Natamycin (pimaricin) is a polyene macrolide antibiotic produced by submerged aerobic fermentation of Streptomyces natalensis and related species. Fermentation is conducted for several days, and the antibiotic is isolated either by broth extraction or by extraction of the mycelium. It is used as a food additive to control the growth of yeasts and moulds on the surface of cheese and other non-sterile products, such as meat and sausages.

Natamycin was evaluated by the Committee at its twelfth and twentieth meetings (Annex 1, references 17 and 41). At its twentieth meeting, the Committee established an ADI of 0–0.3 mg/kg bw. Natamycin was evaluated at the present meeting at the request of the Codex Committee on Food Additives and Contaminants.

The Committee considered information on the current uses of natamycin, biological data not previously evaluated, and data on its intake.

The activity of natamycin against yeasts and moulds, but not bacteria, makes it convenient for use in foods that undergo a ripening period after processing. Its low solubility in water and most organic solvents makes it suitable for the surface treatment of foods. Natamycin is used topically in veterinary medicine to treat mycotic infections, such as ringworm in cattle and horses. Previously, it was used topically against fungal infections of the skin and mucous membranes in humans. Its medical use is now confined to topical treatment of corneal fungal infections and the prevention of such infections in users of contact lens.

The distribution of natamycin was studied by autoradiographic and bioautographic techniques. In the autoradiographic study, five female Wistar rats (TNO, specific pathogen-free) were each given a single dose of 50 mg/kg bw of [¹⁴C]natamycin (50 mg in 5 ml of 1% amylum) orally. In the bioautographic study, four female rats were each given a single dose of 50 mg/kg bw (70 mg in 7 ml of 1% amylum) orally. No information on the purity of the compound was provided. Before treatment, the animals were fasted for 24 h but were given a 5% glucose drinking-water solution. One animal in each group was killed by immersion in a solid CO₂ and acetone mixture under mild ether anaesthesia 1, 2 (autoradiographic study only), 4, 8, and 24 h after treatment. Whole-body sections of the animals were cut in a cryostat at –20 °C. In the autoradiographic study, sections were freeze-dried (48 h) and exposed on photographic plates at –20 °C for 93 days (a few for 150 days). In the bioautographic study, the antibiotic activity of the sections was evaluated by exposure on Whiffen agar plates inoculated with Saccharomyces cerevisiae strain ATCC 9763 for 5, 10, 15, or 20 min (20, 40, 60, and 120 min for sections from the animals killed 24 h after treatment). After exposure, the agar plates were incubated at 30 °C for 20 h and photographed.

In the autoradiographic study, radiolabel was confined to the gastrointestinal tract after 93 days’ exposure (1 h, oesophagus, stomach, small intestine; 2 h, oesophagus, stomach, small intestine, caecum; 4 h, stomach, small intestine, caecum, colon; 8 h, stomach, intestine; 24 h, caecum, colon). After 150 days’ exposure, radiolabel was visible only faintly after magnification of the pictures, in the liver, kidneys, and fatty tissue, in addition to the gastrointestinal tract. In the bioautographic study, the antibiotic activity of natamycin was restricted to the gastrointestinal tract (1 h, stomach, small intestine; 4 h, stomach, small intestine, caecum; 8 h, stomach, small intestine, caecum) and lasted less than 24 h. No antibiotic activity was observed in the colon. The results of the autoradiographic study indicate that natamycin is minimally absorbed into the bloodstream and excreted almost entirely in the faeces. The lack of antibiotic activity and the presence of radiolabel in the caecum and colon 24 h after dosing are consistent with the breakdown of natamycin into microbiologically inactive compounds by bacterial flora in the caecum and colon (Blankwater & Hespe, 1979).

A series of experiments was conducted to study the excretion and resorption of [¹⁴C]natamycin and its degradation products in normal and cholestatic Wistar rats (induced by tying the bile ducts with a ligature). In the first series, the excretion pattern of radiolabelled compound was investigated in groups of three young male Wistar rats given [¹⁴C]natamycin at a dose of 0.1, 1, or 10 mg/kg by quantifying the amount of radiolabel in the urine and faeces at 24-h intervals for 72 h and in expired breath hourly for up to 7 h. Another group received a single dose of 10 mg/kg bw intraperitoneally. A similar experiment was performed in which 10 mg/kg bw of the degradation products of [¹⁴C]natamycin, obtained by acid hydrolysis to simulate that in the stomach, were administered orally or intraperitoneally to three Wistar rats, and their urine, faeces, and expired breath were analysed as described above.

Separate experiments were conducted to determine the elimination of [¹⁴C]-natamycin in the bile by giving 10 mg/kg bw to two rats orally and to four rats by intraperitoneal injection. Bile was obtained via a cannula in the bile duct at 1-h intervals for 7 h and analysed for radiolabel. In a similar experiment, the elimination of [¹⁴C]natamycin via the bile was determined after oral administration of 10 mg/kg bw, in which 0.1 ml of bile obtained from rats not treated with natamycin was placed in the duodenum of treated animals. Bile was collected hourly for 7 h and analysed for radiolabel.

A series of analyses was also carried out to quantify the radiolabel in the stomach, small intestine, caecum, and large intestine of groups of two animals 1, 2, 4, 8, and 24 h after administration of 10 mg/kg bw [¹⁴C]natamycin. Sections of the stomach, small intestine, caecum, and large intestine were extracted in methanol, and the extracts were analysed for radiolabel by thin-layer chromatography. One rat was given the non-radioactive form of the test material and killed after 4 h. The concentration of natamycin was analysed in each section of the gastrointestinal tract by high-performance liquid chromatography. The results were compared with those obtained with the radioactive form. In each experiment, all animals were fasted for 20 h before treatment. Water was available during fasting.

When 10 mg/kg bw [¹⁴C]natamycin were administered orally to normal or cholestatic rats, most of the radiolabel (93–103%) was found in the faeces. Cholestatic rats had about 5% more radiolabel in their urine than normal rats at this dose. The results were similar in rats treated with 0.1 or 1 mg/kg bw. When natamaycin was delivered by intraperitoneal injection at a dose of 10 mg/kg bw, about 16% of the radiolabel was found in urine and about 76% in faeces by 72 h, indicating significant elimination in the bile. Most of the elimination (63%) occurred within 24 h after administration of natamycin. Intraperitoneal administration of acid-hydrolysed [¹⁴C]natamycin resulted in approximately twice as much radiolabel in the urine (61%) as in faeces (30%), showing that hydrolysis transforms natamycin into breakdown products which are more hydrophilic than intact natamycin and thus have less affinity for bile. In contrast, after oral administration of acid-hydrolysed [¹⁴C]natamycin, most of the radiolabel was recovered in faeces (94% as compared with 6.7% in urine); thus, hydrolysis did not appear to result in significant systemic absorption. Little radiolabel associated with either intact or acid-hydrolysed natamycin was eliminated as ¹⁴CO₂ in expired breath (< 1%) after either oral or intraperitoneal administration.

In experiments to determine the amount of radiolabel in bile after an oral or intraperitoneal dose of 10 mg/kg bw [¹⁴C]natamycin, 40% of the total radiolabel was recovered over 7 h from the bile of rats treated intraperitoneally and only 1% from the bile in rats treated orally. When ‘blank’ bile was administered duodenally each hour for 7 h to animals treated orally, the amount of radiolabel recovered in the bile was similar to that recovered in animals not given bile.

In the stomach and small intestine, natamycin was mostly untransformed, as indicated by thin-layer chromatography. Most degradation took place in the large intestine. The degradation products were more hydrophobic than natamycin and were found from about 4 h after treatment. Most of the dose of 10 mg/kg was degraded about 8 h after treatment, suggesting that elimination is relatively rapid. Biotrans-formation was attributed to the bacterial flora in the caecum and small intestine.

Overall, no more than 5–7% of the total radioactive dose was absorbed after oral administration of [¹⁴C]natamycin, and approximately 90% of the administered compound passed through the gastrointestinal tract without resorption and was eliminated in the faeces (Meier & Hespe, 1979).

The resorption and excretion of natamycin were studied by autoradiography in dogs given the compound in plastic coating on cheese at 0.75–0.88 mg/kg, in gelatin capsules at 1.00–1.03 mg/kg, or in a 1% starch suspension at 0.95–1.0 mg/kg. In another experiment, [¹⁴C]natamycin was administered intravenously in 5 ml propylene glycol at a concentration of 1 mg/ml. In the experiments with cheese, a radioactive plastic coating with a natamycin content of approximately 2% was applied to one side of 20-g blocks of cheese. A single batch of natamycin was labelled with ¹⁴C by incorporating labelled sodium acetate as the substrate in the usual fermentation process. The quantity of applied radioactive natamycin was determined by weighing the blocks of cheese before and after application of the radioactive plastic coating. Tests were carried out with blocks that had been stored at 4 °C for various periods. Four female beagle dogs, weighing 10–12.5 kg, were used in these experiments. Three of the four dogs were used in multiple tests, but at least 2 weeks were allowed to elapse between experiments to ensure complete elimination of radiolabelled material from the previous experiment. Before dosing, the animals were fasted for about 16 h but were given drinking-water. Th animals were housed individually in metabolism cages after dosing, and faeces and urine were collected daily for 2–5 days. The samples were processed appropriately, and radiolabel was measured with a liquid scintillation counter. The plastic coating was analysed by thin-layer chromatography to quantify natamycin and reaction products formed during storage.

After oral administration, most of the radiolabel was eliminated in the faeces within 24 h, with less than 4% of the total dose in urine. This pattern of excretion was consistent with all three forms of orally administered natamycin. Storage of the cheese at 4 °C for various lengths of time (1–57 days) had no effect on the pattern of radiolabel observed. Thin-layer chromatographic analysis confirmed the presence of [¹⁴C]natamycin in the cheese coating after up to 56 days at 4 °C. Approximately equal amounts of radiolabel were measured in faeces and urine after intravenous administration of natamycin in propylene glycol, suggesting that resorption occurred via biliary elimination. The amount of radiolabel recovered was < 100% after adminis-tration in an oral capsule or a suspension and was > 100% after administration as a cheese coating. This difference was probably due to uncertainties in the experimental procedure. It is not clear from this study if excretion of natamycin in the faeces after oral administration was due to the lack of absorption from the gastrointestinal tract or to resorption of systemic natamycin via bile (Hespe & Meier, 1980).

Little information was available on the absorption, distribution, excretion, or metabolism of natamycin in humans. No natamycin (< 1 µg/ml) could be detected in the blood after ingestion of 500 mg by human subjects (Anonymous, 1968). This finding corroborates the statement that natamycin is not absorbed from the gut in animals or humans (Raab, 1972).

2.2.1 Acute toxicity

The available data on the acute toxicity of natamycin are summarized in Table 1. They suggest that male rats are more sensitive to the acute toxicity of orally administered natamycin than females (Levinskas et al., 1966). However, in a study by van Eken and Wubs (1976) to determine the LD₅₀ values for natamycin and three of its potential metabolites after intraperitoneal administration to mice, females were more susceptible to the lethal effects (Table 1). The LD₅₀ values of the metabolites of natamycin in this study were higher than that for natamycin, indicating that they are less acutely toxic than the parent compound. LD₅₀ values of 3200, 3700, and > 4000 mg/kg were reported for aponatamycin (n = 2), mycosamine hydrochloride (n = 2), and dinatamycinolidediole (range-finding study), respectively, which are potential metabolites of natamycin.

In rabbits, doses of natamycin ž 500 mg/kg bw caused diarrhoea, and the animals that died had haemorrhagic gastric mucosa. Complexing of natamycin with one-third its weight of a modified polysaccharide increased its toxicity sixfold, and when it was fed to rats natamycin was detected in their blood (Raab, 1972).

2.2.2 Short-term studies of toxicity

Oral administration of natamycin at doses of 50–70 mg/kg bw per day for 5–10 weeks had no effect on the growth, blood, or tissues of rats. A daily oral dose of 150 mg/kg bw for 9 weeks caused some growth inhibition, and a daily dose of 500 mg/kg bw caused 30% of the rats to die within 2 weeks (Struyk, 1958).

Groups of 20 male and 20 female rats were fed diets containing natamycin at a concentration of 0, 125, 500, 2000, or 8000 mg/kg for 94–96 days. None of the five deaths observed could be attributed to treatment. Growth was retarded and food intake was diminished at the two highest concentrations. The results of haemato-logical examinations and organ weights were within normal limits, and no gross or microscopic lesions were found that could be attributed to natamycin (Levinskas et al., 1966).

Groups of three male and three female beagle dogs received diets containing natamycin at a concentration of 0, 125, 250, or 500 mg/kg for 2 years. All but one dog that receiving 250 mg/kg survived for 2 years; the death was unrelated to exposure to natamycin. No effect was seen on food intake, but males receiving the highest concentration did not grow as rapidly as controls initially, and after 15 months, when the dietary intake was reduced, some animals were unable to maintain a satisfactory body weight. The results of haematological and clinical chemical studies revealed no abnormalities. No effects of significance were found on organ weights, and gross and microscopic examination showed no pathological changes (Levinskas et al., 1966).

Groups of two male and two female beagle dogs were given diets containing natamycin at a target concentration of 0, 375, or 750 mg/kg (equivalent to 0, 12, and 25 mg/kg bw per day) for 3 months. The natamycin was obtained in micronized form and was 90.5% pure. The animals were monitored for clinical changes, body weight, food consumption, haematological, clinical chemical, and urinary alterations, electrocardiography (wave intervals and heart rate at weeks 0, 4, 8, and 12), ophthalmology, and pupillary reactions. After being killed by an intravenous overdose of pentobarbital, all animals were necropsied, and the weights of the thymus, heart, liver, kidneys, adrenals, spleen, and testes were measured and gross lesions noted. The tissues preserved in buffered formaldehyde saline and examined microscopically were brain, thyroid, thymus, lung, heart, liver, kidneys, adrenals, spleen, pancreas, lymph nodes, urinary bladder, ovaries, testes, stomach, ileum, colon, jejunum, caecum, and oesophagus. The statistical evaluations included analysis of variance and the Student t test. A signed statement indicated that the study had been conducted in compliance with regulations for Good Laboratory Practice as specified in the Code of Federal Regulations (Title 21, part 58) of the USA and the OECD. A signed and dated quality assurance statement indicated that the findings had been audited throughout the study.

No dose- or treatment-related effects were seen in males or females with respect to mortality rate, food consumption, body weight, haematological, clinical chemical, or urinary end-points, electrocardiography, ophthalmology, absolute and relative organ weights, gross pathology, and histopathology. The only effect attributed to treatment was diarrhoea, which occurred most frequently in animals at the high dose but was also observed in controls and animals at the low dose. The diarrhoea was attributed to local irritation of the gastrointestinal tract. Because of the frequent occurrence of diarrhoea at 750 mg/kg, the authors noted that it would be difficult to expose animals to higher doses. No NOEL could be identified (van Eeken et al., 1984).

2.2.3 Long-term study of toxicity and carcinogenicity

Groups of 35–40 male and female rats received diets containing natamycin at a concentration of 0, 125, 250, 500, or 1000 mg/kg for 2 years. The animals remained in good health, and their survival was unaffected by treatment. Inhibition of growth rate and diminished food intake were seen only for animals of each sex receiving the highest concentration. The results of haematological investigations and determination of organ weights and gross and microscopic lesions showed no differences between treated and control groups. The numbers and types of tumours found in natamycin-treated rats were not significantly different from those in untreated animals (Levinskas et al., 1966).

2.2.4 Genotoxicity

Studies were conducted to evaluate the mutagenic potential of natamycin, its products of degradation (i.e. aponatamycin, natamycinolidediol, and mycosamine hydrochloride), and Delvocid (a 50% suspension of natamycin in water) in Bacillus subtilis, Salmonella typhimurium, and Escherichia coli. B. subtilis was exposed in a standard rec assay (spot diffusion method) according to Kada (citation not provided). E. coli strains WP2uvrA^– and WP2 and S. typhimurium strains TA1535, TA1538, TA98, and TA100 were evaluated in the spot test for reverse mutation. All the tests were carried out by plating a 50-µl spot containing the appropriate dilution of natamycin on a petri dish with the appropriate microbial strain. The spot tests were carried out within 3 h of exposure and after storage for 1, 3, 7, or 14 days and 1, 2, or 4 months (It was not clear whether all tests were conducted at all intervals). The plate incorporation assay was used to evaluate the mutagenicity of Delvocid at concentrations up to 1% alone (without addition of an exogenous metabolic activation system from rodent liver) and in combination with up to 0.2 mol/L nitrite in E. coli WP2uvrA^– and WP2 trp^– and S. typhimurium TA98 and TA100, with or without addition of exogenous metabolic activation. Nitrite was tested with Delvocid, as other studies have shown that nitrite in combination with some food preservatives forms reaction products that interact with DNA. The design of these tests is shown in Table 2. In each spot test, negative controls were included with solvent or buffer alone and positive controls with a known mutagen (N-methyl-N´-nitro-N-nitrosoguanidine) or mixtures of sorbic acid and nitrite at pH 3.5 or 4.5. For the plate incorporation assays, benzidine was used as the positive control. No statistical analyses were reported. The author reported that no positive responses were observed in the spot tests in any of the three test systems. Results for individual plates and summary data were not reported for the spot tests. The authors concluded from the plate incorporation assays that Delvocid did not induce reverse mutation when tested alone or with nitrite in any of the strains of S. typhimurium or E. coli tested. The author commented on the slight positive response with nitrite at about 0.2 mol/L and concluded that Delvocid did not enhance the mutagenic effect (Khoudokormoff, 1977; Khoudokormoff & Gist-Brocades, 1978).

The Committee noted that the reporting of the results of these studies, described as preliminary, had limitations which prevented verification of the author’s conclusions. For example, the bacterial strains used were not assayed for the appropriate phenotypic markers or plasmids, the criteria for a positive response were not reported, summary and individual data were not reported for the spot test, and no statistical analyses were performed. Sufficient information was not provided to indicate that the studies were adequately sensitive to detect positive responses in all strains tested. Furthermore, the assays were conducted in a single trial with one plate per dose. The usefulness of these studies is therefore limited.

Natamycin at a concentration of 1% and its known degradation products (aponatamycin, dinatamycinolidediol, and mycosamine) at 0.5% and at pH and nitrite conditions similar to those in preserved food products such as cheese and sausages, were reported to have no mutagenic activity in B. subtilis under the conditions tested. No actual data were presented to verify this statement (Khoudokormoff, 1978).

Groups of 10 male rats taken from the second litters of the F₁ generation in a three-generation study of reproductive toxicity (see below) were fed on control diet until sexually mature, when they received natamycin at 0, 5, 15, 50, or 100 mg/kg bw daily for 7 days by gastric intubation. Each rat was mated each week for 8 consecutive weeks with two virgin untreated females. Each female was killed and examined 13 days after mating. No differences were found between control and test animals in respect of the numbers of implantation sites or live or dead fetuses or the mutagenic index (Cox et al., 1973).

Five males and five females were selected at random from the five litters produced in the same three-generation study. The animals were given colchicine 3–4 h before being killed, and a bone-marrow preparation was made for examination for aberrant chromatin material. The number of abnormalities in the metaphase chromosomal preparations of test groups did not differ significantly from that in sham-treated controls (Cox et al., 1973).

2.2.5 Reproductive toxicity

Groups of 10 female and five male rats receiving diets containing natamycin at a concentration of 0 or 1000 mg/kg were mated after 181 and 223 days. Other groups were mated after 48, 174, and 250 days on the diets; four control and four test female young from the second mating were fed on the same diet as their parents and mated when 107 days of age. The pups of natamycin-treated animals had lower mean body weights at weaning than control pups, but examination of the results of the 54 matings showed that their fertility, gestation, lactation, and viability indices were similar to or better than those of the controls. There was a low incidence of abnormalities among pups in this study, but none could be attributed to treatment (Levinskas et al., 1963; Levinskas, 1966).

Groups of 10 male and 20 female rats were given a diet containing natamycin providing a dose of 0 (two groups), 5, 15, 50, or 100 mg/kg bw per day for 11 weeks. These formed the F₀ generation of a three-generation study of reproductive toxicity, two litters being produced in each generation. Animals at 100 mg/kg had an increased number of fetuses born dead, a decrease in the number born alive, and a decrease in the number surviving at 21 days. The weight of pups was depressed in the second litters of the F₀ and F₁ generations and both litters of the F₂ generation. However, the fertility, gestation, viability, and lactation indices were within normal limits for both litters of all three generations. The doses of 5, 15, and 50 mg/kg had no detectable effect on growth or reproduction (Cox et al., 1973).

Groups of 20 female rats from the second litters of the F₁ generation of the three-generation study of reproductive toxicity were reared to maturity on control diet and mated with untreated males. The females were given the same dose of natamycin as their parents (0, 5, 15, 50, or 100 mg/kg bw per day) by gastric intubation during the 6–15 days of gestation and were killed and examined on day 20. No differences were found between control and test animals in respect of the numbers of pregnancies, live litters, implantation sites, resorption sites, live and dead fetuses, or skeletal and soft tissue abnormalities (Cox et al., 1973).

Groups of 10–12 female rabbits were given natamycin at a dose of 0, 5, 15, or 50 mg/kg bw per day by gavage on days 6–18 of gestation. They were examined on day 29, and the numbers of corpora lutea, implantation sites, resorption sites, and live and dead fetuses were recorded. No adverse effects on nidation or maternal or fetal survival were found. The number of abnormalities seen in the soft or skeletal tissues did not differ from that occurring spontaneously in controls (Bailey & Morgareidge, 1974).

An aqueous suspension of Delvocid (50% natamycin) was administered to groups of 20–26 mated female Dutch belted rabbits by gavage at a dose of 5, 15, or 50 mg/kg bw per day on days 6–18 of gestation. Two control groups were used: a vehicle control that received an equal volume of sterile saline daily by gavage on days 6–18 of gestation and a positive control group given 2.5 mg/kg bw of 6-aminonicotinamide by gavage on day 9 of gestation. The does were observed daily for signs of toxicity, and body weights were recorded on days 0, 6, 9, 12, 15, 18, and 29 of gestation. On day 29, all surviving does were killed, and the numbers of corpora lutea, implantation sites, resorption sites, and live and dead fetuses, sex of fetuses, and fetal body weights were evaluated at autopsy. The survival rate of the fetuses was determined, and they were examined for external, soft-tissue, and skeletal anomalies. According to the protocol, the study was conducted in compliance with proposed Good Laboratory Practice regulations (21 CFR 3), but there is no indication as to whether quality assurance or quality control procedures were in place.

Treated does showed no clinical signs of toxicity. One at the low dose, two at the intermediate dose, and five at the high dose died or were killed when moribund. Accordingly, the maternal mortality rates were 0% (0/20), 5% (1/20), 9% (2/22), and 19% (5/26) in the four groups, respectively. The cause of these deaths was not indicated in the report. One doe at the intermediate dose delivered young prematurely (the day before scheduled removal). There were no clear treatment-related signs of toxicity. The following parameters were comparable in treated groups and the vehicle control group: mean maternal body weight, pregnancy rate, number of implantation sites, number of resorption sites, numbers of live and dead fetuses, male to female ratio of fetuses, per cent viability, and incidence of soft-tissue anomalies. Maternal body-weight gain was not calculated. In addition, although the number of corpora lutea in each doe and the occurrence of external anomalies were determined, these data were not summarized or analysed statistically. The average body weight of live fetuses in the group at the intermediate dose was significantly lower than that of the vehicle control group. The groups at the two higher doses showed a significant increase relative to the vehicle control group in the incidence of extra sternebrae. The authors noted that the effect on fetal body weight was not dose-related, and they considered the extra sternebrae to be a developmental variation and not an indication of frank teratogenicity (Knickerbocker & Re, 1978, 1979).

The results of this study were difficult to interpret owing to maternal mortality, problems associated with gavage of rabbits, and because the digestive system of rabbits is sensitive to antibiotics. However, there was evidence that the extra sternebrae observed in fetuses of does at the intermediate (15 mg/kg bw per day) and high (50 mg/kg bw per day) doses of natamycin were variations rather than malformations (Manson & Kang, 1994). Consequently, this study was not considered suitable for deriving an ADI.

2.2.6 Special studies

No allergic sensitization occurred among 111 patients being treated with natamycin for a variety of conditions (Gruyper, 1961, 1964). No history of allergic reactions was found in 73 workers engaged for an average of 5 years in the manufacture of natamycin, and no allergic reactions were found in the 71 who were tested with cutaneous or intradermal challenge doses (Malten, 1967). Repeated patch tests on 102 patients with various forms of eczema failed to demonstrate any sensitizing potential of natamycin (Malten, 1968).

Similar breakdown products of natamycin occur in simulated gastric juice, 0.5% citric acid, and urine, and it appears likely that breakdown products in stored apples resemble those produced in gastric juice. The breakdown products are tetraenes related to natamycin, principally aglycone dimerized and/or decarboxylated; whether these are absorbed remains to be tested (Brik, 1975). Approximately 50% natamycin is broken down within 1 h in simulated gastric juice, and the losses from the stomachs of fasted and non-fasted rats were 33–43% and 0–31% respectively (Morgenstern & Muskens, 1975).

The results of studies of the acute toxicity of the decomposition products of natamycin kept under various conditions after intraperitoneal administration to mice are presented in Table 3.

Groups of 15 male and 15 female rats were given diets containing 5% water, 5% of 0.5% citric acid, 500 mg/kg natamycin, or 5% of a solution of acid-degraded natamycin (suspended in 0.5% citric acid until only 14% of the activity remained) for 98 days. No animals died, and their weight gain was unaffected by treatment; no adverse effects were seen in haematological tests or on the absolute weights of the liver and kidneys. Minor differences in relative organ weights were considered to be coincidental and not due to treatment. Microscopic examination of a wide range of organs showed no lesions due to treatment (Hutchison et al., 1966).

Slices of cheese were treated with 0.05% and 5% suspensions of natamycin and left to dry at room temperature. The antimicrobial activity of the two cheeses declined to less than 20% and 60–80% during the 3-week storage period before they were incorporated into rat diet, and the final dietary concentrations of natamycin plus degraded natamycin were 3.6 and 360 mg/kg. Groups of 15 and 30 male and female rats received diets containing fresh cheese dressed with 0, 0.05, or 5.0% natamycin or cheese dressed with 0, 0.05, or 5% suspensions and stored for 3 weeks. The test lasted 7 weeks. No effects that could be attributed to natamycin degradation products were found on behaviour, appearance, morbidity, mortality, food consumption, body-weight gain, haematological indices, liver function, organ weights, or macro- or microscopic appearance of the animals (Wieriks, 1966).

Groups of 10 male and 10 female rats were fed for 3 months on diets containing the peel of apples which had been untreated, freshly treated with natamycin, or treated with natamycin and stored for 2–8 weeks to allow degradation to take place. In a similar experiment, sausage skins untreated, freshly treated, or stored with natamycin were fed to rats. The doses of natamycin and its degradation products cannot be calculated, but the apple-skin diet provided rats with approximately 0, 50, and 1250 times the probable human intake, and the sausage-skin diet provided approximately 0, 1000, and 25 000 times the human intake. Some minor abnormalities were found, but none related to growth rate, mortality rate, haematological indices, serum enzymes, liver function, organ weights, or gross or microscopic appearance could be attributed to the intake of natamycin breakdown products (Wieriks, 1971).

2.2.7 Microbiological effects

Limited information on the microbiological effects of natamycin, including fungal resistance, was included in the previous monograph (Annex 1, reference 42). In that monograph, it was stated that natamycin is active against a wide range of mycotic organisms such as dermatophytes and other fungi, yeasts, and yeast-like organisms (including strains pathogenic to humans, animals, and plants and saprophytic varieties). Standard tests have shown that it has no activity on bacteria or on actinomycetes. There is no evidence that mycotoxin-forming species are unusually resistant to natamycin (Raab, 1972). No yeast or yeast-like organisms have been reported to have primary resistance to natamycin, although some dermatophytes are resistant. It is more difficult to induce resistance to natamycin in yeasts than in bacteria (Khoudokmoff & Petru, 1974), and the resistance that could be obtained appeared to be due to selection of naturally more resistant strains and not to adaptation. The resistant cultures had reduced pathogenicity (Athar & Winner, 1971). No evidence of resistance has been recorded in clinical use of natamycin. In studies of its cross-resistance with other antimicrobials, amphotericin B but not natamycin showed cross-resistance with nystatin, filipin, endomycin, and candidin (Stout & Pagano, 1956; Littman et al., 1958; Bodenhoff, 1968; Walter & Heilmeyer, 1969). Nystatin- and amphotericin-B resistant organisms were susceptible to natamycin (Sörensen et. al., 1959), and a wide selection of nystatin-resistant yeasts were normally susceptible to natamycin (Hejzlar & Vymola, 1970). Cross-resistance between natamycin and nystatin and amphotericin appeared to occur in vitro (Athar & Winner, 1971).

More information on fungal resistance has become available since the previous review, and that pertinent for assessing potential resistance, including a discussion of the mechanism of action of polyene antibiotics and fungi in the human gastrointestinal tract, is summarized below.

The polyenes constitute a large group of antibiotics with various molecular structures, which interact with fungal membranes in an especially interesting way (Franklin & Snow, 1998). The approximately 200 polyenes are all produced by Streptomyces spp. The antifungal activities of natamycin and other polyenes are dependent on their binding to cell membrane sterols, primarily ergosterol, the principal sterol in fungal membranes, thereby making them leaky (Hamilton-Miller, 1974; Norman et al., 1976; McGinnis & Rinaldi, 1985; Carlile & Watkinson, 1994). As polyene macrolide antibiotics like amphotericin B, nystatin, and natamycin have a much greater affinity for ergosterol than for cholesterol, the mammalian membrane sterol, they are selectively antifungal. The polyenes form complexes with sterols and apparently disrupt membrane function by this mechanism. The oomycete fungi and bacteria are insensitive to these antibiotics because their membranes lack sterols. At low concentrations, selective changes in membrane permeability may occur. Leakage of potassium ions is the first detectable event, and, at high concentrations, leakage of amino acids and other metabolites occurs.

The polyenes have a large lactone ring with a rigid lipophilic chain containing three to seven conjugated double bonds and a flexible hydrophilic portion bearing several hydroxyl groups. The length of the chromophore gives the characteristic ultra-violet spectrum for each compound and contributes to the instability of some polyenes to heat, light, and pH. Most polyenes have a sugar unit, typically the amino sugar mycosamine, which is linked by the glycosidic bond to the  carbon atom of the chromophore. Amphotericin B contains seven conjugated double bonds, and natamycin contains four, so these antimicrobial agents are known as heptaenes and tetraenes, respectively. Nystatin is classified as either a pseudoheptaene or a tetraene (McGinnis & Rinaldi, 1985).

The typical polyene structure has both a hydrophobic and hydrophilic face. The polyenes insert themselves into the cell membrane by associating with sterols (the hydrophobic face) and are thought to cause rearrangement of the sterols, so that a group of four to eight polyene molecules forms a ring with the hydrophilic faces in the centre. Thus, they form a polar pore through which small ions like K⁺ and H⁺ can pass freely, disrupting the cell’s ionic control (Griffin 1994; Deacon 1997). Polyenes can also directly affect enzymatic sequences involved in the synthesis of membrane constituents at the level of the early cyclic precursors in the ergosterol biosynthetic pathway (Mukhtar, et al., 1994). The accumulation of these precursors results from a decrease in the trans-methylation reaction that requires S-adenosylmethione as the donor of the methyl group and zymosterol as the substrate for methylation. Bacteria are not susceptible to natamycin as their membranes are devoid of sterols. Accordingly, the reported minimum inhibitory concentrations (MICs) of natamycin against bacteria are high, those for Staphylococcus aureus, Streptococcus faecalis, Streptococcus haemolyticus, Bacillus cereus, Bacillus subtilis, Escherichia coli, Salmonella typhimurium, Proteus mirabilis, and Pseudomonas aeruginosa all being > 250 mg/kg.

The microflora in the human gastrointestinal tract form an extremely complex, yet relatively stable ecological community, populated with over 10¹¹ bacterial cells per gram of content and containing more than 400 bacterial species (Cerniglia & Kotarski, 1999). There are fewer bacteria than fungi. Up to 10⁵ colony forming units of yeasts have been reported in stool samples from healthy subjects (Bernhardt, 1998).

The yeasts found are Candida albicans (the commonest), C. glabrata (Torulopsis glabrata), C. tropicalis, C. guillermondii, C. krusei, C. inconspicua, C. parapsilosis, C. lusitaniae, and C. kefyr (C. pseudotropicalis). Rhodotorula spp., Trichosporon, Saccharomyces cerevisiae, Geotrichum candidum, Aspergillus spp., Cryptococcus spp., and Mucor spp. are rarely found in the intestine (Bernhardt, 1998). The metabolic activity of Candida spp. in the gastrointestinal tract is very low owing to the anaerobic conditions and limited nutrients. Yeasts of the normal flora can invade the tissues of patients whose immune defences have been suppressed by disease or in persons with an altered intestinal microflora. The therapeutic use of antimicrobials can suppress the normal bacterial flora, and this is responsible in part for the increase in the number of yeast infections, particularly gastrointestinal candidosis (Blaschke-Hellmessen et al., 1996; Kreisel, 1999). The polyenes are not absorbed from the gastrointestinal tract, but are sometimes given by mouth to combat fungal growth in the gut, which most commonly results from the use of broad-spectrum antibacterials that deplete the normal bacterial flora of the gut and allow yeasts and fungi to multiply and cause opportunistic infection (Scheurlen, 1996).

The use of natamycin as an antifungal agent in food may result in trace quantities of antimicrobial residues that interact with endogenous microflora. No data were available on the effect of natamycin on the human intestinal microflora. As bacteria are not affected by polyenes, it can be concluded that natamycin residues would have no potentially harmful effects. Furthermore, as yeasts are found in small quantities in the human gastrointestinal tract, the risk of trace exposure of fungi to natamycin would be minimal.

Natural resistance against polyenes such as natamycin does not occur among fungi, because of the mode of action of these chemical agents (Khoudokormoff, 1984). Moreover, in contrast with the main polyenes used clinically, such as amphotericin B and nystatin, the fungistatic and fungicidal minimal concentrations of natamycin differ only negligibly (Table 4), further reducing the opportunity for establishment of resistance (Sorensen et al., 1959). Induction of polyene- and especially natamycin-resistant mutants is difficult (Athar & Winner, 1971). Such mutants invariably show reduced metabolic and growth rates in vitro, and in the absence of polyenes readily revert to normal metabolism, growth, and sensitivity to natamycin. One way of obtaining such resistant isolates is by successive sub-culturing in vitro in the presence of gradually increasing concentrations of a polyene. Typically, such isolates are resistant up to the highest concentration to which they are exposed, and the conditions are not likely to result from technical application of natamycin as a food preservative.

The antifungal action of polyene antibiotics is based on their linkage with sterols in the cytoplasmic membrane of the fungal cell wall, which distends the wall. The sensitivity of fungal cells to the drug depends on the characteristics of the sterol (Littman et al., 1958; Molzahn & Woods, 1972; Subden et al., 1977). Candida strains resistant to nystatin contain more ergosterol than sensitive ones (Athar & Winner, 1971; Safe et al., 1977). Sensitivity to polyene antibiotics is a consistent feature of wild-type fungal strains. Prolonged therapy with an antibiotic results in increased resistance to it. Induced resistance to polyene antibiotics has been observed in Candida, Torulopsis, and Cryptococcus strains (Macura, 1991).

Although there is a potential risk of development of resistance among microbial flora as a consequence of prolonged, repeated application of natamycin, the studies reported indicate that the level of resistance would be low.

Attempts to induce resistance to natamycin in C. albicans by serial passage on Sabouraud maltose agar showed that resistance developed gradually. After 25 passages, the MIC was increased from 2.5–12 to 12–50 µg/ml. Comparison of the polyene antibiotics natamycin and fungicidin indicated that strains that are resistant to fungicidin are sensitive to natamycin (Hejzlar & Vymola, 1970; Table 5).

Natamycin has been given orally for the treatment of intestinal candidosis at a daily dose of up to 400 mg. It was highly active against yeast-like fungi (MIC, 1.5 µg/ml) but less effective against dermatophytes (MIC, 3.0–100 µg/ml). Strains resistant to natamycin are rare, but the effectiveness of this drug in the treatment of vaginal candidosis has decreased (Lovgren & Salmela, 1978). The MIC values were between 2.9 and 31 µg/ml for strains isolated from untreated women but 9.8–64 µg/ml for strains from women who had been treated previously.

Delvocid, a 50% natamycin preparation, has been used for more than 20 years for preserving cheese and sausages (Jay, 1996). Surveys in cheese warehouses and in dry sausage factories where Delvocid had been used for up to 9 years showed no change in the composition or the sensitivity of the contaminating fungal flora (de Boer & Stolk-Horsthuis, 1977; de Boer, 1979; Hoekstra & Van der Horst, 1998).

de Boer and Stolk-Horsthuis (1974) isolated yeasts and moulds from various cheese warehouses in which natamycin was used. All of the isolated fungi but one were inhibited at low concentrations of natamycin (0.5–8 µg/ml). In a similar study in 1976, in which eight warehouses where natamycin had been used and two in which it had never been used were surveyed, 26 strains were isolated and tested for sensitivity to natamycin; no insensitive yeasts or moulds were found.

Laboratory experiments intended to induce tolerance to natamycin in strains isolated from cheese warehouses indicated that after 25–30 transfers to media containing increasing concentrations of natamycin none of the strains had become less sensitive to natamycin (Table 6).

The sensitivity to natamycin of yeasts and moulds isolated in dry sausage factories where natamycin had been used for several years was compared with that of isolates from factories where natamycin had never been applied. No significant differences were found (de Boer et al., 1979).

In experiments with the plant pathogens Cladosporium cucumerinum and Fusarium oxysporum, the frequency of emergence of resistance was 1 in 10⁷. Eighteen resistant strains were selected for further study, and the natamycin-resistant strains were divided into those with lower and higher levels of resistance. Greater resistance appeared to be associated with decreased fitness in vitro (radial growth and sporulation on agar media) and in vivo (pathogenicity). The authors suggested a link between increased resistance and decreased fitness (Dekker & Gielink, 1979).

Reduced sensitivity to polyenes can be induced in vitro, but this may be of no practical significance. The resistance of several subcultures in the presence of increasing concentrations of a polyene antimycotic was associated with slower growth and diminished virulence, so that any resistant cells that appear during polyene antimycotic treatment may succumb to the body defence mechanisms.

Nausea, vomiting, and diarrhoea have been observed occasionally after an oral dose of 300–400 mg of natamycin daily; no changes in peripheral blood cells were observed (Anonymous, 1966). A group of 10 patients with systemic mycoses received oral doses of 50–1000 mg/day for 13–180 days. Nausea, vomiting, and diarrhoea occurred in those receiving 600–1000 mg/day (Newcomer et al., 1960).

Natamycin is proposed in the Codex draft General Standard for Food Additives (GSFA) for use in food groups 1.6 ‘Cheese’ at 40 mg/kg, in 8.2.1.2 ‘Cured and dried non-heat treated processed meats, poultry and game products’ at 6 mg/kg, and in 8.3.1.2 ‘Cured and dried non-heat comminuted meat, poultry and game products’ at 20 mg/kg.

Data on intake of natamycin were submitted by Australia, Germany, New Zealand, and by the manufacturer, whose submission included limited data on intake in the United Kingdom and the USA.

The budget method can be used to assess whether the use of natamycin should be restricted to specific food groups. The calculations indicated that the theoretical maximum level of use of natamycin is 24 mg/kg, assuming it is used in only half the solid food supply and that the ADI is 0–0.3 mg/kg bw. This theoretical level is lower than the proposed level of use in cheese in the draft GSFA but higher than that proposed for meats, indicating that use of natamycin should be restricted.

As natamycin is proposed for use in two single food groups, the reverse budget method can indicate the maximum amount of each food group that can be consumed before the ADI is exceeded, assuming use in only one food group. If use is assumed to be only in cheese, up to 450 g could be consumed per day at a concentration of 40 mg/kg, assuming an ADI of 0–0.3 mg/kg bw and an average body weight of 60 kg. For cured meats, up to 900 g could be consumed per day.

Consumption of these amounts of either food group on a daily basis is unlikely. The maximum amounts reported for cheese consumers were 99 g/day for the Australian population and 108 g/day for the New Zealand population at the 95th percentile of consumption, 62 g/day for the adult population of the United Kingdom at the 97.5th percentile of consumption, and 45 g/day for the population of the USA at the 90th percentile of consumption. Dietary records do not distinguish whether the cheese consumed had been cut, shredded, or grated but report all cheese consumed, either directly or indirectly in mixed foods.

The maximum amounts reported for consumption of comminuted meat were 170 g/day for the Australian population and 210 g/day for the New Zealand population, both at the 95th percentile of consumption.

The intake of natamycin estimated from individual dietary records was available for five countries (Table 7). The estimates were all well below the ADI, when either draft GSFA or national use levels were assumed.

The estimates based on proposed GSFA levels of use from Australia and New Zealand were for cheese and meat sources. The mean intake for Australian consumers was estimated to be 0.026 mg/kg bw per day, or 9% of the ADI; that for New Zealand consumers was 0.022 mg/kg bw per day, or 7% of the ADI. Cheese contributed 72% of the total natamycin intake in Australia and 67% of that in New Zealand.

Estimates of the intake of natamycin in the United Kingdom and the USA were based on the proposed GSFA use in cheese. The mean estimated intakes were slightly lower than those reported for Australia and New Zealand: 0.014 mg/kg bw per day, or 5% of the ADI, for consumers in the United Kingdom and 0.015 mg/kg bw per day, or 5% of the ADI, for consumers in the USA. The intakes of consumers at high percentiles based on draft GSFA levels of use of natamycin ranged from 0.03 to 0.08 mg/kg bw per day (11–27% of the ADI).

National estimates of natamycin intake, submitted by Australia, Germany, New Zealand, and the USA, were all well below the ADI (Table 7), the mean intakes of consumers being 0.008–0.015 mg/kg bw per day (2.5–5% of the ADI), as the national permitted levels of use were much lower than those proposed in the draft GSFA, and use of natamycin was further restricted: 15 mg/kg in cheese and 1.2 mg/kg in salami in Australia and New Zealand; 20 mg/kg in cheese, 6 mg/kg in meats (8.2.1.2), and 20 mg/kg in meats (8.3.1.2) in Germany; and 20 mg/kg in cheese in the USA.

The submissions indicated that the intake of natamycin was well below the ADI and that the ADI was not likely to be exceeded even by consumers at high percentiles. The higher estimates for consumers at high percentiles in Australia and New Zealand (27% and 23% of the ADI, respectively) were due to use of single 24-h recall data, which tend to result in overestimates of the habitual intake of consumers at high percentiles. In the surveys of food consumption in the United Kingdom and the USA, the amounts were averaged over a number of days (3 and 7, respectively), which would tend to decrease the reported daily consumption of all foods and of occasionally consumed foods, such as salami type meats, in particular (Gibney, 1999; Lambe et al., 2000).

Toxicological data

The Committee considered eight studies that had not been evaluated previously; these studies had been conducted more than 20 years earlier. A study of single intraperitoneal administration was considered to be irrelevant to the safety assessment of an ingested substance. The results of two studies of genotoxicity in three bacterial systems (Bacillus subtilis, Salmonella typhimurium, and Escherichia coli) were negative.

Two studies in rats and one in dogs given radiolabelled material for investigation of the distribution and elimination of the compound supported the previous conclusion that natamycin is excreted primarily in the faeces, with minimal absorption. The only adverse effect reported in a short-term study of toxicity in dogs was diarrhoea, which occurred most frequently in animals at the high dose (equivalent to 25 mg/kg bw per day); however, the usefulness of this study was limited as only two dogs were tested.

In a study of developmental toxicity, an aqueous suspension of 50% natamycin was given to groups of 20–26 mated rabbits at a dose of 0, 5, 15, or 50 mg/kg bw per day by gavage on days 6–18 of gestation. The maternal mortality rate was 0%, 5%, 9%, and 19% at the four doses, respectively. No clinical signs of toxicity were observed in the does, and the cause of death was unknown. Mean maternal body weight, pregnancy rate, number of implantation sites, number of resorption sites, numbers of live and dead fetuses, per cent viability, and incidence of soft-tissue anomalies were comparable in treated groups and a control group given the vehicle only. The fetal body weight at the intermediate dose was lower than that of the vehicle control group. The incidence of extra sternebrae was increased at the two higher doses in comparison with the vehicle control group, but not in a dose-related manner. However, in view of the known and unusual sensitivity of the gastrointestinal tract of rabbits to poorly absorbed substances and to compounds with antimicrobial activity, this study was not considered suitable for deriving the ADI.

Microbiological data

The antifungal activities of natamycin and other polyenes depend on their binding to cell membrane sterols, primarily ergosterol, the principal sterol in fungal membranes. Oomycete fungi and bacteria are insensitive to these antibiotics because their membranes lack ergosterol.

Use of natamycin as an antifungal agent in food may result in exposure of the endogenous microflora to trace quantities of antimicrobial residues. The human intestinal microflora are a complex mixture of more than 400 bacterial species, composed primarily of bacterial cells at a concentration of 10¹¹–10¹² colony forming units per gram. Fungi are much less abundant in the human gastrointestinal tract than bacteria, up to 10⁵ colony forming units per gram of yeast being reported in stool samples from healthy subjects. As bacteria are not affected by polyenes, natamycin residues should not harm them, and as yeasts are found in low quantities the consequences of exposure to traces of natamycin would be minimal.

Several studies in experimental animals indicated a lack of antibiotic activity in the colon, suggesting that natamycin was degraded into microbiologically inactive compounds by bacterial flora. However, no data are available on the degradation of natamycin by human intestinal microflora. In one study, natamycin was present in faecal specimens of volunteers who ingested 500 mg of the compound, indicating that the compound is incompletely absorbed or degraded.

As emergence of antibiotic resistance is a concern, the Committee evaluated the possible development of resistance among microflora as a consequence of ingestion of natamycin. A 50% natamycin preparation has been used for over 20 years to preserve cheese and sausages. Surveys in cheese warehouses and in dry-sausage factories where the preparation has been used showed no change in the composition or the sensitivity of the contaminating fungal flora. All but one of the species of yeasts and moulds isolated in cheese warehouses where natamycin was used were inhibited at the same low concentrations (0.5–8 µg/ml). In another study, 26 fungi were isolated in eight warehouses where natamycin was used and two warehouses where it never had been used and tested for sensitivity to the compound; no insensitive yeasts or moulds were found. The results of laboratory experiments intended to induce resistance to natamycin in strains isolated from cheese warehouses indicated that, after 25–30 transfers to media with increasing concentrations of natamycin, none of the strains had become less sensitive. When the sensitivity of yeasts and moulds isolated from dry-sausage factories where natamycin had been used for several years was compared with that of isolates from factories where natamycin had never been applied, no significant differences were demonstrated.

Induction of polyene-resistant and especially natamycin-resistant mutants in vitro is difficult. Such mutants invariably show reduced metabolic and growth rates and, in the absence of polyenes, readily revert to normal metabolism, growth, and sensitivity to natamycin. One means of obtaining resistant isolates is successive sub-culturing in vitro in the presence of gradually increasing concentrations of the polyene. Typically, such isolates are resistant only up to the highest concentration to which they have been exposed. After 25 passages, the microbiological inhibitory concentration of Candida albicans was minimally increased from 2.5–12 to 12–50 µg/ml.

Assessment of intake

Application of the budget method indicated that further assessment of the intake of natamycin was required. The draft GSFA proposes restricted use of natamycin in cheese (category 1.6) and dried, non-heat-treated meat groups (categories 8.2.1.2 and 8.3.1.2) only, so that the intake would not be expected to exceed the ADI.

Submissions from Australia, Germany, New Zealand, the United Kingdom, and the USA indicated that the intakes at the mean and high levels of consumption were well below the ADI, although the estimates for the United Kingdom and the USA covered cheese consumption only. The estimated mean intakes for consumers ranged from 0.01 to 0.03 mg/kg bw per day (representing 3 and 9% of the ADI in Germany and the United Kingdom, respectively), and those for high consumers were 0.03–0.08 mg/kg bw per day (representing 9 and 27% of the ADI in Australia and the United Kingdom, respectively), if it is assumed that natamycin was used at 40 mg/kg in all cheese products and 20 mg/kg in all cured meat products, as proposed in the draft GSFA. The estimated intakes of natamycin were lower when national use levels were assumed.

Although use of natamycin as an antifungal agent in food may result in exposure of the endogenous flora to trace quantities of antimicrobial residues, bacteria in the human gastrointestinal tract are not affected by polyenes, and the Committee concluded that disruption of the colonization barrier is not a concern. Fungi are found in much smaller amounts than bacteria in the human gastrointestinal tract, and the negative results in studies of acquired resistance indicate that selection of natamycin-resistant fungi is not an issue.

The Committee noted the finding of extra sternebrae in the study of developmental toxicity in rabbits, in which a dose-related increase in the mortality rate was reported. It considered that administration of an antimicrobial agent to rabbits by gavage was inappropriate. In addition, extra sternebrae have been described as a skeletal variation rather than a frank indication of teratogenicity. Thus, the Committee did not consider this result to be evidence that natamycin is teratogenic.

The Committee confirmed the previously established ADI of 0–0.3 mg/kg bw for natamycin, which was based on observations of gastrointestinal effects in humans. The Committee noted that the estimated intake of natamycin based on maximum levels of use in cheese and processed meats proposed in the draft GSFA does not exceed this ADI.\

Anonymous (1965) Data on the safety of the use of pimaricin as preservative against mold growth on cheese. Summary of the results of acute and chronic toxicity tests. Unpublished report submitted to WHO by the Royal Netherlands Fermentation Industries Ltd, Delft.

Anonymous (1968) Absorption of pimaricin following oral administration. Unpublished report submitted to WHO by the Royal Netherlands Fermentation Industries Ltd, Delft.

Aparicio, J.F., Colina, A.J., Ceballos, E. & Martin, J.F. (1999) The biosynthetic gene cluster for the 26-membered ring polyene macrolide pimaricin. J. Biochem. Mol. Biol., 274, 10133–10139.

Athar, M.A. & Winner, H. I. (1971) The development of resistance by Candida species to polyene antibiotics in vitro. J. Med. Microbiol., 4, 505–517.

Australia New Zealand Food Authority (2001a) Submission on estimated pimaricin intakes to WHO on behalf of the Australian Government.

Australia New Zealand Food Authority (2001b) Submission on estimated pimaricin intakes to WHO on behalf of the New Zealand Government.

Bailey, D.E. & Morgareidge, K. (1974) Teratogenicity test with pimaricin. Unpublished report No. 1-1052 submitted to WHO by Food and Drug Research Laboratories Inc.

Bernhardt, H. (1998) Fungi in the intestine—Normal flora or pathogens? Z. Arztliche Fortbild. Qual., 92, 154–156.

Blankwater, Y.J. & Hespe, W. (1979) Autoradiographic and bioautographic study of the distribution of oral natamycin in the rat. Unpublished report No. 20.502, dated 8 May 1979 from Gist-Brocades NV, Delft.

Blaschke-Hellmessen, R., Buchmann, H. & Schwarze, R. (1996) Effect of orally administrated polyene antimycotics on the intestinal colonization with yeasts: Possibilities and limitations. Mycoses, 39, 33–39.

de Boer, E. & Stolk-Horsthuis, M. (1977) Sensitivity to natamycin (pimaricin) of fungi isolated in cheese warehouses. J. Food Prot., 40, 533–536.

de Boer, E., Labots H., Stolk-Horsthuis, M. & Visser, J.N. (1979) Sensitivity to natamycin of fungi in factories producing dry sausage, Fleischwirtschaft, 59, 1868–1869.

Bodenhoff, J. (1968) [Resistance of Candida albicans, with besonderer Berücksichtigung von zwei während einer längeren Zeit mittels antibiotica behandelten Patienten.] Scand. Dent. J., 76, 279 (in German).

Brik, H. (1975) Natamycin (pimaricin). New high-molecular decomposition products with intact lactone-ring. Unpublished report submitted to WHO by Gist-Brocades NV, Delft.

Brik, H. (1981) Natamycin. Anal. Profiles Drug Substances, 10, 513–561.

Carlile, M.J. & Watkinson, S.C., eds (1994) Fungal cells and vegetative growth. In: The Fungi, San Diego, CA: Academic Press, p. 148.

Cerniglia, C.E. & Kotarski, S. 1999. Evaluation of veterinary drug residues in food for their potential to affect human intestinal microflora. J. Regul. Toxicol. Pharmacol., 29, 238–261.

Cox, G.E., Bailey, D.E. & Morgareidge, K. (1973) Unpublished report No. 1-1052 submitted to WHO by Food and Drug Research Laboratories Inc.

Deacon, J.W., ed. (1997) Prevention and control of fungal growth. In: Modern Mycology, 3rd Ed., Oxford: Blackwell Science, pp. 289–290.

Dekker, J. & Gielink, A.J. (1979) Acquired resistance to pimaricin in Cladosporium cucumerinum and Fusarium oxysporum f.sp. narcissi associated with decreased virulence. Neth. J. Plant Pathol., 85, 67–73.

DSM Food Specialities (2001) Submission on pimaricin to WHO from DSM Food Specialities USA, Inc., King of Prussia, Pennsylvania, USA.

van Eeken, C.J. & Wubs, W. (1976) Acute intraperitoneal toxicity of natamycin and three potential metabolites. Unpublished report No. 15.465, 11 January 1976, submitted to WHO by Gist-Brocades.

van Eeken, C.J., Birtwhistle R.D.R. & Aboulwafa-wan Velthoven, M.J.E. (1984) Three months study in dogs of the toxicity of natamycin by addition to the food. Unpublished report No. 12.401, 24 October 1984, submitted to WHO by Gist-Brocades Research and Development.

Euchi, H., Kamada, Y., Kanagawa, T., Naito, T. & Shiota, H. (1999) Cases of keratomycosis due to Fusarium species. Rinsho Ganka (Jpn. J. Clin. Ophthamol.), 53, 609–611.

Farkas, J. & Kiss, I. (1976) The effect of pimaricin on the microbial flora of cottage cheese in plastic packaging. Acta Aliment., 5, 107–118.

Food and Drug Administration (2001) Food additives permitted for direct addition to food for human consumption: Natamycin (pimaricin). Final rule. Fed. Reg., 66, 21 CFR Part 172.

Franklin, T.J. & Snow, G.A., eds (1998) Antiseptics, antibiotics and the cell membrane. In: Biochemistry and Molecular Biology of Antimicrobial Drug Action, 5th Ed. Dordrecht: Kluwer Academic Publishers, pp. 55–56.

Gibney, M.J. (1999) Dietary intake methods for estimating food additive intake. Regul. Toxicol. Pharmacol., 30, S31–S33.

Griffin, D.A., ed. (1994) Fungicides. In: Fungal Physiology, 2nd Ed., New York: Wiley-Liss, Inc., pp. 416–417.

Grupper, C. (1961) Personal communication from the Hôpital Saint-Louis, Paris.

Grupper, C. (1964) Pimaricin in the treatment of superficial mucocutaneous monoliasis. In: Proceedings of the International Congress on Tropical Dermatoses, Naples, June 1964.

Hamilton-Miller, J.M.T. (1974) Fungal sterols and the mode of action of the polyene antibiotics. In: Perlman, D., ed., Advances in Applied Microbiology, New York: Academic Press, pp. 109–134.

Hejzlar, M. & Vymola, F. (1970) Comparative study of pimaricin and fungicidin activity in vitro. J. Hyg. Epidemiol. (Praha), 14, 211–213.

Hespe, W. & Meier, A.M. (1980) Studies involving dogs in regard to the resorption of radioactivity following the oral administration of ¹⁴C-pimaricin, applied on cheese, in comparison to other oral forms of administration. Unpublished report No. 20.531, dated 4 February 1980, submitted to WHO by Gist-Brocades NV, Haarlem.

Hoekstra, E.S. & Van der Horst, M.I. (1998) Survey of the fungal flora in Dutch cheese factories and warehouses. J. Food Mycol., 1, 13–22.

Holly, R.A. (1981) Prevention of surface mold growth on Italian dry sausage by natamycin and potassium sorbate. Appl. Microbiol., 41, 422–429.

Holly, R.A. (1986) Effect of sorbate and pimaricin on surface mold and ripening of Italian dry salami. Lebensm-Wiss. Technol., 19, 59–65.

Hutchison, E.B., Ribelin, W.E. & Levinskas, G.J. (1966) 98-day study in the rat. Unpublished report submitted to WHO by American Cyanamid Co.

Jay, J.M. ed. (1996) Modern Food Microbiology, 5th Ed., New York: Chapman & Hall, pp. 293–294.

Khoudokormoff, B. (1977) Short term microbial tests on mutagenicity of pimaricin (natamycin) and its products of degradation. Unpublished preliminary results, archive No. 10.545, 29 June 1977, submitted to WHO by Gist-Brocades NV.

Khoudokormoff, B. (1978) Potential carcinogenicity of some food preservatives in the presence of traces of nitrite. Mutat. Res., 53, 208–209.

Khoudokormoff, B. (1984) Are resistance development and morphological changes possible after use of natamycin? Wein-Wiss., 39, 45–50.

Khoudokormoff, B. & Petru, M. (1974) On the possible development of antibiotic resistance amongst fungi with special reference to the use of pimaricin as a preservative in the food industry. Unpublished report submitted to WHO by Gist-Brocades NV, Delft.

Knickerbocker, M. & Re, T.A. (1978) Teratologic evaluation of pimaricin in Dutch belted rabbits. Unpublished report from Food and Research Laboratories, Inc., Waverly Research Center, NY). FDRL Report No. 5906, 22 November 1978. Submitted to WHO by Gist-Brocades, Delft.

Knickerbocker, M. & Re, T.A. (1979) Teratologic evaluation of pimaricin in Dutch belted rabbits (amendment). Unpublished report from Food and Research Laboratories, Inc., Waverly Research Center, NY). FDRL Report No. 5906, 26 April 1979. Submitted to WHO by Gist-Brocades, Delft.

Kreisel, W. (1999) Fungi in the intestine. Clinical significance. Schweiz. Rundsch. Med. Prax., 88, 5–10.

Lambe, J., Kearney, J., Leclercq, C., Berardi, D., Zunft, H.F.J., De Henauw, S., De Volder, M., Lamberg-Allardt, C.J.E., Karkkainen, M.U.M., Dunne, A. & Gibney, N.J. (2000) Enhancing the capacity of food consumption surveys of short duration to estimate long term consumer-only intakes by combination with a qualitative food frequency questionnaire. Food Addit. Contamin., 17, 177–187.

Lavingia, B. & Dave, D. (1986) Comparative study of amphotericin-B, pimaricin and gentian violet on ocular fungi. Indian J. Ophthalmol., 34, 73–77.

Levinskas, G.J., Shaffer, C.B., Bushey, C., Kinde, M.L., Stackhouse, D.W. & Vidone, L.B. (1963) Two-year feeding to rats. Unpublished report from the Central Medical Department, American Cyanamid Co. Submitted to WHO by

Levinskas, G.J., Ribelin, W.E. & Shaffer, C.B. (1966) Acute and chronic toxicity of pimaricin. Toxicol. Appl. Pharmacol., 8, 97–109.

Littman, M.L., Pisano, M.A. & Lancaster, R.M. (1958) Induced resistance of Candida species to nystatin and amphoteracin B. In: Antibiotics Ann. 981. New York: Medical Encyclopedia.

Lovgren, T. & Salmela, I. (1978) In vitro sensitivity of Trichomonas vaginalis and Candida albicans to chemotherapeutic agents. Acta Pathol. Microbiol. Scand. Sect. B, 80, 155–158.

Lüock, H. & Cheesman, C.E. (1978) Mould growth on cheese as influenced by pimaricin or sorbate treatments. S. Afr. J. Dairy Technol., 10, 143–146.

Macura, A.B. (1991) Fungal resistance to antimycotic drugs. A growing problem. Int. J Dermatol., 30, 131–183.

Majewski, S. & Macua, A.B. (1978) Efficacy of topical antifungal treatment in denture stomatitis. Mykosen, 21, 403–406.

Malten, K.E. (1967) Report of an investigation concerning possible allergic side effects of pimaricin in humans. Unpublished report from the Instituut voor Geneeskunde en Maatschappij, Nijmegen. Submitted to WHO by Gist-Brocades NV, Delft.

Malten, K.E. (1968) Report on investigation into possible sensitising side effects of pimaricin in human beings. Unpublished report from the Instituut voor Geneeskunde en Maatschappij, Nijmegen. Submitted to WHO by Gist-Brocades NV, Delft.

Manson, J.M. & Kang, Y.J. (1994) In: Hayes, J.W., ed., Principles and Methods of Toxicology, 3rd Ed., New York: Raven Press, p. 1003.

Masterton, G., Sengupta, S.M. & Schofield, C.B.S. (1975) Natamycin in genital candidosis in men. Br. J. Vener. Dis., 51, 210–212.

McGinnis, M.R. & Rinaldi, M.G. (1985) Antifungal drugs: Mechanisms of action, drug resistance, susceptibility testing, and assays of activity in biological fluids. In: Lorian, V., ed., Antibiotics in Laboratory Medicine, 2nd Ed., Baltimore: Williams & Wilkins, pp. 223–281.

Meier A.M. & Hespe, W. (1979). The metabolism of pimaricin in rats. II. Investigation, with the help of ¹⁴C-pimaricin, of its resorption, its decomposition in the gastrointestinal tract, and its elimination. Unpublished report 20.504, 28 May 1979.

Mendes, M.V., Aparicio, J.F. & Martin, J.F. (1999) Complete nucleotide sequence and characterization of pSNA1 from pimaricin-producing Streptomyces natalensis that replicated by a rolling circle mechanism. Plasmid, 43, 159–165.

Molzahn, S.W. & Woods, R.A. (1972) Polyene resistance and the isolation of sterol mutants in Saccharomyces cerevisiae. J. Gen. Microbiol., 72, 339–348.

Morgenstern, A.P. & Muskens, G.J.A.M. (1975) Further data on the toxicity of the decomposition products of pimaricin. Unpublished report submitted to WHO by Gist-Brocades NV, Delft

Morris, H.A. & Castberg, H.B. (1980) Control of surface growth on blue cheese using pimaricin. Cult. Dairy Prod. J., 15, 21–23.

Mukhtar, H., Hakkou, A. & Bonaly, R. (1994) Studies on the activity of Kluyveromyces lactis S-adenosylmethionine:  24-sterol methyltransferase in presence of polyenic antifungal agents. Mycopathologia, 126, 75–83.

Newcomer, V.D., Sternberg, T.H., Wright, E.T., Reisner, R.M., McNall, E.G. & Sorensin, L.J. (1960) The treatment of systemic diseases with orally administered pimaricin: Preliminary report. Ann. N.Y. Acad Sci., 89, 240–246.

Nilson, K.M., Shahani, K.M., Vakil, J.R. & Kilara, A. (1974) Pimaricin and mycostatin for retarding cottage cheese spoilage. J. Dairy Sci., 58, 668–671.

Norman, A.W., Spielvogel, A.M. & Wong, R.G. (1976) Polyene antibiotic-sterol interaction. Adv. Lipid Res., 14, 127–170.

Novak, E.K., Barbarics, E., Vincze, I. & Zala, J. (1984) In vitro studies on the food preservative antifungal polyene antibiotic pimaricin (natamycin). In: Kiss, I., Deck, T. & Incaze, K., eds, Microbial Association and Interactions in Food, Dordrecht: Kluwer..

O’Day, D.M., Ray, W.A., Robinson, R.D., Head, W.S. & Savage, A.M. (1987) In vitro and in vivo susceptibility of Candida keratitis to topical polyenes. Invest. Ophthalmol. Visual Sci., 28, 874–880.

Oldenkamp, E.P. (1979) Natamycin treatment of ringworm in cattle in the United Kingdom. Vet. Rec., 105, 554–556.

Oldenkamp, E.P. & Spanoghe, L. (1977) Natamycin-S treatment of ringworm in cattle. Tijdschr. Diergeneeskd., 102, 124–125.

Ottens, H. (1965) Unpublished report submitted to WHO by Royal Netherlands Fermentation Industries, Delft.

Pedersen, J.C. (1992) Natamycin as a fungicide in agar media. Appl. Environ. Microbiol., 58, 1064–1066.

Pugazhenthi, T.R, Dhanalakshmi, B., Narasimhan, R., Shibu, A.V. & Madhan, S. (1999) Effect of anti-mycotic agents on Penicillium citrinum in cheese. Indian Vet. J., 76, 537–539.

Raab, W. P. (1972) Natamycin (Pimaricin). Its Properties and Possibilities in Medicine, Stuttgart: Georg Thieme Publishers.

Reynolds, J.E.F., ed. (1996) Martindale. The Extra Pharmacopoeia, 31st Ed., London: Pharmaceutical Press.

Rusul, G. & Marth, E.H. (1988) Growth and aflatoxin production by Aspergillus parasiticus in a medium at different pH values and with or without pimaricin. Z. Lebensm. Unters. Forsch., 187, 436–439.

Sacjdeva, S. & Singh, S. (1985) Pimaricin—A potential preservative in the cheese industry. Indian Dairyman, 37, 587–596.

Safe, L.M., Safe, S.H., Subden, R.E. & Morris, D.C. (1977) Sterol content and polyene antibiotic resistance in isolates of Candida krusei, Candida parakrusei, and Candida tropicalis. Can. J. Microbiol., 23, 398–401.

Scheurlen, M. (1996) Pathogenicity of fungi in the intestines—Current status of the discussion. Fortschr. Med., 114, 319–321.

Shahani, K.M., Bullerman, L.B., Evans, T.A. & Arnold, R.G. (1977) Prevention of toxin mold growth in cheese by pimaricin. Arch. Inst. Pasteur Tunis, 43, 511–520.

Sörensen, L.J., McNall, E.G. & Sternberg, T.H. (1959) The development of strains of Candida albicans and Coccidioides immitis which are resistant to amphotericin B. In: Antibiotics Annual 1958–1959, New York: Medical Encyclopedia, pp. 920–923.

Stern, G.A. (1978) In vitro antibiotic synergism against ocular fungal isolates. Am. J. Ophthalmol., 86, 359–367.

Stout, H.A. & Pagano, J.F. (1956) Resistance studies with nystatin. In: Antibiotics Ann.: 704. New York

Struyk, A.P., Hoette, I., Drost, G., Waisvisz, J.M., van Eek, T. & Hoogerheide, J.C. (1958) Pimaricin, a new antifungal antibiotic. In: Welch, H. & Marti-Ibanez, F., eds, Antibiotics Annual 1957–1958, New York: Medical Encylopedia, Inc., pp. 878–885.

Subden, R.E., Safe, L., Morris, D.C., Brown, R.G., and Safe S. (1977) Eburicol, lichesterol, ergosterol, and obtusifoliol from polyene antibiotic-resistant mutants of Candida albicans. Can. J. Microbiol., 23, 751–754.

Walter, A.M. & Heilmeyer, L. (1969) Antibiotika Fibel. Stuttgart: Thieme Verlag.

Wieriks, J. (1966) Pimaricin in cheese: A toxicity test of seven weeks in rats. Unpublished report from the Royal Netherlands Fermentation Industries Ltd, submitted to WHO.

Wieriks, J. (1971) Pimaricin in apples: A toxicity test of three months in rats. Unpublished report from the Royal Netherlands Fermentation Industries Ltd, submitted to WHO.

    See Also:
       Toxicological Abbreviations