Question

The DGIII has asked the Scientific Committee for Medicinal Products and Medical Devices (SCMPMD) to express its opinion on the suitability / safety of the "colours permitted for certain uses only" listed in Annex IV of EEC 94/36 [in particular: E 123 (Amaranth); E 127 (Erythrosin); E 161 (Canthaxanthine); E 173 (Aluminium); E 174 (Silver); E 175 (Gold)] for use in pharmaceutical products and the question of whether the use of these agents might represent a consumer health/safety concern.

The Committee has noted that these colouring agents have already been evaluated as food additives by the Scientific Committee for Food (SCF), which established the following limits for acceptable daily intake (ADI): amaranth, 0.8 mg/kg body weight (bw) (SCF, 1983); erythrosin, 0.1 mg/kg bw (SCF, 1987) and canthaxanthine, 0.03 mg/kg bw (SCF, 1997). The Committee is unaware of any recent information that would necessitate revision of these ADIs.

The question to be examined by the SCMPMD in the present "opinion" regards the following colourant:

E 161 Canthaxanthine

Answer

Given the quantities of the colourant allowed in certain foods, which can be consumed without any restriction whatsoever, it seems paradoxical to prohibit its use at levels that are absolutely negligible in pharmaceutical products, the sale and consumption of which are regulated by law or in any case limited.

Main elements of the scientific justification of the answer

The content of E 161 canthaxanthine in pharmaceutical products (capsules or tablets) that are already on the European market ranges from 0.0049-0.042 mg/capsule or tablet. For a 70-kg man, these doses represent 0.00007-0.0006 mg/kg bw.

The pro kg and pro die doses of this dye that have been perfectly tolerated by various animal species are much, much higher than those that can reasonably be expected to be ingested in pharmaceutical products. On a pro kg basis, even the lowest dose tolerated by animals (canthaxanthine 5.4 mg/kg bw/day) is 9000 times greater than the maximum one dosing unit ingested with drugs, i.e., 0.0006 mg/kg/bw/day . This dose is roughly 50 times lower than the ADI proposed by SCF in 1997 [0.025 mg/kg bw rounded up to 0.03 mg/kg bw], which was based on a no-observed-effect level of 0.25 mg/kg bw/day in humans and a safety factor of 10 (Fed. Reg., 1998).

Full opinion

Terms of Reference

The Committee has been asked to respond to the following question:

Would use of the colourants listed in Annex IV ("colours permitted for certain uses only") of Directive 94/36 (in particular: E123 Amaranth, E127 Erythrosin, E161 Canthaxanthine, E173 Aluminium, E174 Silver and E175 Gold) in medicinal products represent a consumer health/safety concern?

Context of the question

EEC Directive 78/25, which deals with colouring agents that can be used in medicinal products, makes reference to the Directive issued on 23 October 1962 regarding colouring agents in food (OJL 115 f 11.11.1962 p. 2645). However, the EEC policy on food-colouring agents has been updated since then by Directive 94/36. Of particular interest in the latter document are Annex I, which lists all substances approved as food colourants, and Annex IV, which contains 10 agents whose use is restricted to certain foods.

The pharmaceutical industry is questioning the scientific justification for excluding the use of Annex-IV colourants in medicinal products, citing in particular the clause in EEC Directive 78/25 that states "Experience has shown that on health grounds there is no reason why the colouring matters authorized for use in foodstuffs intended for human consumption should not also be authorized for use in medicinal products."

Assessment

The question requires an evaluation of the toxicological characteristics of said colourants in relation to: 1) the maximum quantities and concentrations / unit of weight allowed in foodstuffs, 2) those currently found in pharmaceutical products, and 3) the pro kg amounts that have been well tolerated in the various in vivo toxicity tests.

According to Annex IV of Directive 94/36, the colourants in question may not exceed the following concentrations in foods and beverages:

(*Used in external sugar-based coatings; for decoration of cakes and pastries; coating for chocolates and candies; liquors.)

A brief note from the EMEA (completed with data obtained from Drug Department of the Italian Ministry of Health) provides data on the pharmaceutical products sold in European states that contain the colouring agents in question.

In many cases, the drugs have been on the market since the 1960s.

The maximum amounts of amaranth, erythrosin and canthaxanthine found in pharmaceutical products currently marketed in Europe are:

E 123 (amaranth)

Liquid oral preparations (drops, syrups, etc.) - 0.01-0.6 mg / mL

Capsules - 0.3 mg / capsule

E 127 (erythrosin)

Liquid oral preparations (drops, syrups, etc.) - 0.009-0.08 mg / mL

Capsules, tablets, pills - 0.0017- 0.96 mg / capsule, tablet, pill

E 161 (canthaxanthine)

Capsules, tablets - 0.0049-0.042 mg / capsule or tablet

The quantity pro Kg for a 70-Kg adult male one dosing unit for the above agents ranges from 0.00002 mg/kg bw to 0.013 mg/kg bw.

Pharmaceutical products containing E 173 (aluminium powder) are marketed in Denmark, Germany and Spain. E 174 (silver) is used only in Germany, and E 175 (gold) is found in only one medicinal product that is sold exclusively in Germany.

The following table shows the lowest available estimates of amaranth, erythrosin and canthaxanthine likely to be ingested with foods and beverages. The figures in parentheses are the amount in mg pro kg doses for a 70-kg bw adult male.

E 173 (Aluminium), E 174 (Silver) and E 175 (Gold) are used to decorate cakes, candies and other sweets. For all three, Annex IV of Directive 94/36 allows unlimited use ( quantum satis) since there do not seem to be any toxicological problems at all associated with these substances.

Attention will now be focused specifically on the colourant E 161 Canthaxanthine.

E 161 - Canthaxanthine

Canthaxanthine (?-Carotene-4,4'-dione) is a reddish-orange colouring that can be prepared synthetically or extracted from a variety of natural sources, including algae, Daphnia spp, one species of edible mushroom ( Cantharellus cinnabasinus, from which it was first obtained) crustaceans, including brine shrimps, and fish (Cot, 1987; Hallagan et al., 1995), e.g. wild salmonids (Fed. Reg. 1998). It is a nonprovitamin-A xanthophyll carotenoid that is a more potent antioxidant than beta-carotene (Palozza et al., 1996). Canthaxanthine is one of the ten colourants whose use in foodstuffs is restricted by Annex IV of Directive 94/36. According to Storbakken et al. (1987), 4-8 mg/kg of canthaxanthine may be found in the flesh of aquacultured salmonids raised on a diet containing 80 mg/kg of the colourant. In the EC, canthaxanthine is allowed only in Strasburg sausages at a maximum concentration of 15 mg/kg. However, it is still used at much higher concentrations for dermatologic or cosmetic purposes in oral "sun-tanning" agents.

Toxicity

The LD50 for canthaxanthine in mice is more than 10 g/kg (BIBRA, 1991).

Long-term lifetime oral administration of canthaxanthine at a dose of 125-250 mg/kg bw/day (for 1 year in dogs, 2 years in rats and mice) has not been associated with any overt signs of toxicity (BIBRA, 1991, Hallagan et al., 1995). Serum cholesterol levels were elevated after 52-98 weeks of treatment in animals that received 125 mg/kg bw/day or more. Signs of hepatic enzyme activation were observed in rats, and the females presented increases in the weight of the liver that was not associated with any histological changes (BIBRA, 1991, Hallagan, et al. 1995). Doses of 125 mg/kg or more were associated with reddish discolouration of the intestinal contents, faeces, fur, skin, adipose tissue, extremities and liver.

No significant signs of toxicity were seen in dogs fed up 500 mg canthaxanthine/kg bw/day for 13 or 15 weeks or 250 mg/kg bw/day for 52 weeks. The no-effect dose level in rat studies was 5 mg/kg bw/day (Hallagan et al., 1995).

Mutagenicity

Canthaxanthine is not genotoxic (BIBRA, 1991; Ishidate et al., 1981). In other studies, it was reported to inhibit the activity of a known mutagen (He and Campbell, 1990) as well as protect against the chromosomal instability of transformed cells (Pung et al., 1988; Stich et al., 1990).

Carcinogenicity - Antitumour effects

No carcinogenic effects were noted in rats or mice treated, respectively, for two years and 98 weeks with 1000 mg/kg bw/day (Hallagan et al., 1995). A higher incidence of benign and malignant tumors of the colon was observed in rats treated with canthaxanthine beadlets (roughly 500 mg/kg bw/day for 24 weeks) together with subcutaneous injection of a known carcinogen (Colacchio et al., 1989). However, in other studies it seemed to exert an anticarcinogenic activity (Alam et al., 1988; Gensler and Holladay, 1990; Katsumura et al., 1996; Mathews-Roth, 1982; Mathews-Roth and Krinsky, 1987; Palozza et al., 1997, 1998; Santamaria et al., 1982; Schwartz and Shklar, 1988; Schwartz et al., 1988) which might be related to an antioxidant mechanism. In rats and mice, doses of 200-250 mg/kg bw/day administered for 14-30 days reduced the weight of salivary gland tumors (Alam et al., 1988) and displayed antitumoural activity against cutaneous papillomas (Katsumura et al., 1996), both induced by 9,10-dimethyl-1,2-benz (a) anthracene. At much lower doses (7-14 mg/kg bw/day), it displayed anti-tumour activity against transplantable thymomas in BALB/a mice (Palozza et al., 1997). Canthaxanthine also induces apoptosis in human cancer cell lines ( Palozza et al., 1998). In contrast, no protective effects were seen on gastric epithelial proliferation in Helicobacter mustelae infected ferrets treated for over 2 years with 50 mg/kg, 5 d/w (Yo et al., 1995).

Reproductive toxicology

In 3 generations of rats treated with canthaxanthine 250-500-1000 mg/kg bw/day, there were no signs of reproductive toxicity or teratogenic effects. Doses of 100-200-400 mg/kg bw administered to pregnant rabbits by gastric gavage on days 7-19 of pregnancy did not produce any maternal toxicity or teratogenic effects (Hallagan, et al., 1995).

Immunotoxicity

Dietary administration of canthaxanthine (roughly 100 mg/kg bw/day) to rats for up to 20 weeks resulted in an enhanced immune response in a subsequent in vitro assay (Bendich e Shapiro, 1986).

At concentrations of less than 10-6 M, canthaxanthine did not enhance bursal lymphocyte proliferation of newly hatched chicks in vitro (Haq et al., 1996a). When broiler breeder birds were fed with experimental diets containing 0.04% canthaxanthine (used in these cases to colour the skin), there was no enhancement of the immune response in their offspring as there was when the same type of diet was supplemented with vitamin E (0.03%) (Haq et al., 1996b).

Molecular toxicology

At high doses (300 mg/kg diet) canthaxanthine increases the hepatic content of cytochrome P450, the activity of NADH- and NADPH- cytochrome C reductase, and the P-450-dependent enzymes [(ethoxy-(x139), methoxy(x26)-pentoxy- and benzoxyresorufin O-dealdolases)]; phase II activity is also increased. These enhancing effects were clearly detectable at a dose as low as 10 ppm in the diet (Astorg et al., 1994; Gradelet et al., 1996).

Canthaxanthine retinopathy

Microscopic examination of the eyes of rabbits given canthaxanthine by intravenous route 11.4 mg/kg bw or 0.6 mg/kg bw/day for 19 days revealed effects on the photoreceptor cells and retinal pigment epithelium. In some cases, night vision was also affected. Electroretinography revealed slight changes in visual function (Weber et al., 1987a). No crystal deposits were found in the eyes of rats receiving 250-1000-2500 mg/kg bw/day canthaxanthine for 93-98 week (BIBRA, 1991). Administration of 6 mg/kg bw/day for 11 months to rabbits caused morphological changes in the retina and subtle changes in response to light, although no crystals were evident (Weber, et al., 1987b). No crystalline deposits were found in the retina of dogs treated with 100-500 mg/kg bw/day for 13 or 15 weeks (BIBRA, 1991). In five cats given 2-16 mg/kg bw/day for up to 27 weeks, no crystals were seen in the eyes, but an orange screen developed in the retina in animals that received 4 mg/kg bw/day or more, and at 8 mg/kg bw/day or more, there were structural changes in the retinal epithelium. However, in electroretinography performed after 2 months of treatment, all five cats presented normal visual function (Scallon et al., 1988).

Monkeys treated with doses of 5.4-48.6 mg/kg bw/day by gavage for 2.5 years presented birefringent crystalline-like inclusions in the inner and central retina of canthaxanthine and its metabolites (Goralczyk et al., 1997).

In humans long-term administration (ranging between 3 months and several years) of doses ranging from 16 to 74 g -- in any case, higher than 30 mg/person/ die, or roughly 0.4 mg/kg bw/day) (Goralczyk et al., 1997) -- has been associated with deposition of canthaxanthine crystals ("gold particles") in the retina (Arden e Baker, 1991; Bloomenstein and Pinkert, 1996; JEFCA, 1990; BIBRA, 1991). There is no clear relationship between pigment deposition and dose level or duration of canthaxanthine ingestion (BIBRA, 1991). Kopke et al. (1995) identified a clear dose-response relationship between both total and daily dose and crystal formation in the retina. The minimun daily dose at which crystal formation occured was at least 30 mg with a total intake of about 3000 mg. In most cases, the deposits are not associated with visual problems, but there have been occasional complaints of "dazzle" or reduced visual activity (BIBRA, 1991). In some cases, crystalline retinopathy with reduced dark adaptation on electroretinography (ERG) and decreased vision have been observed (Goralczyk et al., 1997). An asymmetric canthaxanthine retinopathy was observed by Chang et al. (1995) after prolonged use of oral canthaxanthine. The most sensitive indicator of this type of retinopathy is alteration of the ERG b wave (Bloomenstein and Pinkert, 1996), which is a parameter for general retinal damage with decreased vision. Arden and Baker (1991) estimated that the low-effect level for ERG b wave changes was between 0.25-1 mg/kg bw/day.

There is evidence that crystal deposition is reversible after canthaxanthine ingestion has ceased (JEFCA, 1990, Leion et al., 1990). The functional impairment is reversible within a few months in humans (Weber et al., 1987c).

Other adverse effects in humans

A fatal case of aplastic anemia associated with long-term use of high doses of canthaxanthine for "tanning" purposes was observed by Bluhm et al. (1990) in a young woman after taking pills containing approximately 35 mg canthaxanthine/day for 4 months. Other adverse effects associated with repeated exposure to canthaxanthine include hepatitis, generalized itching and urticaria (Bluhm et al., 1990; Herbert, 1991).

Allergic reactions have been reported in patients taking drugs containing canthaxanthine (E 161), but it has not been possible to determine whether the reaction was provoked by the colouring agent or the active ingredient of the product (EMEA, 1998).

Opinion

Given the quantities of E 161 canthaxanthine allowed in Strasbourg sausages (15 mg/kg), which can be consumed without any restriction whatsoever, it seems paradoxical to prohibit the use of this colourant at levels that are absolutely negligible in pharmaceutical products, the sale and consumption of which is regulated by law or in any case limited.

The content of canthaxanthine in pharmaceutical products that are already on the European market is ranges from 0.0049-0.042 mg/capsule/tablet, which represents a pro kg dose of 0.00007-0.0006 mg/kg bw for a 70-kg male. The United States Food and Drug Administration recently declared that canthaxanthine is safe and suitable for intended use in food, with a no-observed effects level in humans of 0.25 mg/kg/day and a safety factor of 10 (Federal Register, 1998).

The pro kg and pro die doses of canthaxanthine that have been perfectly tolerated by various animal species are thousands of times higher than those that can reasonably be expected to be ingested in pharmaceutical products and 50 times lower than ADI. On a pro kg basis, even the lowest rat tolerated dose (5.4 mg/kg bw/day) is 9000 times higher than the maximum one dosing unit that can be ingested in medicinal products and 50 times lower than ADI. Since more than one capsule or tablet may be consumed in a day, the actual margin would be lower, i.e., the dose consumed with 5 capsules or tablets would be roughly 1500 times lower than the no-effect daily doses in rats and 10 times lower than ADI.

References

Alam BS, Alam SQ, Weir jr. JC. Effects of excess Vitamin A and canthaxanthin on salivary gland tumors. Nutr. Cancer 11, 233-241 ,1998.

Arden GB, Barker FM. Canthaxantin and the eye: a critical ocular toxicologic assessment. J Toxicol. Cut & Ocular Toxicol. 10, 115-155, 1991.

Astorg P, Gradelet S, Leclerc J, Canivenc MC. Effects of beta-carotene and canthaxanthin on liver xenobiotic-metabolizing enzymes in the rat. Food Chem. Toxicol. 32, 735-742, 1994.

Bendich A., Shapiro SS. Effects of beta-carotene and canthaxanthin on the immune responses of the rat. J. Nutr. 16, 2254-2262, 1986.

BIBRA Toxicity profile-Canthaxantin, 2nd ed. 1991

Bloomenstein MR, Pinkert RB. Canthaxanthine retinopathy. J Am Optom Assoc. 67, 690-692, 1996.

Bluhm R., Branch R., Johnston P., Stein R., A plastic anemia associated with cantaxanthin ingestion for "tanning" purposes J.A.M.A., 264, 1141-1142, 1990

Chang TS, Aylward W, Clarkson JG, Gass JD. Asymmetric canthaxanthin retinopathy. Am. J. Ophthalmol. 119, 801-802, 1995.

Colacchio T, Memoli VA, Hildebrandt L. Antioxidants vs carotenoids. Inhibitors or promoters of experimental colorectal cancers. Arch. Surg. 124, 217-221, 1989.

CoT Food Advisory Committee, Final Report on the review of the colouring matter in food regulations, 1973. FdAC/REP/4. Annex II - Committee on the Toxicity of Chemicals in Food, Consumer Products and the Environment

European Parliament and Council Directive of 12 December 1977 on colouring matters which may added to medicinal products. (78/25 EC).

European Parliament and Council Directive, of 30 June 1994 on colours for use in foodstuffs (94/36/EC).

European Commission Directive of 26 July 1995 laying down specific purity criteria concerning colours for use in foodstuffs (95/45/EC).

EMEA Doc. Ref. 21831, June 16, 1998.

Federal Register, Listing of color additives exempt from certification: canthaxantin 63, 14814-14817, 1998

Gensler HL, Holladay K: Enhanced resistance to an antigenic tumor in immunosuppressed mice by dietary retinyl plamitate plus canthaxanthin. Cancer Letter 49, 231-236, 1990.

Goralcyk R, Buser S, Bausch J, Bee W, Zuhlke U, Barker FM. Occurrence of birefringent retinal inclusions in cynomolgus monkeys after high doses of canthaxanthin. Invest Ophthalmol Vis. Sci. 38, 741-752, 1997.

Gradelet S, Astorg P, Leclerc J, Chevalier J, Vernevaut MF, Siess MH. Effects of canthaxanthin, astraxanthin, lycopene and lutein on liver xenobiotic-metabolizing enzymes in the rat. Xenobiotica. 26, 49-63, 1996.

Hallagan JB, Allen DC and Borzelleca JF. The safety and regulation status of food, drug cosmetics colour additives exempt from certification. Fd Chem. Toxicol 33, 515-528, 1995.

Haq A, Chinnah A, Bailey CA. Effect of beta-carotene, canthaxanthin, or lutein on lymphocyte proliferation ( in vitro) of newly hatched chicks. Avian Dis, 40, 823-827, 1996a

Haq A, Bailey CA, Chinnah A. Effect of beta-carotene, canthaxanthin, lutein, and vitamin E on neonatal immunity of chicks when supplemented in the broiler breeder diets. Poult. Sci. 75, 1092-1097, 1996b.

He Y, and Campell TC. Effects of carotenoids on aflatoxin B1-induced mutagenesis in S. typhimurium TA 100 and TA 98. Nutrition and Cancer 13, 243-253, 1990.

Herbert V. Canthaxanthin toxicity. Amer. J. Clin. Nutrition 53, 573, 1991.

JEFCA (Joint 35 35th Report of the Joint FAO/WHO Expert Committee on Food Additives), WHO Fd Add. Ser. n. 26.

Katsumura N, Okuno M, Onogi N, Moriwaki H, Muto Y, Kojima S. Suppression of mouse skin papilloma by canthaxanthin and beta-carotene in vivo: possibility of the regression of tumorigenesis by carotenoids without conversion to retinoic acid. Nutr Cancer 26, 203-208, 1996.

Kopcke W, Barker FM, Schalch W. Canthaxanthin deposition in the retina: a biostatistical evaluation of 411 patients. J. Toxicol. - Cut. & Ocular Toxicol 14, 89-104, 1995.

Mathews-Roth MM. Antitumour activity of ß-carotene, canthaxanthin and phytoene. Oncology 39, 33-37, 1982.

Mathews-Roth MM, and Krinsky NI. Carotenoids affect development of UV-B induced skin cancer. Photochemistry and Photobiology 46, 507-509, 1987.

Palozza P, Calviello G, Serini S, Moscato P, Luberto C, Bartoli GM. Antitumor effect of an oral administration of canthaxanthin on BALB/c mice bearing thymoma cells. Nutrition and Cancer. 28, 199-205, 1997.

Palozza P, Luberto C, Ricci P, Sgarlata E, Calviello G, Bartoli GM. Effect of beta-carotene and canthaxanthin on tert-butyl hydroperoxide-induced lipid peroxidation in murine normal and tumor thymocytes. Arch. Biochem. Biophys. 325, 145-151, 1996.

Palozza P, Maggiano N, Calviello G, Lanza P, Piccioni E, Ranelletti FO, Bartoli GM. Canthaxanthin induces apoptosis in human cancer cell lines. Carcinogenesis. 19, 373-376, 1998.

Pungh A, Rundhaug JE, Yoshizawa CN and Bertram JS. ß-Carotene and canthaxanthin inhibit chemically- and physically-induced neoplastic transformation in 10T1/2 cells. Carcinogenesis 9, 1533-1539, 1988

Santamaria L, Bianchi A, Arnaboldi A and Andreoni L. Prevention of the benzo(a)pyrene photocarcinogenic effect by beta-carotene and canthaxanthin. Médecine Biologie Environment 9, 113-120, 1981.

Scallon LJ, Burke JM, Mieler WF, Kies JC, Aaberg TM.Canthaxanthine-induced retinal pigment epithelial changes in the cat. Curr. Eye Res. 7, 687-693, 1988.

SCF (1987), Colouring matters, 21st Report Series, Commission of the European Communities, EUR 11617. Opinion expressed on 10 December 1987.

SCF (1997), Opinion on Canthaxanthin expressed on 13 June 1997.

Scwartz J and Shklar G. Regression of experimental oral carcinomas by local injection of ß-carotene and canthaxanthin. Nutrition and Cancer 11, 35-40, 1988.

Scwartz J and Shklar G, Reid S and Trickler D. Prevention of experimental oral cancer by extracts of spirulina-duanaliella algae. Nutrition and Cancer 11, 127-134, 1988.

Stich HF, Hornby AP, and Dunn BP. Betacarotene levels in exfoliated mucosa cells of population groups at low and elevated risk for oral cancer. Intern. J. of Cancer 37, 389-393, 1986.

Stichk HF, Tsang SS, and Palcic B. The effect of retinoids carotenoids and phenolics on chromosomal instability of bovine papillomavirus DNA-carrying cells. Mutation Research 241, 387-393, 1990.

Storebakken T, Foss P, Schiedt K, Austreng E, Liaaen-Jensen S, Mainz U. Carotenoids in diets for salmonids IV. Pigmentation of Atlantic Salmon with ataxanthin, astaxanthin dipalmitate and canthaxanthin. Aquaculture, 65, 279-292, 1987.

Weber U, Michaelis L, Kern W, Goerz G. Experimental carotenoid retinopathy. II. Functional and morphological alterations on the rabbit retina after acute canthaxanthin application with small unilamellar phospholipid liposomes. Graefes Arch. Clin. Exp. Ophthalmol, 225, 346-350, 1987a.

Weber U, Kern W, Novotny GE, Goerz G, Hanappel S. Experimental carotenoid retinopathy. I. Functional and morphological alterations of the rabbit retina after 11 months dietary carotenoid application. Graefes Arch. Clin. Exp. Ophthalmol. 225, 198-205, 1987b.

Weber U, Goerz G, Kern W, and Michaelis L. Clinical and experimental findings in carotenoid retinopathy. Concepts in Toxicology 4, 105-109, 1987c.

White WS, Stacewicz-Sapuntzakis M, Erdaman JW jr, Bowen PE. Pharmacokinetics of beta-carotene and canthaxanthin after ingestion of individual and combined dosed by human subjects. J. Am. Coll. Nutr. 13, 665-671, 1994.

Yu J, Fox JG, Blanco MC, Yan L, Correa P, Russel RM. Long-term supplementation of canthaxanthin does not inhibit gastric epithelial cell proliferation in Helicobacter mustelae-infected ferrets. J Nutr. 125, 2493-2500, 1995.