/MS/MS. Fd. Add. Contam., 21, 825-832. SCF (2002). Opinion of the Scientific Committee on Food on Bisphenol A. Document SCF/CS/PM/3936 Final. 3 May 2002. (Expressed on 17 April 2002). SCF (2003a). Opinion of the Scientific Committee on Food on Carrageenan. Document SCF/CS/ADD/EMU/199 Final. 21 February 2003. (expressed on 5 March 2003). SCF (2003b). Opinion of the Scientific Committee on Food on the 23rd additional list of monomers and additives for food contact materials. Expressed on the 4th April 2003. Available at: http://europa.eu.int/comm/food/fs/sc/scf/out181_en.pdf Shephard, T.H. (1989). Catalog of Teratogenic Agents. 6th Edition. The Johns Hopkins University Press, Baltimore and London. Simpson M.N., Freeman S.J. (1986). The effects of the lathyrogen semicarbazide on the development of rat embryos cultures in vitro . Human Toxicology 5, 408-409. Steffek, A.J., Verrusio, C. & Watkins, C.A. (1972). Cleft palate in rodents after maternal treatment with various lathyrogenic agents. Teratology 5, 33-40. Sugie S., Yoshimi N., Mori H., Shimizu H. (1987) Genotoxicity of hydrazine derivatives in the hepatocyte primary culture/DNA repair test using rat hepatocytes. Mutat. Res. 182: 379 (abstract) TNO (2004a) Bacterial reverse mutation test with semicarbazide. TNO Report n. V5005/11, 6 February 2004 TNO (2004b) Gene mutation test at the TK locus of L5178Y cells with Semicarbazide. TNO Report n. V5003/09, 22 January 2004 TNO (2004c) Chromosomal aberration test with Semicarbazide in Chinese hamster Ovary (CHO) cells. TNO Report n. V 5002/09, 9 February 2004 Toth, B. and Shimizu, H. (1974). 1-Carbamyl-2-phenylhydrazine tumorigenesis in Swiss mice. Morphology of lung adenomas. J. Natl. Cancer Inst., 52, 241-251. Toth, B., Shimizu, H. and Erickson, J. (1975). Carbamylhydrazine hydrochloride as a lung and blood vessel tumour inducer in Swiss mice. Europ. J. Cancer, 11, 17-22.
Ulland, B., Weisburger, E.K. and Weisburger, J.H. (1973). Chronic toxicity and carcinogenicity of industrial chemicals and pesticides. Toxicol. Appl. Pharmacol., 25, 446. van Rhijn, J. A., Kennedy, D. G., Mulder, P.P.J., Castle, L., Gatermann, R. & Bock, C. (2004) Record of the special workshop held on Nitrofurans. Held in connection with the Euroresidues V Conference, Holland, on 12th May 2004. Weisburger, E.K., Ulland, B.M., Nam, J., Gart, J.J. and Weisburger, J.H. (1981). Carcinogenicity tests of certain environmental and industrial chemicals. J. Natl. Cancer Inst., 67, 75-88. Wiley MJ and Joncja MG (1978). Neural tube lesions in the offspring of hamsters given single oral doses of lathyrogens early in gestation. Acta Anat. 100, 347-353.
SCIENTIFIC PANEL MEMBERS R. Anton, S. Barlow, D. Boskou, L. Castle, R. Crebelli, W. Dekant, K.-H. Engel, S. Forsythe, W. Grunow, M. Heinonen, J.-C. Larsen, C. Leclercq, W. Mennes, M.-R. Milana, I. Pratt, I. Rietjens, K. Svensson, P. Tobback, F. Toldr.
ANNEX to Semicarbazide in food
Annex 1
The EFSA advice on baby foods from October 2003 Exposure data The ad hoc expert group received additional information from industry on the range of concentrations of semicarbazide in foods packed in glass jars and bottles and on consumption of the relevant foods by infants of 4-12 months of age. The number of food samples analysed for semicarbazide is still limited, but confirm that amounts ranging up to 25 g/kg have been found in baby foods. The additional information confirms that the preliminary estimate of exposure of infants to semicarbazide, contained in the AFC Panel statement of 1 October 2003, of around 2 microgrammes/kg body weight/day is a reasonable worst case. Toxicological aspects No additional information on toxicological aspects was presented to the ad hoc expert group since the update provided by the AFC Panel on 1 October 2003. The Panel commented that semicarbazide has weak genotoxic activity in vitro and weak carcinogenic activity in female but not male mice, but because of the limited evidence, it is not possible to conclude whether SEM may pose a carcinogenic risk to humans. There is no risk of immediate illness to adults, children or infants from consumption of foods containing SEM. The concern relates to health in the long term because of the possibility that SEM may cause cancer. The risk, if any, is judged to be very small, both for infants and adult consumers. Although there are uncertainties in the risk assessment due to limitations in the data, these relate only to how to characterise a very small risk. Further details on toxicological aspects can be found in the advice issued by EFSA on 28 July 2003 and in the recent statement of the AFC Panel dated 1 October 2003. Microbiological aspects Jars of baby food have an excellent microbiological safety record. Additional assurance for the consumer about the integrity of the seal is provided by the incorporation of the pop-up button. Current manufacturing practice for baby foods in jars includes controlled heat treatment to ensure that the product is microbiologically safe for the consumer. The airtight seal maintains microbiological stability and ensures a long shelf-life. It prevents contamination with microorganisms, dirt, or insects, during storage and protects the nutritional integrity of the product. If alternative technologies are going to be used to reduce or eliminate SEM from baby foods, it is crucial that the same rigorous microbiological standards are upheld. The ad hoc expert group was informed that industry is making significant progress on the development of new seal technology. Any future changes to the current gasket would need careful consideration and evaluation of seal integrity before being launched as a replacement for the existing technology. Nutritional aspects Processed baby foods are products which are carefully regulated with respect to ingredients, low levels of nitrate, pesticide residues and contaminants, and baby foods in glass jars and bottles make an important nutritional contribution to the diets of many infants. Although not an obligatory part of the diet of infants, they are widely used primarily for reasons of convenience,
ANNEX to Semicarbazide in food
quality and nutritional safety. If these products were to become suddenly unavailable, or caregivers of infants chose not to use them any longer, it should be borne in mind that the nutritional adequacy of the diets of those infants could be compromised, particularly where caregivers have been relying extensively on these products. There is a possibility that the foods which would be chosen instead might be either not nutritionally equivalent to the foods used before or not suitable for the feeding of infants.
European Food Safety Authority

Brussels, 28 July 2003 AFC/adhoc SEM/1
Advice of the ad hoc expert group set up to advise the European Food Safety Authority (EFSA) on the possible occurrence of semicarbazide in packaged foods
PRELIMINARY ADVICE ON THE POSSIBLE OCCURRENCE OF SEMICARBAZIDE IN PACKAGED FOODS
Background The European Food Safety Authority has been informed by industry1 that semicarbazide has been found in a number of food products from different manufacturers that are packaged in glass jars with metal lids sealed with plastic PVC gaskets. Examples of the foods involved include some types of baby food, fruit juices, jams and conserves, honey, pickles and sterilized vegetables, mayonnaise, mustard, sauces and ketchups.
The presence of semicarbazide has recently been linked by industry to the permitted use of azodicarbonamide as a blowing agent to make a foamed plastic which is suitable for gaskets, but this link has not yet been clearly established. The amounts that have been detected in foods are variable but low, ranging from not detectable up to 20 microgrammes per kg of food (20 ppb). Levels of 1 - 7 mg semicarbazide per kg of gasket material have been detected in extracts of the gaskets themselves.1
An ad hoc expert group of the EFSA Scientific Panel on Additives, Flavourings, Processing Aids and Food Contact Materials (AFC Panel) has been asked to advise on the significance of these findings for human health. Its preliminary advice is given below.
AFC/adhoc SEM/1
Possible origin of semicarbazide in food packaging materials Azodicarbonamide is structurally related to semicarbazide as shown below. Azodicarbonamide Semicarbazide H2N-CO-N=N-CO-NH2 H2N-NH-CO-NH2
Azodicarbonamide is authorized for use in the EU as a blowing agent for plastics in contact with food. Blowing agents are added to polymers during processing to form minute gas cells throughout the plastic. During high temperature processing, azodicarbonamide decomposes to form gases, mainly nitrogen, together with carbon monoxide, carbon dioxide and ammonia, and non-volatile residues such as biurea. Semicarbazide has recently been detected in food contact materials made using azodicarbonamide.1 Such materials are typically used in sealing gaskets for metal lids on glass bottles and jars.
Azodicarbonamide was recently re-evaluated by the EC Scientific Committee on Food when it was confirmed as acceptable for use in food contact materials, but only as a blowing agent and not for materials in contact with alcoholic beverages. It was classified in List 3 (substances for which an ADI or a TDI could not be established, but where the present use could be accepted).2
Semicarbazide is not specifically regulated by EU food packaging directives but if it were present in food packaging materials, for instance as an impurity or a reaction or degradation product, its presence in food would be covered by the Council Directive 89/109/EEC.3 Under Article 2 of this Directive, it could be present in food contact
AFC/adhoc SEM/1
materials provided it did not transfer into foodstuffs in quantities which could endanger human health.
Although semicarbazide has been detected in foods packaged in glass bottles and jars sealed with lids containing foamed plastic gaskets and in the gaskets themselves, it is not yet clear whether semicarbazide itself is present in the gaskets and the food, or is formed from a precursor during the analytical procedure. The expert group understands that the industry is currently working to clarify this issue as soon as possible.
Other possible origins of semicarbazide in food Semicarbazide is one of the breakdown products of veterinary drugs known as the nitrofurans. The presence of residues of nitrofurans and/or their breakdown products in foods is illegal in the EU.4 The foods in which semicarbazide has been found in the recent tests include foods not of animal origin and its presence is unlikely to be explained by nitrofurans use. Azodicarbonamide is also used as an additive in some non-EU countries for the treatment of flour used in bread-making. It is not permitted for this use in the EU.
Toxicity of semicarbazide Semicarbazide has not been extensively tested for toxic effects. However, it belongs to a family of chemicals (hydrazines) which are known to cause cancer in animals5 and has itself been shown to cause cancer in female but not male mice when given via the diet or via drinking water.
6-9
The types of tumour found with semicarbazide were lung and
vascular tumours, which have also been found with other hydrazines. However, semicarbazide was found to be one of the least potent carcinogens among several
AFC/adhoc SEM/1
hydrazines.5,10,11 A study in rats given semicarbazide in the diet did not show treatmentrelated tumours, but the study was potentially flawed due to poor survival.9,12,13
In acellular systems, semicarbazide has been shown to bind to the cytosine residues of RNA, to the deoxycytosine residues of DNA, and to cytosine and deoxycytosine nucleosides.14,15 It has also been shown to induce chromosome damage in spermatocytes of grasshopper.16 However, it showed little or no mutagenic activity in the more
commonly used Ames test in Salmonella.10 Moreover it was negative in studies of DNA damage, as measured by alkaline elution in liver and lung tissue of mice.10 Recently, it has been shown to cause DNA damage at the thymine and cytosine residues of DNA fragments obtained from the human p53 tumour suppressor gene and the c-Ha-ras-1 protooncogene but only in the presence of copper (Cu(II)) in vitro.17 The amount of DNA damage observed increased with increasing concentration of semicarbazide (10-100M). While copper is known to be present in the chromatin in cell nuclei, it is normally tightly bound and not available in ionic form. Thus the significance of these in vitro results for the in vivo situation is unclear. The study also investigated the possible mechanism(s) for the observed sequence-specific damage to DNA. The results suggested that it was oxidative damage caused by reactive oxygen species formed from hydrogen peroxide and Cu(I) or by carbamoyl (CONH2) radicals. Overall, the available data are insufficient for the evaluation of the genotoxicity of semicarbazide. In particular, no data are available on the induction of gene mutations or chromosome aberrations in mammalian cells.
Semicarbazide has been shown to cause embryo-fetal death and cleft palate in surviving fetuses in rats given high doses by oral gavage during pregnancy.18 These effects were
AFC/adhoc SEM/1
seen at doses of 25-100 mg/day; no adverse effects were seen at doses of 5 or 10 mg/day. The highest dose (100 mg/day) also caused death of some of the pregnant rats. The fetuses were only examined for gross malformations and were not subjected to a full teratological examination. Malformations of the brain and kidney, and haemorrhages have been reported in rats given semicarbazide by subcutaneous injection at a dose of 17 mg/kg.19
Conclusions There is an urgent need first to establish whether the semicarbazide detected in food packaging gaskets and in foods in contact with those gaskets is actually present in foods as consumed, or whether the detection of semicarbazide is a result of the method of analysis used. If its presence in food is proven then a better understanding of the conditions under which semicarbazide is formed would be helpful for risk management.
The toxicological data available at present are very limited. Two studies have indicated that it may cause cancer in one sex of one species of animal. There is some indication that semicarbazide is genotoxic but the data are insufficient to draw any firm conclusion. These are the main issues in the safety evaluation of semicarbazide. The production of birth defects occurs only at exposure levels that are very high compared to the maximum levels of semicarbazide detected so far in foods and would not be a matter for concern.
In view of the uncertainties in both the analytical and toxicological aspects it is not possible at present to give scientifically based risk assessment advice on the possible occurrence of semicarbazide in foods in contact with food packaging gaskets.
AFC/adhoc SEM/1
References 1. Communication Possible contaminant in gaskets of metal lids from industry to the European Food Safety Authority, dated 25 June 2003. Communicated to EFSA on 8 July 2003. 2. SCF (2003). Opinion of the Scientific Committee on Food on the 23rd additional list of monomers and additives for food contact materials. Expressed on the 4th April 2003. Available at: http://europa.int/comm/food/fs/sc/scf/outcome_en.html 3. Council Directive 89/109/EEC (1989). Council Directive of 21 December 1988 on the approximation of the laws of the Member States relating to materials and articles intended to come into contact with foodstuffs (89/109/EEC). Official Journal of the European Communities No. L 40, 11.2.89, 38-44. 4. Article 5 of the Council Regulation 2377/90 laying down a Community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin and article 6 of the Directive 2001/82/EC on the Community code relating to veterinary medicinal products. 5. IARC (1974). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man: Some Aromatic Amines, Hydrazine and Related Substances, NNitroso Compounds and Miscellaneous Alkylating Agents. Volume 4. International Agency for Research on Cancer, Lyon. 6. Mori K, Yasuno A & Matsumoto K (1960). Induction of pulmonary tumors in mice with isonicotinic acid hydrazid. Gann 51, 83-89. 7. Toth B & Shimizu H (1974). 1-Carbamyl-2-phenylhydrazine tumorigenesis in Swiss mice. Morphology of lung adenomas. Journal of the National Cancer Institute 52, 241251. 8. Toth B, Shimizu H & Erickson J (1975). Carbamylhydrazine hydrochloride as a lung and blood vessel tumour inducer in Swiss mice. European Journal of Cancer 11, 17-22. 9. IARC (1976). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man: Some Carbamates, Thiocarbamates and Carbazides. Volume 12. International Agency for Research on Cancer, Lyon. 10. Parodi S, Flora SD, Cavanna M, Pino A, Robbiano L, Bennicelli C & Brambilla G (1981). DNA-damaging activity in vivo and bacterial mutagenicity of sixteen hydrazine derivatives as related quantitatively to their carcinogenicity. Cancer Research 41, 14691481. 11. Cheeseman MA, Machuga EJ & Bailey AB (1999). A tiered approach to threshold of Regulation. Food and Chemical Toxicology 37, 387-412.
AFC/adhoc SEM/1
12. Ulland B, Weisburger EK & Weisburger JH (1973). Chronic toxicity and carcinogenicity of industrial chemicals and pesticides. Toxicology and Applied Pharmacology 25, 446. 13. Weisburger EK, Ulland BM, Nam J, Gart JJ & Weisburger JH (1981). Carcinogenicity tests of certain environmental and industrial chemicals. Journal of the National Cancer Institute 67, 75-88. 14. Hayatsu H, Takeishi KI & Ukita T (1966). Modification of nucleosides and nucleotides. III. A selective modification of cytidine with semicarbazide. Biochimica Biophysica Acta 123, 445-457. 15. Hayatsu H, & Ukita T (1966). Modification of nucleosides and nucleotides. IV. Reaction of semicarbazide with nucleic acids. Biochimica Biophysica Acta 123, 458-470. 16. Battacharya K (1976). Chromosome damage in spermatocytes of a grasshopper. Mutation Research 40, 237-242. 17. Hirakawa K, Midorikawa K, Oikawa S & Kawanishi S (2003). Carcinogenic semicarbazide induces sequence-specific DNA damage through the generation of reactive oxygen species and the derived organic radicals. Mutation Research 536 91-101. 18. Steffek AJ, Verrusio C & Watkins CA (1972). Cleft palate in rodents after maternal treatment with various lathyrogenic agents. Teratology 5, 33-40. 19. De La Fuente M (1986). Teratogenic effect of semicarbazide in Wistar rats. Biology of the Neonate 49, 150-157.
AFC/adhoc SEM/1
European Food Safety Authority

Brussels, 9 October 2003 EFSA/AFC/adhoc SEM/2-final
Additional advice on semicarbazide, in particular related to baby food Ad hoc expert group meeting 9 October 2003
Additional advice on semicarbazide, in particular related to baby food Ad hoc expert group meeting 9 October 2003
Background Preliminary advice on the possible occurrence of semicarbazide in certain packaged foods was issued by EFSA on 28 July 2003 (available on the Internet at http://www.efsa.eu.int/pdf/p_afc_doc_01.pdf). The foods concerned are those packaged in glass jars and bottles closed with metal lids sealed with plastic gaskets that are foamed using the blowing agent, azodicarbonamide. The Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food (AFC Panel) considered further data at its meeting on 1 October 2003 and prepared an update on the earlier advice (see Panel statement available at http://www.efsa.eu.int/p_foodadd_en.html). In the following week, an ad hoc group of experts was convened, including members of EFSAs AFC Panel, its Panel on Contaminants in the Food Chain and its Panel on Dietetic Products, Nutrition and Allergies. It heard a presentation from the food and packaging industries updating their findings since July. This new information covered industry progress on development of alternative technologies and strategies to reduce or eliminate semicarbazide in foods, further information on the range of concentrations of semicarbazide found so far in foods packaged in glass jars and bottles, and exposure estimates for infants. The ad hoc expert group was specifically asked by EFSA to advise further on possible risks to infants, given that this is the consumer group for which
Ad hoc expert group 9 Oct. 2003. Advice on SEM and baby food
potential exposure to semicarbazide on a body weight basis is likely to be the highest. The advice of the ad hoc group is given below and includes discussion of microbiological and nutritional aspects of baby foods which need to be considered alongside the toxicological aspects.
The advice on baby foods Exposure data The ad hoc expert group received additional information from industry on the range of concentrations of semicarbazide in foods packed in glass jars and bottles and on consumption of the relevant foods by infants of 4-12 months of age. The number of food samples analysed for semicarbazide is still limited, but confirm that amounts ranging up to 25 ppb have been found in baby foods. The additional information confirms that the preliminary estimate of exposure of infants to semicarbazide, contained in the AFC Panel statement of 1 October 2003, of around 2 microgrammes/kg body weight/day is a reasonable worst case.
Toxicological aspects No additional information on toxicological aspects was presented to the ad hoc expert group since the update provided by the AFC Panel on 1 October 2003. The Panel commented that semicarbazide has weak genotoxic activity in vitro and weak carcinogenic activity in female but not male mice, but because of the limited evidence, it is not possible to conclude whether semicarbazide may pose a carcinogenic risk to humans. There is no risk of immediate illness to adults, children or infants from consumption of foods containing semicarbazide. The concern relates to health in the long
term because of the possibility that semicarbazide may cause cancer. The risk, if any, is judged to be very small, both for infants and adult consumers. Although there are uncertainties in the risk assessment due to limitations in the data, these relate only to how to characterise a very small risk. Further details on toxicological aspects can be found in the advice issued by EFSA on 28 July 2003 and in the recent statement of the AFC Panel dated 1 October 2003.
Microbiological aspects Jars of baby food have an excellent microbiological safety record. Additional assurance for the consumer about the integrity of the seal is provided by the incorporation of the pop-up button. Current manufacturing practice for baby foods in jars includes controlled heat treatment to ensure that the product is microbiologically safe for the consumer. The airtight seal maintains microbiological stability and ensures a long shelflife. It prevents contamination with microorganisms, dirt, or insects, during storage and protects the nutritional integrity of the product. If alternative technologies are going to be used to reduce or eliminate semicarbazide from baby foods, it is crucial that the same rigorous microbiological standards are upheld.
The ad hoc expert group was informed that industry is making significant progress on the development of new seal technology. Any future changes to the current gasket would need careful consideration and evaluation of seal integrity before being launched as a replacement for the existing technology.
Nutritional aspects Processed baby foods are products which are carefully regulated with respect to ingredients, low levels of nitrate, pesticide residues and contaminants, and baby foods in glass jars and bottles make an important nutritional contribution to the diets of many infants. Although not an obligatory part of the diet of infants, they are widely used primarily for reasons of convenience, quality and nutritional safety. If these products were to become suddenly unavailable, or caregivers of infants chose not to use them any longer, it should be borne in mind that the nutritional adequacy of the diets of those infants could be compromised, particularly where caregivers have been relying extensively on these products. There is a possibility that the foods which would be chosen instead might be either not nutritionally equivalent to the foods used before or not suitable for the feeding of infants.
Conclusions and recommendations Semicarbazide is present in certain foods packaged in glass jars and bottles in very low quantities (parts per billion - ppb - or microgrammes/kg of food) due to migration from the plastic gasket which forms the airtight seal in the metal lid. Amounts of up to 25 ppb have been found in baby foods, which give an estimate of intake of around 2 microgrammes/kg body weight/day as a worst case for infants.
There is evidence from animal studies that semicarbazide is a weak carcinogen but it is not possible to conclude whether semicarbazide poses a carcinogenic risk to humans. The risk, if any, is judged to be very small, both for infants and for adult consumers.
Against this background and the microbiological and nutritional issues discussed above, it would be prudent to reduce exposure to semicarbazide as swiftly as technological progress safely allows, and unwise to take any immediate actions on baby foods which could trigger potentially hazardous consequences of a different nature. In the interim, there is no reason for consumers, including infants, to change their dietary habits because of the possible presence of semicarbazide in certain foods.

EFSA Journal 219

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

EFSA Journal 219

Încărcat de

Drepturi de autor:

Formate disponibile

The EFSA Journal (2005) 219, 1-36

The EFSA Journal (2005) 219, p. 2 of 36

The EFSA Journal (2005) 219, p. 3 of 36

The EFSA Journal (2005) 219, p. 4 of 36

The EFSA Journal (2005) 219, p. 5 of 36

The EFSA Journal (2005) 219, p. 6 of 36

The EFSA Journal (2005) 219, p. 7 of 36

The EFSA Journal (2005) 219, p. 8 of 36

The EFSA Journal (2005) 219, p. 9 of 36

The EFSA Journal (2005) 219, p. 10 of 36

The EFSA Journal (2005) 219, p. 11 of 36

The EFSA Journal (2005) 219, p. 12 of 36

The EFSA Journal (2005) 219, p. 13 of 36

The EFSA Journal (2005) 219, p. 14 of 36

The EFSA Journal (2005) 219, p. 15 of 36

The EFSA Journal (2005) 219, p. 16 of 36

The EFSA Journal (2005) 219, p. 17 of 36

The EFSA Journal (2005) 219, p. 18 of 36

The EFSA Journal (2005) 219, p. 19 of 36

The EFSA Journal (2005) 219, p. 20 of 36

The EFSA Journal (2005) 219, p. 21 of 36

The EFSA Journal (2005) 219, p. 22 of 36

The EFSA Journal (2005) 219, p. 23 of 36

The EFSA Journal (2005) 219, p. 24 of 36

The EFSA Journal (2005) 219, p. 25 of 36

The EFSA Journal (2005) 219, p. 26 of 36

The EFSA Journal (2005) 219, p. 27 of 36

The EFSA Journal (2005) 219, p. 28 of 36

The EFSA Journal (2005) 219, p. 29 of 36

The EFSA Journal (2005) 219, p. 30 of 36

The EFSA Journal (2005) 219, p. 31 of 36

The EFSA Journal (2005) 219, p. 32 of 36

The EFSA Journal (2005) 219, p. 33 of 36

The EFSA Journal (2005) 219, p. 34 of 36

ANNEX to Semicarbazide in food

The EFSA Journal (2005) 219, p. 35 of 36

ANNEX to Semicarbazide in food

The EFSA Journal (2005) 219, p. 36 of 36

European Food Safety Authority

PRELIMINARY ADVICE ON THE POSSIBLE OCCURRENCE OF SEMICARBAZIDE IN PACKAGED FOODS

The types of tumour found with semicarbazide were lung and

European Food Safety Authority

S-ar putea să vă placă și