Advisory Committee on Novel Foods and Processes

Salatrim

Mr A Klepsch
European Commission DG III
Rue de la Loi 200
B-1049
Brussels
Belgium

8 November 1999

Dear Mr Klepsch

The UK Competent Authority has assessed an application from Danisco-Cultor for clearance of Salatrims as a novel food and has recommended approval for the limited uses described in the application, that is in bakery and confectionery products only. The full UK initial opinion is attached to this letter. Any extension in the range of uses of this novel food ingredient would require a further application for approval.

The labelling provisions of the Novel Food Regulation require that consumers be provided with information if any properties of the food are significantly different from those of the foods that it would replace. Foods containing Salatrims will provide significantly fewer calories than foods containing conventional fats and this information should therefore be provided to consumers. However, this requirement is in conflict with the requirements of the Nutrition Labelling Directive regarding the declared caloric value of any fat component of foods. At the same time one of the main provisions of Directive 79/112/EEC is to ensure that labelling does not mislead the consumer to a material degree as to the characteristics of the foodstuff. This generic issue, which will apply to an increasing number of fat replacers, needs to be resolved at Community level as a matter of urgency to ensure that the labels of such products do not mislead the consumer.

Finally, we wish to point out that, as a general principle, young children do not need to restrict their energy intake. Therefore, on nutritional grounds, foods containing Salatrims, in common with any other low fat foods, should not be targeted at this age group. In the UK we are considering ways of providing appropriate nutritional advice to this group of consumers, we would encourage other Member States to do the same.

Yours sincerely

Mrs S J Hattersley
For the UK Competent Authority

1. The UK Competent Authority accepted an application from Cultor Food Science ¹ for clearance of Salatrims, a family of low calorie fat replacers, under the EU Novel Food Regulation on 28 June 1999. The UK had previously considered data on Salatrims under the voluntary scheme for novel foods that existed in the UK prior to the introduction of the EU Regulation in May 1997. A number of concerns had been identified in the original consideration of these materials and these concerns have been addressed in the current application. During the consideration of the data, the Advisory Committee on Novel Foods and Processes (the UK assessment body) sought clarification of a number of issues and supplementary papers to answer these points were submitted by the company on 24 and 27 August 1999 ^2,3.

2. Salatrims (Short And Long chain Acyl TRIglyceride Molecules) are a family of structured triglycerides composed predominantly of mixtures of long chain fatty acids (principally stearic) and short chain organic acids (acetic, and/or propionic and/or butyric) esterified with glycerol. They are prepared by the interesterification of glycerol esters using a process routinely used by the edible oil industry. The materials are designed to provide fewer calories than conventional fats, but to have similar technological properties. This application is restricted to the use of Salatrims in bakery and confectionery food products.

3. Much of the information summarised in the application has been published in the Journal of Agricultural and Food Chemistry⁴, and is also contained in the GRAS affirmation petition submitted to the UD Food and Drug Administration⁵ and in the report of an expert panel of the Life Sciences Research Office of the Federation of American Societies for Experimental Biology⁶. Salatrims have also been evaluated by the FAO/WHO Joint Expert Committee on Food Additives⁷.

4. According to the EC guidance on the classification of novel foods, Salatrims are classified as Class 2.1 - complex novel food from non-GM sources, where the source of the novel food has a history of use in the Community. The information provided on Salatrims is addressed below, according to the requirements for this class of novel food. The location of the information in the summary document and the full application dossier are given for each of the information areas detailed below.

5. A proposed generic specification has been provided by the company, which is attached to this opinion (Annex I). The fatty acid composition of Salatrims can be varied to give differing technological properties and certificates of analysis for five Salatrim variants are included in the application.

6. The proposed specification states that Salatrims contain >87% triacylglycerols. The fatty acid composition can be varied to confer differing technological properties. The proportion of short chain organic acids (acetic, propionic and butyric acids) can vary between 33 and 70% and that of long chain fatty acids (mainly stearic acid) varies between 30 and 67%. The fatty acids used in the manufacture of Salatrims can be sourced from a number of hydrogenated oils (canola, soya bean, and cottonseed). These oils may be derived from approved genetically modified plant varieties, where such oils meet the normal specifications for edible oils. Further details of the composition of the Salatrims used in the safety studies supporting this application are given in the first statement from the Committee on Toxicity (Annex II to this opinion).

7. The production process used to manufacture Salatrims is commonly used in the edible oils and fats industry to modify the physical properties of edible oils, and the raw materials used are of food grade quality. The process consists of the interesterification of long chain saturated fatty acids (derived from hydrogenated soya, canola and cottonseed oils) with an excess of triacetin and/or tripropionin and/or tributyrin, using a sodium methoxide catalyst. The catalyst is deactivated with distilled water when the reaction is complete and the oil is then washed, bleached and filtered. Any residual mono- and diglycerides are converted to the corresponding triglycerides, as necessary, and any residual short chain triglyceride is reduced to less than 1% by vacuum and steam stripping followed by vacuum distillation. The composition of the particular Salatrim variant is determined by the molar ratios of the starting materials used (1.5-12 molar excess of the short chain fatty acids) and actual composition is highly correlated to that predicted (see published paper in Journal of Agricultural and Food Chemistry⁴).

8. Salatrims are derived from hydrogenated soya, canola and cottonseed oils, all of which has a long history of safe use within the European Community, as sources of edible oils. The oils may be derived from genetically modified varieties that have obtained approval as novel foods, where the composition of such oils meets the normal requirements for edible oils.

9. This application is limited to seeking approval for use of Salatrims in bakery and confectionery food products only. The company has stated that Salatrims are not intended for use in infant formulas and their technical properties preclude their use as frying oils/fats. They are intended for sale to the food processing industry for use in manufactured products and will not be available directly to the consumer. The range of uses permitted in the USA is wider than that applied for in the Community.

10. The main food products in which it is envisaged that Salatrims will be used are chocolate, chocolate confectionery, other chocolate products, buns and pastries, and cookies and brownies. The company has used data from the UK National Dietary and Nutrition Surveys for adults’ food consumption, together with information on the fat content of foods, to estimate a mean intake of Salatrims of 11g/day, with a 97.5^th percentile intake of 33g/day. Products containing Salatrims will be aimed at those people choosing a restricted calorie diet. The Company has suggested that such products are likely to carry a price premium and will not be aimed at young children. For this reason, intake estimates have not been generated for young children.

11. Salatrim structured triglycerides contain at least one short chain organic acid (acetic and/or propionic and/or butyric) and at least one long chain saturated fatty acid (primarily stearic). These fatty acids and glycerol are common components of the normal diet or are formed by colonic bacteria and the levels of intake anticipated from consumption of Salatrim food products are not associated with any adverse health effects (see also section VIII on toxicological information).

12. Data on the composition of Salatrim fats have not shown the formation of any unexpected by-products of the production process, nor has the presence of any unexpected, non-acylglycerol components been demonstrated. Levels of phytosterols, tocopherols, unsaponifiable components and organic constituents are found at levels comparable to, or lower than, those found in conventional oils and fats. Pesticide residue levels in Salatrim fats are comparable to the levels found in the commercial hydrogenated fats from which they are derived.

13. Post-market data from self-reporting systems in the USA (toll-free telephone calls, internet and mail), where Salatrim food products have been on the market since 1995, indicate only a low incidence of possible gastrointestinal tract effects following consumption.

The constituents of Salatrims are common components of the normal human diet and the levels of intake arising from the consumption of Salatrim food products are not associated with any known adverse health effects. The results of toxicological studies in animals and of human volunteer studies are addressed later in this document.

14. The use of Salatrim in food products in place of conventional fats can result in an increase in the dietary intake of saturated fats. However, the total dietary intake of calories from fat will decrease. This is because although the absolute intake of stearic acid is increased following substitution of conventional fats by Salatrim, not all of this stearic acid is absorbed.

15. In addition, the use of Salatrim will displace from the diet other saturated fats, such as palmitic, myristic and lauric acids, whose consumption is known to be associated with increases in serum cholesterol levels. The company has reviewed the literature on the cholesterolaemic effect of stearic acid and has suggested that it is neutral in this respect in relation to fatty acids such as lauric and myristic. The company also noted that stearic acid does not induce changes in platelet phospholipids, nor does it increase the tendency for thrombosis.

16. The ACNFP, in its initial consideration of the safety of Salatrims, had identified thrombogenic potential as a possible area of concern. In response to this concern, the company has conducted studies in humans to address this issue and the results are included in the current application. A single dose study was conducted in volunteers, who were at moderately increased risk of coronary heart disease, to investigate post-prandial changes in various markers of fibrinolytic activity. Subjects were middle-aged and had moderately raised plasma cholesterol levels (between 5.2 and 7.8mmol) and were also moderately overweight. These subjects consumed test meals rich in Salatrim or one of two control meals (containing equivalent amounts of oleate-rich sunflower oil or cocoa butter). The post-prandial increase in serum triglycerides seen after consumption of the Salatrim meal was lower than those seen after the control meals. Fibrinolytic activity (as measured by tissue plasminogen activator activity and plasminogen activator type I activity) was not affected by the different fats present in the test meals. However, plasma factor VII coagulant activity and the plasma concentration of activated factor VII were increased to a lesser extent after the Salatrim meal in relation to the two control meals. The company also conducted a five-week crossover study in hypercholesterolaemic subjects on margarines enriched with Salatrim or palm oil. This study showed no evidence of any increases in factor VII coagulant activity or fibrinogen following consumption of Salatrim.

17. The company has considered whether consumption of Salatrims could impair the absorption of fat-soluble vitamins. The results from a number of studies in both animals and human volunteers support the view that Salatrim does not affect the absorption of vitamins A and E. In addition, the physico-chemical properties of Salatrim would not suggest that it would have an adverse consequence on fat-soluble vitamins.

18. The Salatrim family of structured triglycerides has been designed to provide 4-6 calories/g rather than the 9 calories/g supplied by traditional fats. This reduced caloric value is partially attributable to the presence of the short chain organic acid constituents, which are less energy dense than long chain fatty acids, and partially to the incomplete absorption of the stearic acid component, which is the predominant long chain fatty acid constituent. The company has provided evidence from a two- week growth assay in rats to support a reduced caloric intake from Salatrim, in relation to conventional fats. The company is claiming that this study, and other supporting data, would suggest a caloric value of Salatrim of the order of 5-6 calories/g.

19. Salatrims are manufactured under the normal Good Manufacturing Practices guidelines and the oil phase precludes the growth of typical food borne micro-organisms.

20. A number of animal toxicology and human clinical have been conducted to investigate the safety of Salatrims. The studies have tested a number of representative Salatrim variants, covering the range of commercial products. The ACNFP sought the specialist advice of it’s sister committee, the Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) on the testing strategy adopted for Salatrims and on the results obtained. The COT first considered Salatrims in 1997, before the Novel Food Regulation came into effect, and issued a statement summarising the toxicological and clinical data then available. The COT identified a number of concerns at that time, which have been addressed in the application made under the Regulation. The new data and arguments have also been considered by the COT and that committee has now issued a further statement of it’s views. These statements are attached as Annexes II and III to this opinion.

21. Salatrims are a series of structured triglycerides that have been designed to provide fewer calories than conventional fats. They are manufactured from traditional food sources using existing food processing technologies. The company has sought clearance for the use of Salatrims in a restricted range of food products, namely baked goods and confectionery, and has estimated average and extreme intakes of Salatrims in adults arising from such uses, assuming maximum substitution of conventional fats in these products by Salatrims.

22. The company has provided data from a range of animal and human studies to support the safety-in-use of Salatrims. They have addressed a number of concerns that had been identified in the course of the safety evaluation, namely possible thrombogenic effects, possible toxicity of the short chain organic acid component of Salatrims, possible hepatotoxicity and gastro-intestinal tract symptoms.

23. On the basis of the totality of the data now provided to us, we are satisfied that Salatrims are acceptable for use in the food product categories described in the application, namely baked goods and confectionery. Any extension of the range of use of Salatrims to other food categories would require a further approval. We note that there are no specific safety data relating to possible effects of Salatrims in children. However, data from a recently completed, but as yet unpublished, diet and nutrition study in the UK, has indicated that older children consume approximately equivalent amounts of the food products that Salatrims would be used in as adults, with the exception of confectionery, where intake is somewhat higher than in adults. As a general principle, young children do not need to restrict their energy intake and therefore, on nutritional grounds, foods containing Salatrims should not be aimed at this age group. However, the question of the applicability of any low calorie/low fat food products for this young age group is a generic one, that is not restricted just to consideration of foods containing Salatrims.

24. We accept that Salatrims provide fewer calories than conventional fats and we recommend that the consumer should be provided, via labelling on the products themselves, with information on the true caloric value of foods containing Salatrims. This is in conflict with the provisions of the present Nutrition Labelling Directive and this issue will need to be resolved at European Commission level.

25. We have considered the information presented in the application dossier and the supplementary information provided in response to specific questions raised during our evaluation and we conclude that Salatrims, as specified, are acceptable for use in the food products described, that is in baked goods and confectionery. A further application would need to be made before the use of Salatrims could be extended to other food products.

26. We accept that Salatrims supply fewer calories than conventional fats. We recommend that consumers should be supplied with appropriate information on this aspect, and that this may require amendment of the provisions of the Nutrition Labelling Directive.

1. Submission for approval under Regulation (EC) No 258/97 of the European Parliament and the Council of 27^th January 1997 concerning novel foods and novel food ingredients, for the use in bakery and confectionery food products of the family of low calorie fats known as "SALATRIMS". Cultor Food Science (Xyrofin UK Limited). 26^th May 1999.

2. Supplementary information on serum hepatic enzymes and dropouts from the free-living clinical study. Cultor Food Science. 24^th August 1999.

3. Supplementary information on thrombogenic potential. Cultor Food Science. 27^th August 1999.

4. Journal of Agricultural and Food Science, 42, 2, pages 432-604. 1994

5. US GRAS affirmation petition. 1994.

6. The evaluation of the health aspects of using certain triacylglycerols as food ingredients. Life Sciences Research Office of the Federation of American Societies for Experimental Biology. 1993

7. World Health Organisation, 1998. Salatrim (short and long chain acyltriglyceride molecules). In: Safety Evaluation of Certain Food Additives and Contaminants. WHO Food Additive Series, No 40, pages 111-141. Geneva: WHO.

8. National Diet and Nutrition Survey for young people: 4-18 years. 1999. In Press.

1. We have been asked by the Advisory Committee on Novel Foods and Processes (ACNFP) to comment on specific aspects of a large submission of data received by the ACNFP in respect of Salatrims, a family of low calorie fat products. The ACNFP reviewed the available toxicological and clinical safety data on Salatrim at its 32nd meeting on 26 September 1996. The ACNFP requested advice from the Committee (COT) in respect of:

i) the adequacy of the animal toxicological database and, in particular, the arguments proposed by Cultor Food Science, who wish to market these products in the UK, regarding limited testing of Salatrims in animals.

ii) an evaluation of the increases in serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) documented in human clinical studies and, for comments regarding the lack of predictivity of the animal toxicology studies in respect of these effects on the liver.

2. Salatrims comprise a family of structured glycerides composed predominantly of mixtures of long chain fatty acids (LCFAs; principally stearic acid) and short chain fatty acids (SCFAs; acetic, propionic and/or butyric) which are intended as low caloried fats for use in soft sweets, coatings (eg wafers and confections), dairy products, shortening and potentially in table spreads. Predicted intakes will vary according to the uses of Salatrims, level of substitution for existing fats, and the extent to which these fatty acids are consumed. We have considered several reports provided by Cultor Food Science, which present calculations regarding potential intakes. The estimated intake for the whole population assuming selected uses varied between 11-29 g/day (mean), and 18-65 g/day (97.5th percentile). Estimates of intake using all potential uses varied between 18-46 g/day (mean) and 30-88 g/day (97.5th percentile). We also note that the Joint Food and Agriculture Organisation/World Health Organisation Expert Committee on Food Additives estimated intakes in children aged 3-5 years to be approximately 26 g/day (90th percentile). We note that there are considerable uncertainties regarding the methods and precision of the calculated estimates of potential Salatrim intake and that the various figures provided by the company show a wide variation. We are also aware that children have a caloric intake, which on a body weight basis is higher than that in adults, on which basis their exposure to Salatrim would also be higher. Thus it would have facilitated interpretation of the intake data if the latter had been calculated and expressed in terms of grams Salatrim/kg bw/day and grams Salatrim/energy intake/day. However, we consider that the available figures provided by the company can be used as a guide to the evaluation of the clinical and toxicological data provided in the submission.

3. An abbreviated nomenclature has been used by the manufacturers and also throughout this statement to describe the various Salatrims which have been evaluated in toxicological and clinical safety studies. An account of this nomenclature and details regarding all Salatrims tested are given in tables 1 and 2 of the Annex to this statement. (For example, in Salatrim 43SO tributyrin and tripropionin are the SCFAs and the LCFA source is hydrogenated Soybean Oil.)

4. The specific questions raised by the ACNFP with regard to Salatrims were considered by the COT during 1996 and at a joint ACNFP/COT Working Group in early 1997 where representations from the company were heard. Following this latter meeting, additional information regarding intakes and evaluation of the animal toxicity and human clinical studies was submitted to the COT. A short summary of the available animal and human toxicology data on Salatrims is given below for information so that our consideration of the questions raised by the ACNFP can be placed into context.

5. Initial in-vitro studies of the hydrolysis of Salatrim products using porcine pancreatic lipase demonstrated that a wide range of Salatrim triacylglycerides underwent rapid hydrolysis. In-vivo metabolism experiments in rats were designed to compare the metabolism of a specified Salatrim (23CA) with triolein and the results showed that Salatrim 23CA was metabolised in an analogous way to triolein (a normal dietary fat).

6. Five 90-day feeding studies were undertaken in rats using a range of Salatrim products selected to include different combinations of short chain fatty acids (ie acetate, propionate, butyrate) and different sources of stearate (ie canola oil, cottonseed oil and soybean oil). Details regarding the Salatrim products tested are given in table 2 of Annex 1. The results of the toxicity studies were consistent, showing no toxicologically significant effects at up to 10% w/w in the diet (ie approximately 6-8 g fat/kg bw/day). The observed changes in clinical chemistry and histopathological findings noted in these studies, which occurred predominantly at 10% w/w in the diet and consisted of alterations in bone mineral levels (eg increased bone zinc concentrations), and of renal mineralisation in female rats, were considered by the authors to be consistent with alterations in mineral metabolism induced by a reduced proportion of polyunsaturated fats in the diet. The authors of the animal toxicology studies noted that published evidence is available which reported diet-induced alterations in mineral metabolism, and in particular reduced bone zinc concentrations in rats following administration of diets containing high levels of polyunsaturated fatty acids. (Lusak & Johnson, 1992) Dietary administration of 10% w/w Salatrim 23SO to rats for up to 17 days had no effect on serum activity of aspartate aminotransferase (AST), alanine aminotransferase (ALT) or g -glutamyltransferase (GGT) activities, although increases in the levels of serum enzyme have been documented in the human clinical studies. No toxicologically significant effects were documented in a 28-day study in minipigs fed a diet containing 10% w/w Salatrim 23SO. Samples of caecum were taken during routine necropsies of the rats fed Salatrim 23CA or 32CA for 13 weeks, frozen and then subsequently analysed for gut microflora and for evidence of any Salatrim induced changes in caecal pH, and primary/secondary metabolites of bile acids and phytosterols and for cholesterol and coprostanol levels. There was no evidence of any effects on gut microflora from these experiments. We note that the morphological methods used in these investigations to assess effects on gut microflora were insensitive. Salatrim products do not contain any structural alerts for potential mutagenicity. There was no evidence of genotoxicity in an adequate range of in-vitro or in-vivo studies. No studies to evaluate potential carcinogenicity or effects on reproduction are available.

7. Four clinic-based studies using adult volunteers have been undertaken which utilised double blind study designs. The experimental protocols used exposures up to 60 g Salatrim/day for 1, 4 or 7 days and study designs that included a 4-day triple crossover experiment. This type of design allowed the effects of three different fat sources on clinical chemistry and recording of symptomatology to be compared in the same individuals, but we note that only a very short exposure period of 4 days was used. The results of the clinic based studies with tested Salatrim products showed similar effects on gastrointestinal function (ie nausea, stomach cramps, diarrhoea, flatulence were reported) at 60 g/day. These effects subsided when the Salatrim administration was stopped. Overall, the authors concluded that 30 g Salatrim/day in the clinic-based studies did not result in any gastrointestinal effects. We note, however, the limited duration of these studies and that clinical investigations were only performed on healthy adults and thus potential effects in children or individuals with compromised gastrointestinal function were not considered.

8. A 28-day free living study was undertaken using groups of at least 12 male and 12 female volunteers consuming a range of food products containing one of three Salatrim products (23SO, 43SO, 4SO) at levels of up to 60 g Salatrim/day. The purpose of this study was to allow an extended evaluation of the Salatrim products and to confirm the effects seen in the clinic-based studies and also to determine if any of these effects were reversible. The adverse effects reported in this study, predominantly at 60 g Salatrim/day, were consistent with those documented during the clinic based studies (ie gastrointestinal disturbances and increases in serum transaminase levels). The authors considered that the effects on serum transaminase levels were transient with noticeable decreases in serum activities of AST and ALT occurring towards the end of the study period. We note, however, that convincing evidence of a decline in the activities of serum AST and ALT to baseline levels was not provided. Although increases in serum AST and ALT activities rarely reached clinical significance in the 28 day free living study, we note that there were concomitant increases in the activities of some liver function enzymes (ie AST and ALT) and reduced serum cholesterol documented in the clinic-based studies at 60 g Salatrim/day. There is limited published evidence to support the view that an increased load of high caloric materials such as proteins or sucrose in the diet may induce alterations in serum transaminase enzyme levels (Schimke, 1962; Porikos & Van Itallie, 1983).

9. The authors noted that none of the subjects in the 28-day free living study reported severe gastrointestinal effects, which might impair normal function. The effects were considered to be "mild" or "annoying". They considered that 30 g Salatrim/day would have little or no effect on the health of individuals. We note that complaints of adverse effects on gastrointestinal function persisted for at least 10 days in a small number of individuals at all dose levels, including two who consumed 30 g Salatrim/day.

10. There are a number of proposals concerning the mechanism(s) for the effects of Salatrim on gastrointestinal function. Increased stool weight and faecal water content as reported in the 4-day triple cross over study could have resulted in a bulking effect and hence may have contributed to the gastrointestinal symptoms. (Besselaar Clinical Research Unit, 1993). The authors have also proposed that the introduction of Salatrim, a poorly digested fat, into the diet may have caused transient gastrointestinal disturbance since similar effects have been documented following an abrupt increase in dietary fibre intake. (Finley et al., 1994a; Pilch, 1987). The authors of the clinic-based studies have also speculated that the level of short chain fatty acids at high Salatrim doses (ie 60 g/day) might temporarily overwhelm the ability to utilise acetate which might be sufficient to induce some adverse gastrointestinal symptoms. (Finley et al., 1994b). Overall, we conclude that there is no convincing explanation regarding the mechanism of the adverse gastrointestinal effects of Salatrims seen in the clinical studies. Additionally there is no information available regarding the likelihood of adverse gastrointestinal effects in children or individuals with compromised gastrointestinal tract function.

11. We consider that the animal toxicity studies were adequately conducted and can be used in the safety assessment of Salatrims. However, we note, that there are inadequate data available in respect of the potential effects of bolus doses of SCFAs on reproduction following their release from Salatrims during the metabolism of Salatrims in the gastrointestinal tract. We are aware that butyrate and propionate have been shown to have teratogenic potential in-vitro (Coakley et al. 1986, Brown et al. 1987). We reviewed the additional data provided by Cultor Food Science including calculations regarding potential blood levels of butyrate following consumption of a meal containing 30 g of Salatrim 4SO and conclude that additional pharmacokinetic studies to evaluate blood levels of SCFAs in volunteers following consumption of individual Salatrim products are required before any conclusions regarding the teratogenic risk of Salatrim(s) can be drawn. We agree that there is no requirement for additional mutagenicity studies or for the provision of carcinogenicity bioassays with Salatrim(s).

12. We have considered all of the data available from the clinical studies with regard to the potential effects of Salatrim(s) on serum enzymes which can be used to evaluate potential adverse effects on liver function and have also considered the further information supplied by Cultor Food Science regarding an evaluation of the individual clinical chemistry for all of the clinical studies. We conclude that increases in both AST and ALT occurred in a higher proportion of individuals consuming 30 g or 60 g Salatrim per day for 28 days compared to controls, although only a few of these reported increases can be regarded as reaching the level of clinical significance. Whereas there were differences between individuals in respect of the magnitude of the increase in AST and ALT in response to the ingested dose level of Salatrim and also between that induced by different Salatrim products, we conclude that the evidence is consistent with a weak treatment related effect which did not appear to decline during the 28 day treatment period. The analysis of individual clinical chemical data for all of the clinical studies is complicated by the absence of detailed background information on the individuals included in the clinical studies and also, in respect of the 28 day study, by the identification of inconsistencies and transcription errors both in the original and published reports of this study. There was also additional evidence of concurrent increases in other serum enzymes consistent with liver dysfunction, such as alkaline phosphatase, lactate dehydrogenase and GGT, in a small number of individuals who ingested Salatrims. Overall, we consider that a No Observable Adverse Effect Level (NOAEL) with respect to clinical chemical markers of liver function cannot be identified from the clinical safety studies undertaken with Salatrims and that no conclusions can be drawn with regard to the mechanism or biological significance of the Salatrim-induced effects on AST and ALT.

13. We consider that there are insufficient data available to derive any firm conclusions regarding the reasons for an absence of effects on serum liver enzymes in the animal studies. We conclude that it would be appropriate to proceed on the basis that the effects on serum liver enzymes documented in the clinical safety studies were treatment-related and thus require additional evaluation, particularly with respect to the identification of an appropriate NOAEL. In this regard the discrepancies in the reporting of the results from the 28-day free living study limit the value of this investigation. We recommend that a suitable long-term clinical evaluation study of the individual Salatrims to be marketed should be undertaken with particular reference to the identification of a NOAEL.

14. We have, during our consideration of the questions raised by the ACNFP, evaluated all of the available toxicological and clinical safety studies on Salatrim and have considered the representations made by Cultor Food Science at the joint ACNFP/COT Working Group. The following two conclusions respond to the specific requests made by the ACNFP. An additional conclusion based on a consideration of all the toxicological and clinical data submitted to the COT is given in paragraph 15.

i) Animal toxicological data alone on Salatrims are insufficient to evaluate the proposed use of these materials as fat replacers. We recommend that additional pharmacokinetic studies to evaluate blood levels of SCFAs in volunteers following consumption of individual Salatrim products are required before any conclusions regarding the teratogenic risk of Salatrim(s) can be drawn.

ii) The clinical safety studies demonstrate a weak treatment related effect of Salatrims on serum levels of marker enzymes for liver dysfunction. There are insufficient data available to derive any firm conclusions regarding the reasons for an absence of effects on serum liver enzymes in the animal studies. The documented discrepancies in the reporting of the results from the 28 day free living study limit the value of this investigation with regard to evaluating the potential effects of Salatrims on liver function. We recommend that a suitable long-term clinical evaluation study of the individual Salatrims to be marketed should be undertaken with particular reference to the identification of a NOAEL for this effect.

15. Regarding overall conclusions, we note that evidence of adverse effects on gastrointestinal function and on marker enzymes for liver dysfunction in humans were reported following the consumption of 30 g Salatrim 23SO for 28 days and that there was clear evidence of adverse effects at 60 g/day regarding several Salatrim products. We are concerned that there would appear to be no margin of safety between these levels of consumption and the calculated potential intakes reported above in paragraph 2. Additionally children and individuals with compromised gastrointestinal function might be more susceptible to these particular effects associated with Salatrim consumption. We therefore recommend that the additional clinical studies requested in paragraph 14 (ii) should further investigate the potential effects of the Salatrim(s) to be marketed on gastrointestinal function with a view to also identifying a NOAEL for this effect.

December 1997

Table 1.Typical molar ratios of short- and long-chain acid sources used to prepare the SALATRIM family of edible oils*

* The SALATRIM family name defines the sources of the short-chain and long-chain fatty acids with the numerals representing the carbon chain lengths of the short chain acids in decreasing proportion in the mix; the letters define the oil which provides the source of the long-chain fatty acids. (E.g. in SALATRIM 43SO tributyrin and tripropionin are the SCFAs and the LCFA source is hydrogenated Soybean Oil. The molar ratio of the mix that is used to prepare the SALATRIM is 11 parts tributyrin : 1 part tripropionin : 1 part hydrogenated soybean oil).

A listing of those products that have been used in safety evaluation studies is also given below. There are only very minor differences in composition between Salatrim products prepared from different long chain fatty acid sources. However, different batches of a product may have used differing molar ratios of the starting products.

Table 2. Materials used in metabolism and toxicity studies

Ames tests 4CA, 23CA, 23SO, 32CA, 234CA, 234CS

In vitro mammalian tests 23CA

In vivo bone marrow 234CA, 234SO

micronucleus assays

In vitro metabolism 4CA, 23CA, 32CA, 234CA

(Porcine pancreatic lipase)

Metabolism in rats 23CA

90 day feeding studies (rats) 4CA, 23CA, 32CA, 23SO, 234CA, 234CS*

(*and supplementary 17 day test of effects on transaminases)

28 day mini-pigs 23SO

Effects on gut microflora: rats 23CA, 32CA

Studies I & II in volunteers 23CA

Studies III & IV in volunteers 23SO

Free living study in volunteers 4SO, 23SO, 43SO

Besselaar Clinical Research Unit (1993a) Randomised, 3-way crossover, double blind tolerance study of fat replacement compound TAG A9300 versus soybean oil administered to non-sedentary subjects by substituting 30 g/day or 60 g/day at 1800 or 2500 Kcal/day diets. Unpublished report No. 8024 from Besselaar Clinical Research Unit.

Brown NA, Coakley ME and Rawlings SJ (1987). Structure teratogenicity relationships of valproic acid congeners in whole embryo culture, chapter 2 in Approaches to Elucidate Mechanisms in teratogenesis. Ed Welsch F.; Hemisphere Publishing Corporation, pp17-29.

Coakley ME, Rawlings SJ, and Brown NA (1986). Short chain carboxylic acids, a new class of teratogens: Studies of potential biochemical mechanisms. Env. Health Perspect., 70, 105-111.

Finley, J.W., Walchak, C.G., Sourby, J.C. and Leveille, G.A. (1994a). Clinical study of the effects of exposure of various SALATRIM preparations to subjects in a free-living environment. J. Agric. Fd. Chem. 42, 597-604.

Finley, J.W., Leveille, G.A., Dixon, R.M., Walchak, C.G., Sourby, J.C., Smith, R.E., Francis, K.D., & Otterburn, M.S. (1994b). Clinical assessment of SALATRIM, a reduced-calorie triacylglycerol. J. Agric. Fd. Chem. 42, 581-596.

Lusak, H.C. & Johnson, P.E. (1992). Dietary fatty acids and minerals. In: Chow, C.K. (ed.), Fatty acids and their health implications, Dekker, New York, pp. 501-516.

Pilch, S.M. (1987). Physiological effects and health consequences of dietary fiber. Life Science Research Office of the Federation of American Societies for Experimental Biology (LSRO - FASEB). Unpublished report prepared for US FDA CFSCAN, DHEW, contract number 223-84-2059.

Porikos, K.P. & Van Itallie, T.B. (1983). Diet induced changes in serum transaminase and triglyceride levels in healthy adult men. Am. J. Med., 75, 624-630.

Schimke, R.T. (1962). Adaptive characteristics of urea cycle enzymes in the rat. J. Biol. Chem., 237, 459-468.

1. In 1997 we issued a statement (Committee on Toxicity, 1997) on specific aspects of a submission made to the Advisory Committee for Novel Foods and Processes (ACNFP) on Salatrims, a family of low calorie fat materials. We had been asked by ACNFP to comment on the toxicological aspects of the data provided at that time under the voluntary system that existed for the evaluation of novel foods.

2. We concluded that there were three areas of concern that needed to be addressed further, namely:

i) additional pharmacokinetic studies were required to evaluate blood levels of short chain fatty acids in volunteers following consumption of individual Salatrim products before any conclusions could be drawn regarding the teratogenic risk of Salatrims;

ii) a No Observable Adverse Effect Level (NOAEL) should be determined for effects of Salatrims on enzyme markers for liver dysfunction in humans; and

iii) a NOAEL should be determined for effects of Salatrims on gastrointestinal function in humans.

3. The EC Regulation on Novel Foods and Novel Food Ingredients was introduced in May 1997, after the initial consideration of Salatrims in 1995-1997. As Salatrims were not on the market in Europe at that time, they are regarded as novel foods and therefore they require clearance under the Regulation before they can be sold in Europe. An application has now been made to the UK Competent Authority for such clearance (Cultor Food Science, 1999) and ACNFP has asked for our further advice on the additional information and analysis included in the submission to meet the concerns that we identified earlier.

4. Our first statement described the toxicological and clinical data available on Salatrims in 1997. This statement, which needs to be read in conjunction with our previous opinion, discusses the additional data and analyses provided in the submission made under the Novel Food Regulation in response to the concerns that we had identified.

5. The Committee had previously expressed a concern that the release of butyric acid from Salatrim, which is rich in this short chain fatty acid, might be sufficient to cause a rise in plasma butyrate levels. This possibility needed to be examined in the light of in vitro data on possible teratogenic effects of high levels of butyric acid. A new study has been conducted (Pronczuk, Lipinski and Hayes, 1999) in which the post-prandial response to a load of 30 grams (g) of butyrate-based Salatrim (known as 4SO) was investigated to determine, amongst other effects, whether free butyrate would reach the general circulation. Plasma was analysed for butyric acid at regular intervals up to 360 minutes after dosing and at no time could free butyric acid be detected.

6. The Committee was satisfied by these data.

Effects on enzyme markers for liver dysfunction in humans

7. The Committee had asked that further information be provided to enable a NOAEL to be set for the effects of Salatrims on enzyme markers of liver dysfunction seen previously in the free-living clinical study. The company has provided further statistical analyses of the serum enzyme data from this study and expert assessments of the implications of those statistical analyses.

8. Small, statistically significant increases were recorded in the mean activities of the enzymes aspartate aminotransferase (AST) and alanine aminotransferase (ALT) at the beginning of the exposure period in the free-living clinical study, although the mean values for the treatment groups did not fall outside the normal ranges for these enzymes at the laboratory conducting the study. The Committee noted that there was some concordance in the increases in liver enzymes AST and ALT within individuals and that some temporal trends were apparent. The Committee had also noted that it was possible that the individuals who withdrew from the study may have been atypical and that this could have contributed to the variability of the data. However, the company submitted information that shows that those individuals who withdrew from the study, did so mainly because of adverse gastro-intestinal effects resulting from consumption of 60 g/day Salatrim. In addition, further analysis of the enzyme activities for those withdrawing from the study in comparison to those remaining shows that there were no significant differences in changes above baseline for serum hepatic enzyme activities whether or not those subjects withdrawing from the study were included.

9. In addition, data from a further study conducted by Nestel and colleagues (1998) were provided. In this five-week crossover study, following a low-fat control phase, subjects consumed diets containing margarine rich either in Salatrim or in palm oil (‘palmitate’). These subjects were selected on the basis of their having an elevated blood cholesterol level and the study was designed primarily to provide information on effects of Salatrim on blood lipid levels. However, additional analyses of serum enzymes from blood samples taken pre-dose and at the end of each dietary phase showed that mean activities of AST and ALT, and other indices of hepatic function (g -glutamyl transferase, alkaline phosphatase and lactic dehydrogenase) were no different at the end of the Salatrim and ‘palmitate’ dietary phases. Small increases in serum enzymes were noted after the ‘palmitate’ dietary phase in some subjects, which were of a similar magnitude to those seen in some subjects consuming Salatrim in the free-living clinical study described above.

10. The Committee concluded that consumption of high doses of Salatrim did result in slight increases in AST and ALT activities in serum, but noted that these increases were within the normal reference range. The further analyses of the enzyme data satisfied the Committee that there were no differences in the trends seen whether those withdrawing from the study were included or not. These enzyme changes, in the absence of any other indications of liver damage, were not considered to represent a clear toxic effect. The data from the five-week crossover study (Nestel et al., 1998) provided reassurance of a lack of any clear adverse hepatic effect. However, the Committee noted that the reported studies did not include investigations in children or in those with pre-existing liver disease.

Effects on gastrointestinal function in humans

11. The Committee had noted previously that gastrointestinal effects had been recorded in a number of subjects consuming Salatrim and had asked that further information be provided to enable a NOAEL to be determined. A further analysis of these observations was submitted by the company. The company suggested that the gastrointestinal disturbances reported was the consequence of sudden changes from an absorbable diet to one containing significant amounts of unabsorbable material. Metabolism of Salatrims rich in short chain fatty acids might result in rapid release of cholecystokinin, which would lead to slow gastric emptying, nausea and bloating. Analysis of the reasons for subjects withdrawing from the study shows that adverse gastrointestinal symptoms (such as nausea, cramps and gas) were the predominant reason for withdrawal from the study in the 60-g/day Salatrim group. However, there was no clear evidence of any increase in the incidence of such effects at intakes of 30 or 45 g/day Salatrim.

12. The Committee considered that the gastrointestinal disturbances seen after consumption of high doses of Salatrim could be the result of individual intolerances, noting the considerable variation between individuals and their tolerance to fibre and other poorly digested substrates. Nevertheless, these effects were seen in a significant proportion of individuals after consumption of 60 g/day Salatrim and needed to be considered in the context of estimates for adults of a mean intake of 11 g Salatrim/day and of a 97.5^th percentile intake of 33g Salatrim/day. Salatrims are intended for use in reduced calorie foods aimed at individuals choosing a diet for the control of weight (Cultor Food Science, 1999). The main target consumers would be adults over 16 years of age and the intake estimates therefore were derived only for adults using commercially available databases on UK food consumption in conjunction with information on the types of food in which Salatrim would be used and the likely levels of inclusion. The application under the Novel Food Regulation is restricted to use in confectionery and baked goods only.

13. The Committee was of the view that the gastrointestinal effects seen were likely to be linked to intolerances to large amounts of this material rather than any specific toxic effect. However, gastrointestinal effects might be more common in children because of their relatively higher nutrient requirement and dietary intake. The Committee further concluded that it was not possible to set any no effect level on the basis of such subjective end points.

Conclusions

i) consumption of Salatrim did not result in any elevation in plasma butyrate levels and thus did not pose any teratogenic risk;

ii) consumption of high doses of Salatrim did result in slight increases in AST and ALT activities in serum, although these increases were within the normal reference range. There were no differences in the trends seen whether those dropping out of the study were included or not. These changes in enzyme activities, in the absence of any other indications of liver damage, were not considered to represent a clear toxic effect. The data from the five-week crossover study in humans provided an additional reassurance of the lack of any clear adverse effect;

iii) the gastrointestinal effects reported were likely to be linked to individual intolerances to large amounts of this material rather than to any adverse toxic effect. Furthermore, it is not possible to set any no effect level on the basis of such subjective end points.

15. The Committee noted that neither children nor those with pre-existing liver disease were included in the trial groups. However, the Committee understands that products containing Salatrim would be aimed at adults choosing a diet for the control of weight.

September 1999

COT Statement 1999/09]

Committee on Toxicity (1997). Statement for the Advisory Committee on Novel Foods and Processes on short and long chain triacyl glycerol molecules (Salatrims) - a family of low calorie fats. Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment, December 1997.

Cultor Food Science (1999). Submission for approval under Regulation (EC) No 258/97 of the European Parliament and of the Council of 27th January 1997 concerning novel foods and novel food ingredients, for the use in bakery and confectionery food products of the family of low calorie fats known as "SALATRIMS". Submitted to the Advisory Committee on Novel Foods and Processes by Cultor Food Science, Redhill, Surrey, U.K., June 1999, with additional material submitted in August 1999.

Nestel PJ, Pomeroy S, Kay S, Sasahara T and Yamashita T (1998). Effect of a stearic acid-rich, structured triacylglycerol on plasma lipid concentrations. Am J Clin Nutr, 68:1196-1201.

Pronczuk A, Lipinski B and Hayes WJ (1999). Butyrate-rich Salatrim induces minimal lipemia without altering fibrinogen or fibrinolysis in men (submitted for publication). In: Cultor Food Science, 1999.

Go to top of page

Back to ACNFP page

This page was last updated on 25 November 1999