First draft prepared by
    Dr B. Priestly
    University of Adelaide


         Furfural has not been previously evaluated by the Committee.
    Furfural occurs naturally and is formed during the processing and
    domestic preparation of a broad range of foods. It is also carried
    over into food from its use as an extraction solvent or as a
    component of flavour mixtures. Food additive or processing use
    provides an intake estimated to be no more than 0.5-1% of the intake
    from other food sources.


    2.1  Biochemical aspects

    2.1.1  Absorption, distribution, and excretion

         Furfural is well absorbed by all routes of exposure. In rats,
    85% of 14C-furfural administered by gavage in corn oil was
    recovered in urine within 72 h (NTP, 1987). In humans, approximately
    78% was retained by the lungs during 8 h inhalation of vapour
    (7-30 mg/m3), and 20-30% of this amount was absorbed through the
    skin of a hand immersed in the liquid (Flek & Sedivec, 1978). The
    biological half-life in man was estimated to be 2-2.5 h.

    2.1.2  Biotransformation

         The major route of metabolism is oxidation to furoic acid,
    followed by conjugation with glycine (NTP 1987, Flek & Sedivec,
    1978). A minor pathway involves condensation with acetate, followed
    by conjugation with glycine. The resultant furfurylacryluric acid
    has been detected in human urine after exposure to furfural (Flek &
    Sedivec, 1978).

    2.1.3  Effects on enzymes and other biochemical

         No data available.

    2.2  Toxicological studies

    2.2.1  Acute toxicity

         Furfural can irritate exposed mucosa and cause degenerative
    effects on epithelial cells. Acute toxic doses cause CNS depression,
    lung congestion and haemorrhage, and eye/nasal discharge. Delayed
    toxic effects and those associated with repeat dosage include
    hepatic and renal tubular necrosis, hypochromic anaemia and
    leukopenia (Castellino  et al., 1963; Jenner  et al., 1964).
    Markedly increased hepatic mitogenesis (without necrosis) was noted
    within 6 h after a single oral dose of 50 mg/kg furfural in rats
    (Shimizu & Kanisawa, 1986).

         The acute lethal dose is comparable across several species and
    by different routes of exposure.

    Table 1. Acute toxicity of furfural

    Species        Route           Lethal dose                Reference
                                   mg/kg bw

    Rat            oral            127                        Jenner et al., 1964

    Rat            i.p.            120                        Tiunov et al., 1970

    Mouse          oral            333                        Boyland, 1940

    Mouse          s.c.            200                        Tiunov et al., 1970

    Mouse          s.c.            223 (1 day survival)       Castellino et al.,
                                   119 (10 day survival)      1963

    Rabbit         dermal          > 310                      Moreno, 1976

    Hamster        inhalation      2500 mg/m3                 Kruysse, 1972

    Avian          oral            > 98                       Schafer et al., 1983
    2.2.2  Short-term studies  Mice

         In range-finding dosage studies for the NTP carcinogenesis
    bioassay, B6C3F1 mice were dosed by gavage with 25, 50, 100,
    200, or 400 mg/kg bw/day for 16 days, and 75, 150, 300, 600 or
    1200 mg/kg bw/day for 13 weeks. Relative liver weights were
    increased in males at 300 and in females at 75, 150 and
    300 mg/kg/day (NTP, 1990). The following table shows incidence of
    centrilobular coagulative necrosis observed:

    Table 2. Incidence of centrilobular coagulative necrosis

    Dose level                    0      75     150     300     600      1200
    mg/kg bw/day)

    males                         0       0      1       1       9        8
    females                       0       0      0       0       0        2

    (n=10 for all groups)
         Mortality at the top two dose rates was 100 per cent for both
    males and females. Most deaths occurred during the first week of
    dosing.  Rats

         A group of 48 male Wistar rats were fed a dietary regime of
    furfural (20 ml/kg, equivalent to approximately 23 000 ppm or
    1150 mg/kg bw/day) for 7 days; 30 ml/kg for 7 more days; 40 ml/kg
    for a further 76 days; then 40 ml/kg for 5 days/week for a further
    30 days. Two rats died, four rats were killed at 90 days, one at 12
    days, and all were found to have developed a hepatic cirrhosis,
    characterized by red nodules interspersed with white fibrous
    bundles. The white areas displayed marked proliferation of atypical
    cholangioepithelial cells and the lesion was described by the
    authors as "cholangiofibrosis". There were increased numbers of
    cells undergoing mitosis; none of the rats developed ascites or
    hyperbilirubinaemia (Shimizu & Kanisawa, 1986).

         In range-finding dosage studies for the NTP carcinogenesis
    bioassay, F344/N rats were dosed by gavage with 15, 30, 60, 120, or
    240 mg/kg bw/day for 16 days, and 11, 22, 45, 90 or 180 mg/kg bw/day
    for 13 weeks. Only the top dose rates were significantly lethal.
    Centrilobular hepatocytic vacuoles were found in all treated rats in
    the 13 week study (and in 4/10 controls as well), the absolute and
    relative liver and kidney weights were higher at the 90 mg/kg/day
    dose level, but there was no other compound-related toxicity (NTP,
    1990).  Hamsters

         Groups of 10 Syrian golden hamsters of each sex were exposed to
    furfural vapour (0, 77, 448, 2165 mg/m3) for 6 h/day, 5 days/week
    over 13 weeks. The main findings were mild growth retardation,
    irritation of the eyes and nose, and hyperplastic atrophy of the

    nasal epithelium (all at the highest dose). The NEL was 77 mg/m3,
    since some mild nasal epithelial degeneration was observed at 448
    mg/m3 (Feron  et al., 1979)

    2.2.3  Long-term/carcinogenicity studies  Mice

         Groups of 50 B6C3F1 mice of each sex were dosed by gavage
    with 0, 50, 100 or 175 mg/kg bw/day furfural (99% pure) for 2 years.
    Growth and survival were not affected by the treatment. The main
    non-neoplastic findings related to treatment were an increased
    incidence of hepatic multifocal pigmentation and subserosal
    inflammation in both sexes (m>f) (NTP, 1990).

         The principal neoplastic effects are summarized in Table 3.

        Table 3. Neoplastic effects observed in mice

                             Dose group (n=50 except for male 100 where n=49)

                                    Males                          Females

                              0    50    100   175           0     50     100   175

    Hepatocellular            9    13     11    19           1      3       5     8

    Hepatocellular            7    12      6    21           4      0       2     4

    Adenoma/                 16    22     17    32           5      3       7    12

    Renal cortical            0     1      1     1

    Forestomach                                              0     5        5     3

    Squamous cell                                            1     0        1     6
         The NTP conclusions (confirmed by the Peer Review Panel) were:
    there was  clear evidence of carcinogenicity in male mice, based on
    the increased incidence of hepatocellular adenomas and carcinomas;
    and there was  some evidence of carcinogenicity in female mice
    based upon the increased incidence of hepatocellular adenomas.  Rats

         Groups of 50 F344/N rats of each sex were dosed by gavage with
    0, 30, or 60 mg/kg bw/day furfural (99% pure) for 2 years. Growth
    was not affected by the treatment, but survival was reduced in the
    female high-dose group due to gavage-related deaths. The main
    non-neoplastic findings related to treatment were an increased
    incidence of hepatic centrilobular necrosis in males (3/50, 9/50,
    12/50 in controls, low- and high-dose groups, respectively). Two
    high-dose males had bile duct dysplasia with fibrosis and two others
    had cholangio-carcinomas. The historical incidence of bile duct
    neoplasms was 3/2,145, so that these lesions were considered to be
    treatment-related (NTP, 1990).

         The NTP conclusions (confirmed by the Peer Review Panel) were
    that there was  some evidence of carcinogenicity in male rats,
    based on the rarity of the biliary pathology. There was  no
    evidence of carcinogenicity in female rats. It was noted that poor
    survival in the high-dose female rats may have compromised the
    detection of carcinogenesis in this group.

    2.2.4  Reproduction studies

         No data available.

    2.2.5  Special studies on carcinogenicity

         Citral and heptaldehyde are two aldehydes found to inhibit the
    growth of spontaneous mammary tumours in mice. However, furfural and
    its furfuracrylic acid metabolite (at a daily dose rate of 2.5 mg -
    approximately 25% of the LD50) had no effect on the growth of
    spontaneous mammary tumours in mice, although they were weakly
    active against grafted sarcomata (Boyland, 1940).

         Syrian Golden hamsters were exposed to furfural vapour for
    1 year (1550 mg/m3 weeks 0-9; 1280 mg/m3 weeks 10-20;
    970 mg/m3 weeks 21-52; 7 h/day; 5 days/week). Dosage reduction was
    necessary to avoid substantial toxicity. Some groups were also
    treated with carcinogens (benzo[a]pyrene by weekly intratracheal
    instillation; total dose 18.2 or 36.4 mg; diethylnitrosamine
    injected s.c every three weeks; total dose 2.1 l). All rats were
    observed for a further 29 weeks after cessation of dosing. While
    furfural produced marked nasal irritation, cysts of the propria and

    degeneration of the olfactory epithelium and Bowmans glands, it did
    not induce respiratory tract cancers by itself, nor did it
    potentiate the effects of either of the respiratory carcinogens
    (Feron & Kruysse, 1978).

         Half of the surviving rats from an experiment in which furfural
    was used to produce hepatic cirrhosis were fed a diet containing
    0.03% N-2-fluorenylacetamide for three weeks, followed by one week
    of normal diet. The other received only the normal diet. All
    surviving rats were killed 12 weeks later and their livers were
    examined. Rats which had received furfural only did not have any
    hyperplastic changes in the liver. Rats which had been treated with
    N-2-fluorenylacetamide developed multiple hyperplastic nodules that
    stained positive for alpha-fetoprotein, and this response was
    markedly potentiated in the rats previously treated with furfural.
    The authors concluded that furfural-induced hepatic cirrhosis
    increases susceptibility to a potent hepatocarcinogen in the rat
    (Shimizu & Kanisawa, 1986).

         Samples of mouse liver tumours from the NTP Carcinogenesis
    Bioassay had their transforming gene activity assessed by Southern
    blot analysis of H- ras oncogene sequences. It was found that the
    pattern of mutations in the  ras gene differed in spontaneous
    tumours compared to furfural-induced tumours. While all of the
    activated oncogenes in tumours from the control mice were H-ras,
    with point mutations at codon 61, 50% of the oncogenes activated in
    furfural-treated mouse liver tumours were K-ras or H-ras genes
    bearing point mutations other than at codon 61. The authors
    concluded that furfural has a direct genotoxic effect in mouse liver
    (NTP 1990).

    2.2.6  Special studies on genotoxicity

         Furfural is generally negative in bacterial mutagenicity
    studies, but is positive in  Drosophila and in some mammalian cells
     in vitro. The results are summarised in Table 4.

         Few  in vivo genotoxicity studies have been reported for
    furfural. Furfural did not induce SCEs or chromosomal aberrations in
    bone marrow cells of B6C3F1 mice injected i.p. with furfural
    at 50, 100 or 200 mg/kg bw (NTP 1990).

        Table 4. Genotoxicity studies on furfural

    Test system         Test object             Concentration or    Result                 Reference

    Ames test           TA98, 100               up to 8 l          negativea              Zdienicka et
                                                (cytotoxic)                                al., 1978

    Ames test           TA98, 1535              0.05-60 mol        negativea              Loquet et
                                                                                           al., 1981

    Ames test           TA104                   1 mole             negative               Marnett et
                                                                                           al., 1985

    Ames test           TA97, 98, 199,                              equivocal positive     Mortelmans et
                        1535, 1537                                  in TA100 in 1 of 2     al., 1986
                                                                    test centres

    Ames test           TA100, 102, 104         up to 500 g        negative               Shane et al.,

    Ames test           TA98, 100, 104                              negative               Kato et al.,
                        E. coli reversion                                                  1984

    Umu gene            S. typhimurium          1932 g/ml          negative               Nakamura et al.,
    expression          TA1535/pSK 1002                                                    1987

    Rec-assay           B. subtilis             up to 1 mg          negative               Osawa & Namiki,
    Ames test           TA100                                                              1982

    Clastogenicity      CHO cells               up to 40 mM         positiveb              Stich et al.,


    Table 4. cont'd

    Test system         Test object             Concentration or    Result                 Reference

    SCE                 human lymphocytes       up to 140 M        positive               Gomez-Arroyo &
                                                                                           Souza, 1985

    Gene conversion     S. cerevisiae D7        16 mg/ml            positive               Stich et al.,

    Mouse               lymphoma                200 g/ml           positive               McGregor et al.,
                        L5178Ytk+/tk                                                       1988

    Forward mutation

    DNA strand break    calf thymus DNA                             positive               Hadi et al.,
    and alkaline                                                                           1989

    Sex-linked          Drosophila                                  positive               Woodruff et al.,
    recessive           melanolgaster                               (recessive)            1985

    Lethal mutation                                                 lethals only
    and reciprocal
    a)   results in two studies with TA100 were positive (Zdzienicka  et al., 1978;
         Loquet et al, 1981), both in the presence and absence of metabolic
         activation. This strain is supposedly more sensitive to aldehydes. However,
         negative results have also been recorded in TA100, and in T104, a strain
         even more sensitive to aldhyde mutagens (Marnett  et al., 1985)

    b)   positive without metabolic activation. In contrast, the parent alcohol,
         furfuryl alcohol, is more potent as a clastogen and its activity was
         enhanced by S9 activation.

    2.3  Observations in humans

         The frequency of SCE in lymphocytes cultured from 6 Mexican
    factory workers purportedly exposed occupationally to furfural and
    furfuryl alcohol was not different from that in 6 controls
    (Gomez-Arroyo & Souza, 1985).


         The Committee considered data on the absorption, metabolism,
    acute and chronic toxicity, genotoxicity, and carcinogenicity of

         Furfural is absorbed by all routes of exposure. It is
    metabolized by oxidation to furoic acid and by subsequent
    conjugation with glycine. The half-life in humans is 2-2.5 h.

         The liver is the primary target for furfural toxicity in rats
    and mice. In short-term studies, it produced liver enlargement at
    doses > 90 mg/kg bw/day in rats and at 75-300 mg/kg bw/day in
    mice. At higher doses, it produced centrilobular necrosis and

         In a 2-year gavage study, using dose levels of 0, 50, 100, or
    175 mg/kg bw/day in B6C3F1 mice and 0, 30 or 60 mg/kg bw/day
    in F344/N rats, furfural induced a statistically significant
    increase in the incidence of hepatocellular adenomas and carcinomas
    in male mice (34-64% compared with 32% in controls), hepatocellular
    adenomas in female mice (6-16% compared with 2% in controls), and
    bile-duct dysplasias (4%) plus cholangiocarcinomas (4%) in male rats
    at the highest dose level (compared with zero incidence in controls
    and a historical control incidence of 3 in 2145 (0.14%) for
    cholangiocarcinomas). While furfural was generally negative in
    bacterial mutagenicity tests, it was positive in a range of other
    tests for genotoxicity. The pattern of oncogene expression in liver
    tumours in furfural-treated mice differed from that in the
    spontaneous liver tumours of the controls.


         While taking into consideration the relatively high
    concentrations of furfural in some foods as normally prepared and
    consumed, the Committee considered that it could not allocate an ADI
    to furfural because of the evidence of genotoxicity and
    carcinogenicity. The Committee considered that its direct addition
    as a flavour was not appropriate, and that its use as a solvent
    should be restricted to situations when alternatives were not
    available, e.g., for the purification of food oil extraction of
    unsaturated components. Carry-over into food should be reduced to
    the lowest extent technologically feasible.


    BOYLAND, E. (1940). Experiments on the chemotherapy of cancer. 4.
    Further experiments with aldehydes and their derivatives.
     Biochem. J., 34: 1196-1201.

    CASTELLINO, N, ELMINO, O. & ROZERA, G. (1963). Experimental research
    on the toxicity of furfural.  Arch. Environ. Health, 7: 574-582.

    FERON, V.J. & KRUYSSE, A. (1978). Effects of exposure to furfural
    vapour in hamsters simultaneously treated with benzo[a]pyrene or
    diethylnitrosamine.  Toxicology, 11: 127-144.

    Repeated exposure to furfural vapour: 13-week study in Syrian Golden
    hamsters.  Zbl. Bakt. Hyg. I Abt. Orig. B., 168: 442-451.

    FLEK, J & SEDIVEC, V. (1978). The absorption, metabolism and
    excretion of furfural in man.  Int. Arch. Occupat. Environ. Health,
    41: 159-168.

    GOMEZ-ARROYO, S. & SOUZA, V. (1985).  In vitro and occupational
    induction of sister-chromatid exchanges in human lymphocytes with
    furfuryl alcohol and furfural.  Mutation Res., 156: 233-238.

    HADI, S.M., SHAHABUDDIN & REHMAN, A. (1989). Specificity of the
    interaction of furfural with DNA.  Mutation Res., 225: 101-106.

    (1964). Food flavourings and compounds of related structure. I.
    Acute oral toxicity.  Fd. Cosmet. Toxicol., 2: 327-343.

    KATO, F., ARAKI, A., NOZAKI, K. & MATSUSHIMA, T. (1984).
    Mutagenicity of aldehydes and diketones.  Mutation Res., 216:

    KRUYSSE, A. (1972). Acute inhalation toxicity of furfural in
    hamsters. Central Institute for Nutrition and Food Research TNO,
    Report R 3853.  cited in: Feron & Krysse (1978).

    LOQUET, C., TOUSSAINT, G. & LeTALAER, J.Y. (1981). Studies on
    mutagenic constituents of apple brandy and various alcoholic
    beverages collected in Western France, a high incidence area for
    oesophageal cancer.  Mutation Res., 88: 155-164.

    MAARSE, H & VISSCHER, C.A. (1989). Volatile compounds in food.
    Qualitative and quantitative data. Vol. III. TNO-CIVO Food Analysis
    Institute, Zeist, The Netherlands.

    H. & AMES, B.N. (1985). Naturally occurring carbonyl compounds are
    mutagens in  Salmonella tester strain TA104.  Mutation Res., 148:

    CASPARY, W.J. (1988). Responses of the L5178Y tk+/tk- mouse
    lymphoma forward mutation assay. II: 18 coded chemicals.  Environ.
     Molec. Mutagenesis, 11: 91-118.

    MORENO, O. (1976). Dermal toxicity in rabbits. Report No. MB 75-941
    from MB Research Laboratories Inc. Submitted to WHO by the Flavor &
    Extract Manufactures Association of the United States.

    ZEIGER, E. (1986).  Salmonella mutagenicity tests: II. Results from
    the testing of 270 chemicals.  Environ. Mutagenesis, 8, Suppl. 7.

    NAKAMURA, S., ODA, Y., SHIMADA, T., OKI, I. & SUGIMOTOT, K. (1987).
    SOS-inducing activity of chemical carcinogens and mutagens in
     Salmonella typhimurium TA1535/pSK 1002: Examination with 151
    chemicals.  Mutation Res., 192: 239-246.

    NAS (1989). 1987 poundage and technical effects update of substances
    added to food. Report of the Committee on Food Additives Survey
    Data, Food & Nutrition Board, Institute of Medicine, National
    Academy of Sciences.

    NTP (1987). National Toxicology Program. Studies of chemical
    disposition in mammals: The effect of dose on the chemical
    disposition of furfural in rats after oral administration. NIEHS
    Contract No. N01-ES-66138. Arthur D. Little, Inc.

    NTP (1990). National Toxicology Program. Toxicology and
    carcinogenesis studies of furfural in F344/N rats and B6C3F1
    mice (gavage studies). NTP Technical report No. 382. US Departmant
    of Health & Human Services, Public Health Service, National
    Institutes of Health, Research Triangle Park, NC.

    OSAWA, T. & NAMIKI, M. (1982). Mutagen formation in the reaction of
    nitrite with the food components analogous to sorbic acid.  Agric.
     Biol. Chem., 46: 2299-2304.

    AARONSON, S.A. & ANDERSON, M.W. (1987). Activated oncogenes in
    B6C3F1 mouse liver tumors: Implications for risk assessment.
     Science, 237: 1309-1316.

    SCHAFER, E.W., BOWLES, W.A. & HURLBUT, J. (1983). The acute oral
    toxicity, repellency, and hazard potential of 998 chemicals to one
    or more species of wild and domestic birds.  Arch. Environ. Contam.
     Toxicol., 12: 355-382.

    SHANE, B.S., TROXCLAIR, A.M., McMILLIN, D.J. & HENRY, C.B. (1988).
    Comparative mutagenicity of nine brands of coffee to  Salmonella
     typhymurium TA100, TA102, and TA104.  Environ. Molec. Mutagenesis,
    11: 195-206.

    SHIMIZU, A. (1986). Experimental studies on hepatic cirrhosis and
    hepatocarcinogenesis. II. Influence of cirrhotic liver on 2-FAA
    hepatocarcinogenesis in rats.  Acta Pathol. Jpn., 36: 1039-1048.

    SHIMIZU, A & KANISAWA, M. (1986). Experimental studies on hepatic
    cirrhosis and hepatocarcinogenesis. I. Production of hepatic
    cirrhosis by furfural administration.  Acta Pathol. Jpn., 36:

    STICH, H.F., ROSIN, M.P., WU, C.H. & POWRIE, W.D. (1981a).
    Clastogenicity of furans found in food.  Cancer Letters, 13: 89-95.

    STICH, H.F., ROSIN, M.P., SAN, R.H.C., WU, C.H. & POWRIE, W.D.
    (1981b). Intake, formation, and release of mutagens by man. Banbury
    Report 7. Gastrointestinal cancer. 247-266.

    TIUNOV, L.A., VASIL'EV, G.A. & IVANOVA, L.V. (1970). Effect of
    frequency of injection on the toxic action of furfural.  Bull. expt.
     Biol. Med., 70: 999-1000.

    Chemical mutagenesis testing in  Drosophila. V. results of 53 coded
    compounds tested for the National Toxicology Program.  Environ.
     Mutagenesis, 7: 677-702.

    Mutagenic activity of furfural in  Salmonella typhimurium TA100.
     Mutation Res. 58: 205-209.

    See Also:
       Toxicological Abbreviations
       Furfural (ICSC)
       Furfural (WHO Food Additives Series 42)
       Furfural (WHO Food Additives Series 46)
       FURFURAL (JECFA Evaluation)
       Furfural (IARC Summary & Evaluation, Volume 63, 1995)