First draft prepared by
    Dr K. B. Ekelman and Dr G. A. Dannan, Additives Evaluation Branch
    Division of Health Effects Evaluation
    Center for Food Safety and Applied Nutrition
    Food and Drug Administration, Washington, DC, USA


         Konjac flour, commonly referred to as konjac mannan, is a
    -D-(1->4)-linked linear copolymer of glucose and mannose
    substituted with  O-acetate every 9-19 sugar units.  Konjac flour
    is derived from the tubers of  Amorphophallus konjac.  The flour,
    which constitutes 60-80% of the dried root tuber, is obtained by a
    dry milling process of thin tuber slices.  Carbohydrates (as water-
    soluble fiber) make up approximately 75% of konjac flour, the
    remainder being protein (2-8%), fat (<1%), ash (3-5%), and moisture

         Konjac flour has not been evaluated previously by the
    Committee.  The current evaluation was undertaken because of
    anticipated new food additive uses of konjac flour as a gelling
    agent, thickener, emulsifier, and stabilizer in such foods as soup,
    gravy, mayonnaise, and jam.  Nevertheless, there is a long history
    of use of konjac (containing approximately 4% konjac flour) in
    traditional Japanese and Chinese foods; the average consumption of
    konjac flour from these uses is estimated to be 2-3 g/person/day,
    and occasionally as high as 4 g/person/day.  The anticipated maximum
    consumption of konjac flour from food additive uses is about
    3 g/person/day.

         Numerous physicochemical studies have been done on konjac.  The
    primary polysaccharide constituent in konjac flour is a high-
    molecular-weight glucomannan (200 000 to 2 000 000 daltons depending
    on the strain and place of cultivation) in which  D-glucose and
     D-mannose, in a molar ratio of 1.0:1.6, are linked by -1->4
    glycosidic bonds.  Branching from the C3 of either hexose is
    estimated to occur every ten repeating units through 1,3 linkages. 
    Acetyl groups bound every 9-19 units along the glucomannan backbone
    are thought to contribute to the high solubility characteristic of
    konjac flour (Nishinari  et al., 1992).

         Doi and coworkers (1982) reported that the viscosity of konjac
    flour is greater than that of guar gum, one of the most viscous of
    the dietary fibers.  Ebihara and coworkers (1981) reported that,
    compared to dietary fibers such as carboxymethylcellulose (CMC) and
    pectin, the relative viscosity of konjac flour increases
    extraordinarily as its concentration in water increases.  When fiber
    concentration in water is increased from 1 to 3 g/l, the relative

    viscosity of pectin increases from 1.8 to 3.2, the relative
    viscosity of CMC increases from 4.6 to 10.4, and the relative
    viscosity of konjac flour increases from 5.7 to 171.2 (water was
    arbitrarily assigned a viscosity of 1.0).


    2.1  Biochemical aspects

         Due to the -glycosidic linkages between the glucose and
    mannose building blocks (-1->4 linkages in the main chain and
    -1->3 linkages at the branch points) konjac flour is commonly
    regarded as a non-digestible polysaccharide.  Because of its high
    water solubility, conferred mainly by attached acetyl groups, konjac
    flour is also classified as a soluble fiber.  Following alkali
    treatment or heating, konjac flour loses acetyl groups and forms a
    gel.  Gelling is thought to result from cross-linking, mainly
    through hydrogen-bonding between konjac flour moieties that are
    deficient in acetyl groups (Nishinari  et al., 1992).

         Since konjac flour's polymeric structure is assumed to render
    it unavailable for intestinal degradation or absorption, konjac
    flour is commonly believed to pass through the gastrointestinal
    tract unaltered.  Although no studies have been performed for
    evaluating the possible degradation or hydrolysis of konjac flour in
    the intestinal tract, formation of oligo- and mono-saccharides from
    konjac flour has been observed in the presence of 1N sulfuric acid
    at 100C (Japan Food Research Laboratories, 1984).

    2.2  Toxicological studies

    2.2.1  Acute toxicity studies

         Table 1 summarizes the results of acute toxicity studies with konjac flour:

    Table 1. Acute toxicity studies with konjac flour


    Species    Sex     Route         LD50                Reference


    Mouse      M & F   oral1         >2800 mg/kg bw     Oketani  et al., 1984

    Rat        M & F   oral2         >5 000 mg/kg bw3   Kotkoskie  et al., 1992

    Rat        M & F   inhalation2   >0.0015 mg/l       Kotkoskie  et al., 1992

    Rabbit     M & F   dermal2       >2 000 mg/kg bw    Kotkoskie  et al., 1992

    (1)  Test substance was identified as konjac mannan.
    (2)  Test substance was identified as konjac flour.
    (3)  This was the only dose tested.
    2.2.2  Short-term toxicity studies

         In a study conducted to determine appropriate doses for a
    subsequent embryotoxicity study, Burger and coworkers (1992) fed
    diets containing 2% carob gum (controls) or 2% konjac flour (Test
    substance identified as konjac flour from  Amorphophallus
     oncophyllus from Thailand) to two groups of 15 adult female
    British domestic short-hair cats for eight weeks.  Feed consumption
    and body weight were monitored throughout the study.  Although no
    data were presented, report stated that both diets had digestibility
    coefficients similar to typical canned cat foods.  In addition, body
    weights of cats fed diets containing 2% konjac flour increased more
    than body weights of cats fed the control diet during the study. 
    Mean body weight gain in adult female cats during the 8-week period
    was 593 g/wk for controls (2% carob gum) and 865 g/wk for konjac
    flour-fed cats; however, the respective mean feed consumption was
    1908 g/day and 18112 g/day (Burger  et al., 1992).

    2.2.3  Long-term toxicity/carcinogenicity studies

         No information was available.

    2.2.4  Reproduction studies

         No information was available.

    2.2.5  Special studies on anti-carcinogenic effects

         A diet containing refined konjac was evaluated for its effects
    on the incidence of spontaneous liver tumours in C3H/He mice; these
    tumours generally occur in 60-70% of one-year-old mice of this
    strain.  At seven weeks of age, groups of 30 male mice were fed
    either a powdered commercial diet (control group) or the same diet
    to which 10% konjac flour (identified as konjac mannan) had been
    added (konjac flour group).  At one year of age, all animals were
    necropsied and the number and size of liver tumour nodules were
    determined.  There was a slight decrease in the number of animals
    with liver tumours in the konjac flour group (control: 63% of 24
    mice; konjac flour: 48% of 23 mice) and a statistically significant
    decrease (p<0.05) in the mean number of tumour nodules per mouse in
    the konjac flour group (control: 1.1; konjac flour: 0.5).  However,
    mean tumour size was not altered.  Weight gain in the 10% konjac
    flour diet group was lower (p<0.05) than that in the control diet
    group throughout the experiment, but there was no change in total
    feed intake between the control and konjac flour-treated mice. 
    While feed efficiency was decreased in konjac flour-treated mice
    compared to controls (control: 2.9%; konjac flour: 2.3%), the
    decrease was not statistically significant.  In this study,
    spontaneous liver tumours in C3H/He mice were inhibited by
    maintaining the mice on a diet containing 10% konjac flour, although

    the reviewers note that animals maintained on this diet consumed
    approximately 10% fewer calories per day than control animals 
    (Mizutani and Mitsuoka, 1982).

         In another study by the same authors the effect of a diet
    containing 5% konjac flour on the incidence of colon tumours induced
    by 1,2-dimethylhydrazine (DMH) in rats was studied.  Five-week old
    male Fisher 344 rats (20/group) were fed either a commercial diet
    (414 kcal/100 g) or a similar diet containing 5% konjac flour
    (identified as konjac mannan, diet had an energy content of 393
    kcal/100 g).  At six weeks of age, and weekly thereafter for a total
    of 13 weeks, all rats were injected i.p. with 20 mg DMH/kg bw.  Feed
    consumption was measured weekly for 20 weeks (duration of the study
    was approximately 27 weeks).  Rats were necropsied 13 weeks after
    the last injection of DMH; the intestine (small and large) and other
    organs (unspecified) were examined grossly and microscopically for
    numbers and types of tumours.  Throughout the study, body weights of
    konjac flour-fed rats were significantly lower than those of rats
    fed the control diet; however, there was no significant difference
    in feed efficiency between konjac flour-fed and control rats.  The
    incidence of DMH-induced colon tumours was significantly lower in
    the konjac flour-fed group (39%) compared to the control group
    (75%).  The number of colon adenocarcinomas per rat was also
    significantly lower in konjac flour-fed rats (0.22) than in control
    rats (0.75).  However, the mean diameter of colon tumours was not
    significantly different in the two groups of rats (konjac flour-fed
    rats: 5.81.3 mm; control rats: 6.93.6 mm).

         In contrast to the effects reported for colon tumours, dietary
    konjac flour had no significant effect on the incidence of tumours
    of the small intestine, all of which were adenocarcinomas in this
    study (control: 45%; konjac flour: 33%); mean diameters of
    adenocarcinomas of the small intestine were not significantly
    different in the two groups (control: 84 mm; konjac flour: 62 mm). 
    Dietary konjac flour did not appear to have a significant effect on
    the incidences of ear duct or pancreas tumours in rats in this study
    (Mizutani and Mitsuoka, 1983).

    2.2.6.  Special studies on embryotoxicity

         Pregnant British short-hair domestic cats were fed diets
    containing 2% carob gum (9 control cats) or 2% konjac flour (6 cats;
    test substance identified as konjac flour from  Amorphophallus
     oncophyllus from Thailand) during gestation.  Body weights were
    recorded weekly until parturition and feed consumption was recorded
    daily during the week prior to parturition.  Actual intake of konjac
    flour during the week prior to parturition ranged from 0.98 to
    3.08 mg/kg bw/day.  All pregnant females completed a normal
    gestation period and that there were no significant differences in
    body weight changes of females fed control and konjac flour-
    containing diets.  Mean birth weight of kittens born to control cats

    was statistically significantly lower (p<0.01) than mean birth
    weight of kittens born to konjac flour-fed cats, but mean litter
    size for control cats was less than mean litter size for konjac
    flour-fed cats: A total of 32 kittens (mean birth weight 10417 g)
    were born to 9 control cats and 36 kittens (mean birth weight
    95 g22 g) were born to 6 konjac flour-fed cats; mean litter size
    was 3.51.6 for controls and 5.11.2 for konjac flour-fed cats (mean
    control litter size for cats in the same colony was reported to be
    3.31.5).  All cats in the study completed lactation and reared
    their progeny successfully.  The study also reported that
    biochemical and haematological parameters were within normal ranges
    throughout the study (no data were provided) (Burger  et al.,

    2.2.7  Special studies on gastrointestinal effects.

         There was no change in total faecal microflora count in 30 male
    C3H/He mice fed a diet containing 10% konjac flour (test substance
    was identified as konjac mannan) for one year compared to 30 male
    mice fed a powdered control diet.  Of the 11 specific types of
    microflora that were examined, however, two were significantly
    changed: the frequency of bifidobacteria increased from 30% in
    control mice to 100% in konjac flour-fed mice and the log count of
    enterobacteriaceae increased from 6.0 in control mice to 6.6 in
    konjac flour-fed mice (Mizutani and Mitsuoka 1982).

         Five-week-old male C3H/He mice bearing human flora were either
    maintained on a control diet or fed a diet containing 10% konjac
    flour (test substance was identified as konjac mannan, sterilized by
    gamma-radiation from 60Co).  When the mice were five months old,
    microflora, enzymes, and putrefactive products were analyzed in
    faecal samples from animals in the control and konjac flour-fed
    groups.  Total bacterial counts were nearly identical in control and
    konjac flour-fed mice; however, streptococcus bacteria were
    significantly reduced in the konjac flour-fed mice.  Of the soluble
    enzymes measured, -glucuronidase and nitroreductase activities were
    significantly reduced in the konjac flour-fed mice (67% and 19% of
    activities in control mice, respectively) while azoreductase
    activity was slightly increased (139% of activity in control mice). 
    Several putrefactive metabolites (p-cresol, indole, and skatole)
    were decreased in konjac flour-fed mice compared to control mice. 
    Dietary konjac flour, through material sequestration and lowering of
    substrate concentration, might lead to suppression of bacterial
    enzyme activities and intestinal metabolism without significantly
    affecting microflora composition (Fujiwara  et al., 1991).

         Three-month-old Fisher 344 male rats were switched from a basal
    diet to a similar diet containing 10% konjac flour (test substance
    was identified as konjac mannan, sterilized by gamma-radiation from
    60Co), to which they were allowed  ad libitum access for two
    months.  Soluble enzyme activities were measured in fresh faecal

    samples collected two days before and 19, 29, and 39 days after rats
    were placed on the 10% konjac flour diet.  Putrefactive products
    were analyzed in fresh faecal samples collected the day before and
    on days 20, 30, and 40 after rats were placed on the konjac flour
    diet.  Mean faecal -glucuronidase activity (n = 13 rats) initially
    rose, then decreased significantly to approximately one-third of its
    original level of activity at the end of the experimental period. 
    Two faecal reducing enzyme activities significantly decreased
    following introduction of the konjac flour diet: nitroreductase
    activity was decreased approximately 2-fold and azoreductase
    activity was decreased approximately 5-fold by the end of the study. 
    As well, the gastrointestinal (microflora) metabolites tyrosine and
    tryptophan were significantly altered in rats that consumed a konjac
    flour-containing diet for 2 months.  The reviewers note that
    concurrent control animals were not included in this study, and
    effects attributed to consumption of konjac flour could also be due
    to factors such as age or changes in the test animals' environment
    (Fujiwara  et al., 1991).

         Groups of weanling male Wistar rats (six/group) were fed a
    basal diet (control group: 67% corn starch, 21% casein, 7% corn oil,
    4% salt mix and 1% vitamins) or diets containing 20% konjac flour or
    20% cellulose (test diets were prepared by substituting the test
    substance for an equal weight of corn starch in the basal diet,
    konjac flour was identified as having been derived from
     Amorphophallus konjac.  After eight weeks, total protein, DNA,
    RNA, and the activity of (Na+K)ATPase were determined in homogenates
    of caecal and colonic mucosa.  Compared to rats fed the basal diet,
    feed intake of the konjac flour-fed rats was unchanged but feed
    intake of the cellulose-fed rats was significantly increased (data
    not shown).  After eight weeks, the konjac flour-fed group had a
    significantly reduced mean body weight (90% of control); wet weights
    of the caecum and colon in konjac flour-fed rats were significantly
    increased compared to control rats (approximately 300% and 25%,
    respectively).  In contrast, the average body weight of cellulose-
    fed rats was not significantly different from control rats, but wet
    weights of the caecum and colon in cellulose-fed rats were increased
    by approximately 25% and 60%, respectively.  Total mucosal DNA in
    the caecum and colon of konjac flour-fed rats was significantly
    increased compared to control rats (263% and 159%, respectively);
    total mucosal DNA in the caecum and colon of cellulose-fed rats was
    also increased compared to control rats (148% and 187%,
    respectively).  Ratios of mucosal RNA/DNA and protein/DNA in the
    caecum were significantly increased in konjac flour-fed rats
    compared to control rats.  Caecal and colonic mucosal (Na+K)ATPase
    activities (expressed per mg protein) were significantly increased
    in konjac flour- or cellulose-fed rats compared to control rats;
    konjac flour-fed rats had a greater increase in mucosal ATPase
    activity in the caecum than in the colon (180% vs. 150%,
    respectively) while the opposite was true for cellulose-fed rats
    (150% vs. 219%, respectively).  Based on these results, this report

    suggests that caecal enlargement in rats due to ingestion of konjac
    flour results from both increased number (hyperplasia) and size
    (hypertrophy) of mucosal cells, but that colonic enlargement due to
    ingestion of konjac flour and colonic and caecal enlargement due to
    ingestion results from hyperplasia only (Konishi  et al., 1984).

    2.2.8  Special studies on genotoxicity

         Konjac flour was non-mutagenic in five tester strains (TA98,
    TA100, TA1535, TA1537, and TA1538) of  Salmonella typhimurium in
    the presence or absence of liver microsomal metabolic activation
    (Kotkoskie  et al., 1992).

    2.2.9  Special studies on lipid metabolism

         Venter and coworkers (1990) studied the effect of konjac flour
    on plasma fibrinogen, serum and liver lipid, glucose tolerance,
    insulin response, and liver glycogen in baboons fed a "Western"
    diet.  Twelve male baboons (mean weight of 193 kg) were fed a
    "Western" diet (approximately 400 g/day) with or without konjac
    flour (5%) or sodium propionate (2%) supplements for periods of 9
    weeks in a crossover, randomized order with period of stabilization
    between treatment periods.  The "Western" diet consisted of 38.3 g
    corn meal, 13.4 g beef tallow, 10 g sucrose, and 38.3 g of a
    commercially prepared dietary supplement containing protein,
    vitamins, and minerals per 100 g diet (the supplement results in
    normal growth and excellent health when fed in combination with corn
    meal to young baboons).  Parameters were measured before and 4 and 9
    weeks after the beginning of each treatment period.  After 9 weeks,
    serum total cholesterol levels were statistically significantly
    higher than pretest values in baboons fed the unsupplemented
    "Western" diet, and that konjac flour supplementation of the
    "Western" diet prevented this increase.  Although serum levels of
    high-density lipoprotein increased with all "Western" diets, the
    percentage of total cholesterol as high-density lipoprotein was
    statistically significantly greater in baboons fed the konjac flour-
    supplemented diet for 9 weeks compared to other diets.  Konjac
    flour-supplementation also was reported to statistically
    significantly increase levels of serum triglycerides and circulating
    free fatty acids after 9 weeks.  Finally, liver cholesterol
    concentration was approximately 30% lower and the area under the
    glucose tolerance curve was smaller when baboons were fed konjac
    flour-supplemented diets for 9 weeks.  Because similar effects were
    seen with the sodium propionate-supplemented diet, the effects
    observed in baboons fed konjac flour-supplemented "Western" diets
    may be due to colonic production and absorption of propionate from
    this soluble fiber (Venter  et al., 1990).

    2.2.10  Special studies on nutrient absorption

         Kiriyama and coworkers (1974) studied the effects of chemically
    pure, water-soluble konjac flour (test article was "purified by the
    method of Sugiyama" and purity was tested) on transport of bile
    acids in everted ileal sacs of rats.  Small intestines (not
    including the duodenum) from adult male and female Wistar rats were
    excised, cut into segments, everted, and tied into sacs; these sacs
    were used to evaluate the uptake of 14C-labelled cholate or
    taurocholate  in vitro.  Results of this experiment confirmed that
    the rat ileum actively transports cholic acid against a
    concentration gradient.  Active transport of cholic or taurocholic
    acid was significantly inhibited (2-3-fold) when 0.25% konjac flour
    was added to the media outside of sacs made from the ileum (distal
    small intestine) but not when 0.25% konjac flour was added to the
    media outside of sacs made from the jejunum (proximal small
    intestine).  However, there was no inhibition of transport in the
    presence of 0.05% konjac flour.  No inhibition of transport was seen
    when everted ileal sacs were incubated in a konjac flour-free bile
    acid medium after the sacs had been pre-immersed in 0.25% konjac
    flour.  This suggests that the binding of konjac flour to the
    surface of the intestinal mucosa is not strong enough to effectively
    inhibit the active transport of bile acids.  This report suggests
    that konjac flour does not bind, sequester, or adsorb bile acids
    since, in a simple dialysis experiment, the equilibrium of bile
    acids across a cellophane membrane was not altered by the presence
    of konjac flour on one side of the membrane (Kiriyama  et al.,

         Because previous studies (Reinhold  et al., 1975 and 1976)
    have shown that plant fibers bind minerals, including calcium, zinc,
    and iron, thereby rendering them unavailable for intestinal
    absorption, Oku  et al. (1982) studied the influence of dietary
    konjac flour and other nutritionally unavailable carbohydrates on
    intestinal absorption of calcium.  Six male Wistar rats/group
    (initial body weight, 40-50 g) were fed  ad libitum a basal diet
    containing 67% corn starch (control) or the same diet in which 20%
    cellulose I, 20% cellulose II, 10% konjac flour, 20% konjac flour,
    or 20% pullulan had been substituted for an equal amount of corn
    starch (konjac flour was identified as having been derived from
     Amorphophallus konjac).  Rats were fed control or experimental
    diets for 7 or 8 weeks.  Each rat was placed in a metabolic cage for
    three days at the end of the study.  Body weights were recorded
    weekly; feed consumption was determined daily during the last 3 days
    of the study.

         Compared to rats fed the control diet, rats fed diets
    containing 10% and 20% konjac flour had significantly reduced mean
    body weights (control: 328 g; 10% konjac flour: 296 g; 20% konjac
    flour: 258 g).  Feed consumption of rats fed diets containing 10% or
    20% konjac flour, however, was approximately the same as feed

    consumption of control rats.  Serum levels of calcium and inorganic
    phosphorus were not statistically significantly altered in animals
    fed diets containing konjac flour, although serum calcium levels
    showed a declining trend.  In  in vitro investigation of calcium
    transport using everted duodenal sacs from treated and control rats,
    calcium transport was significantly reduced (to 60% of control
    value) in the 20% konjac flour-fed group only.  Calcium binding
    activity in the supernatant of homogenized duodenal mucosa was also
    significantly depressed in all groups of rats fed diets containing
    20% carbohydrate diets, with the greatest effect associated with
    consumption of diets containing 20% konjac flour (50% of control
    binding).  Based on these results, consumption of unavailable
    carbohydrates, including konjac flour, may be associated with
    decreased function of intestinal epithelial protein(s) essential for
    transport of calcium.  As well, because consumption of konjac flour
    appeared to be associated with a functionally compromised
    gastrointestinal mucosal surface, absorption of minerals other than
    calcium may also be affected (Oku  et al., 1982).

    2.3  Observations in humans

         Several experiments were performed to evaluate the effects of
    konjac fiber on glucose metabolism in normal and diabetic subjects. 
    After a 12-hour fast, 3.9 g konjac flour (test substances were
    identified as "powdered glucomannan", viscosities were 100-150 000
    cP [konjac flour], 52 000 cP [low-viscosity konjac flour], or
    194 800 cP [high-viscosity konjac flour]) was consumed either
    simultaneously with or 15 min. before a glucose load or test meal
    (This dose was selected because of reports that Japanese subjects
    experienced abdominal pain and complications following consumption
    of 5.2 g or more konjac flour.)  Venous blood samples were drawn at
    0 (fasting), 30, 60, 90, 120, and 180 min. after glucose or test
    meal consumption.  When meals containing konjac flour were fed to
    seven non-diabetic subjects, mean serum glucose was significantly
    below levels for control, non-diabetic subjects at 30, 60, 120, and
    180 min.  Serum insulin levels were also significantly decreased at
    30, 60, and 90 min. in these subjects.  Less pronounced effects were
    observed when konjac flour was administered 15 min. before meals;
    under these circumstances, serum glucose levels were significantly
    reduced at 30 and 180 min. only.  In a similar experiment involving
    six non-insulin-dependent diabetics, significant reductions in mean
    serum glucose levels were observed at 30 and 60 min. following
    consumption of konjac flour.  However, no significant reductions in
    serum glucose levels were observed when guar gum, another
    gel-forming fiber consisting of galactose and mannose in a molar
    ratio of 1:2, was fed to subjects instead of konjac flour.

         In a study with nine subjects, low-viscosity konjac flour
    delayed the increase in serum glucose from 30 min. (seen in subjects
    fed a meal without konjac flour) to 60 min (when the same subjects
    were fed an identical meal containing konjac flour).  In addition, a

    high-viscosity konjac flour was more effective in delaying the rise
    in serum glucose following consumption of a konjac flour-containing
    meal than the low viscosity konjac flour.

         Absorption of xylose was measured following administration of
    25 g xylose and 50 g glucose to each of five healthy volunteers. 
    Co-administration of 3.9 g konjac flour caused a significant
    decrease in the excretion of xylose after two hours and appeared to
    prolong the time required for absorption of xylose, since total
    xylose excretion after 6 hours was similar in konjac flour-fed and
    control subjects (no data presented).

         When each of 21 diabetic subjects was fed 7.2 g konjac flour
    daily for 17 days, mean fasting serum glucose levels were
    significantly decreased throughout the study.  When observed for 90
    days after daily feeding of 7.2 g konjac flour for 17 days (no data
    presented), mean serum cholesterol levels decreased significantly
    for the first 38 days, then gradually increased; triglyceride and
    HDL-cholesterol levels, however, were reported not to have been
    affected (Doi  et al., 1982).

         Following an overnight fast, seven young men (22-32 years old;
    1106% of ideal Japanese body weight) were given a 500 ml solution
    containing 80 g glucose (controls) or a similar solution containing
    80 g glucose and 5 g konjac flour (konjac flour was identified as
    having been prepared fresh from  Amorphophallus konjac after prompt
    inactivation of tuber mannanase I and mannanase II by homogenization
    with ethanol).  One week later, the same experiment was performed,
    but treatments for each group of subjects were reversed.  Plasma
    glucose and insulin levels were measured in venous blood samples
    collected from each subject at 0 (fasting sample), 30, 60, 90, 120,
    and 180 min after consuming control and konjac flour-containing
    glucose solutions.  In control and konjac flour-fed subjects, plasma
    glucose and insulin levels peaked within 30 min.; however, both
    parameters were lower (the decrease in serum insulin was
    statistically significant) when subjects consumed the glucose
    solution containing konjac flour.  At 60, 90, and 120 min. following
    administration of the konjac flour-containing glucose solution,
    serum glucose and insulin levels were decreased compared to samples
    following administration of the control glucose solution, although
    the decreases were not statistically significant; at 180 min.,
    however, serum glucose and insulin levels were increased in subjects
    administered the konjac flour-containing glucose solution.  Thus,
    when subjects were challenged with glucose, plasma glucose and
    insulin levels returned more slowly to fasting levels when konjac
    flour was co-administered with the glucose.  Despite the time course
    differences in plasma glucose in subjects following administration
    of a glucose solution or a konjac flour-containing glucose solution,
    the total areas under the plasma glucose curves for 0-180 min. were
    identical.  However, the total area under the plasma insulin curve

    was significantly smaller after administration of the konjac flour-
    containing glucose solution than after administration of the control
    glucose solution.  These results suggest that soluble dietary fibers
    such as konjac flour have beneficial effects on serum glucose levels
    and that these effects may be due to delayed stomach emptying and
    delayed glucose diffusion in the intestinal lumen (Ebihara  et al.,

         The effects of konjac flour on serum glucose levels were
    evaluated in 72 type II diabetic subjects (mean age 55, range 39-76)
    by Huang and coworkers; subjects were grouped as mild, moderate, or
    severe diabetics.  Meals containing 2% refined konjac flour in the
    form of konjac toast or konjac noodles were consumed by test
    subjects (average intake was 8.6 g konjac flour/day) for
    approximately 65 days.  Weekly food intakes were recorded three
    times: before konjac flour-ingestion began and during weeks three
    and seven of the study.  Fasting venous blood samples and 2-hr
    postprandial blood samples (before and after breakfast,
    respectively) were drawn once before konjac flour-ingestion began
    and on days 30 and 65 of the study; the following determinations
    were made: fasting blood glucose (FBG), 2-hour post-prandial blood
    glucose (PBG), glycosylated haemoglobin (GHB), total cholesterol
    (TC), high-density lipoprotein cholesterol (HDL-C), low-density
    lipoprotein cholesterol (LDL-C), and triglyceride (TG).  Compared to
    levels measured before konjac flour-ingestion began, there were
    statistically significant reductions in FBG levels on days 30 and 65
    and in GHB on day 65.  Konjac flour appeared to be particularly
    beneficial to subjects with higher levels of blood glucose because 
    the extent of konjac flour-associated decreases in FBG and PBG
    appeared to be directly proportional to the severity of subjects'
    pre-test levels of FBG and PBG.  For instance, mild, moderate, and
    severe diabetic groups had the following initial vs. final FBG
    levels: 129.410.5 vs. 124.026.8; 165.910.9 vs. 142.635.3; and
    227.025.3 vs. 171.437.4, respectively.  Finally, the study
    reported a positive correlation between konjac flour-associated
    decreases in FBG and PBG.

         In general, food containing konjac flour did not appear to have
    a significant effect on blood lipids.  Several subjects were
    reported to have experienced weight loss during the study: the mean
    weight loss of 42/59 subjects at 30 days was 1.8 kg (range
    0.2-4.7 kg) and the mean weight loss for 41/51 subjects at 65 days
    was 2.2 kg (range 0.5-6.0 kg).

         In this study, subjects reported the following symptoms
    associated with consumption of meals containing konjac flour: 69-90%
    of the subjects reported improved appetite, polyuria at night,
    thirst, and constipation or soft stool; on the other hand, 40
    subjects reported a total of 45 symptoms, such as loose stool,
    flatulence, diarrhoea, and abdominal pain, sounds, or distension
    (Huang  et al., 1990).

         A recent study compared the effects of four non-starch
    polysaccharides, including konjac flour, on glucose tolerance,
    insulin secretion, gastric emptying, and gut hormone secretion in 12
    healthy male volunteers (mean age = 19.5 years).  Subjects were
    divided into two groups; on three separate occasions (at least one
    week apart) and following an overnight fast, members of each group
    were given a test meal (white bread, honey, and orange squash
    containing 100 g carbohydrate) either with or without 10 g soya-bean
    cotyledon fiber (SCF) or 5 g konjac flour.  Soluble paracetamol
    (1.5 g dissolved in 150 ml water) was consumed simultaneously with
    each meal; over the following 180 min., plasma levels of paracetamol
    were measured as an index of liquid gastric emptying.  Plasma
    concentrations of glucose, insulin, and gastric inhibitory
    polypeptide (GIP) were measured during the same time period.  Konjac
    flour-supplemented meals had a significant depressive effect on
    post-prandial insulin levels, where peaks for control and konjac
    flour-supplemented meals were 96.211 and 64.32.6 mU/l,
    respectively.  The 0-90 min. incremental area under the insulin
    curve was significantly reduced following consumption of a konjac
    flour-supplemented meal (82% of area following consumption of a
    control meal), although consumption of a meal containing SCF
    increased this parameter (119% of area following consumption of a
    control meal).  Post-prandial plasma glucose, GIP and paracetamol
    levels were not affected by consumption of meals containing konjac
    flour or SCF (Morgan  et al., 1990).

         Results of this study are consistent with results reported in
    similar experiments by Ebihara  et al. (1981) but were inconsistent
    with results reported by Doi  et al. (1982), which  Morgan and
    coworkers attributed to possible differences in the composition of
    test meals (not specified in the study by Doi  et al., 1982) that
    may have affected gastric emptying.

         Zhang and coworkers studied the effects of consuming konjac
    flour-supplemented diets (5 g konjac flour/day) for 45 days on human
    lipid metabolism in elderly subjects with hyperlipidaemia.  Subjects
    were described as having borderline risk levels of serum
    triglycerides (200-230 mg/dl), hypercholesterolaemia (>230 mg/dl),
    borderline risk levels of serum triglycerides (130-150 mg/dl),
    hypertriglyceridaemia (>150 mg/dl), or both hypercholesterolaemia
    and hypertriglyceridaemia.  Subjects were randomly divided into two
    groups: the konjac flour group (66 subjects) consumed normal
    Japanese diets supplemented with foods containing konjac flour
    (3 g/day for the first 2 or 3 days, then 5 g/dy for the rest of the
    study); the control group (44 subjects) consumed normal Japanese
    diets without konjac flour supplementation; a recovery group (46 of
    66 subjects in the konjac flour group) consumed normal Japanese
    diets for an additional 45 days.  Venous blood samples were drawn
    before breakfast before the study began, at the end of the study
    (45 days) and at the end of the recovery period (an additional 45
    days).  Konjac flour used in this study was described as "konjac

    meal...in the form of noodles, breads, cakes, etc., to which it was
    added, or after it was cooked directly with milk, soya-milk, soup,
    water, or stuffing."

         After consumption of konjac flour for 45 days, subjects had
    statistically significantly decreased levels of serum total
    cholesterol, triglycerides, and low-density lipoprotein and
    statistically significantly increased levels of high-density
    lipoprotein and apoprotein; however, changes in these parameters in
    the control group during the study were not significant.  In
    addition, konjac flour subjects had statistically significantly
    decreased body weight, serum total cholesterol, serum triglycerides,
    and serum low-density lipoprotein and increased serum high-density
    lipoprotein compared to control subjects after 45 days.  After
    consumption of konjac foods, 27 subjects had statistically
    significant decreases in body weight (0.5 to 4.7 kg at 45 days);
    subjects consuming konjac flour reported decreases in dizziness and
    headaches (9/66) and chronic constipation (13/66); however, subjects
    also reported that they excreted more bulk faeces (17/66),
    experienced diarrhoea (3/66), and had increased hunger (4/66). 
    After the 45-day recovery period, subjects (46 of 66 subjects who
    had previously consumed konjac flour for 45 days) had statistically
    significant increases in serum total cholesterol and low-density
    lipoprotein and a statistically significant decrease in high-density
    lipoprotein; serum triglyceride levels remained the same.  This
    report concluded that beneficial effects observed in
    hypercholesterolaemic patients following consumption of 5 g konjac
    flour/day for 45 days were reversible upon discontinuance of konjac
    flour in the diet (Zhang  et al., 1990).

         The efficacy of konjac flour (test substance identified as
    "glucamannan") for weight loss in overweight osteoarthritic patients
    was evaluated in a double-blind crossover study.  Twenty outpatients
    (3 men and 17 women; mean age 53 years, range 30-68 years) were
    divided into two groups.  For the first two months of the study,
    each subject was administered 1.5 g of konjac flour or a placebo
    before breakfast and dinner; for the following two months,
    treatments were reversed.  Sixteen subjects completed the study; one
    subject stopped because of abdominal discomfort during the konjac
    flour-treatment period.  Body weight was recorded monthly. 
    Decreases in body weight during konjac flour-phases of the study
    (first group of subjects: mean body weight decreased from 73.7 to
    70.1 kg; second group of subjects: mean body weight decreased from
    79.0 to 76.4 kg) were reported to be statistically significant and
    were attributed to konjac flour administration.  Study stated that
    subjects experienced no changes in blood pressure, heart rate, or
    clinical chemistry, but no data were provided (Biancardi  et al.,

         The effect of konjac flour on weight loss and blood levels of
    total cholesterol, total triglycerides, and low-density lipoprotein
    was investigated in a double-blind study of obese women for 8 weeks. 
    Twenty obese women (weights were at least 20% over ideal body
    weights) were randomly assigned to two groups (10/group) so that
    weight and height distributions for each group were similar.  The
    konjac flour group took two capsules of a supplement containing
    500 mg "purified glucomannan" three times a day (with 8 oz water),
    one hour before each meal, for 8 weeks; control subjects took a
    similar capsule containing 500 mg starch under the same conditions. 
    At the beginning of the study and at 4 and 8 weeks, subjects were
    weighed and blood samples were drawn.  There were statistically
    significant decreases in weight, serum cholesterol and serum
    low-density lipoproteins in the konjac flour group compared to the
    control group at weeks 4 and 8.  A statistically significant
    decrease in serum triglycerides occurred at four weeks, but the
    decrease was not statistically significant at 8 weeks.  This study
    suggests that konjac flour can be used for weight reduction and
    serum cholesterol reduction in obese patients (Walsh  et al.,

          The efficacy of konjac flour was studied in 93 patients with
    chronic constipation; 78 patients completed the study.  In this
    non-controlled, open study, patients were treated with 1 g konjac
    flour/day for one month, followed by 1 g konjac flour every other
    day for one month.  Statistically significant improvements were seen
    in several measures of bowel function (number of days per week with
    bowel movements and number of enemas given) that lasted until the
    end of the second month.  Konjac flour was well accepted by patients
    and without significant side effects; in addition, konjac flour
    reduced (by approximately 50%) the number of patients reporting
    abdominal disturbances by the end of the study (Passaretti  et al.,

         The influence of dietary konjac flour (test substance
    identified as konjac mannan) on absorption of vitamins E (tocopherol
    acetate) and B12 (mecobalamin) was evaluated in six normal
    volunteers (aged 20-61) and five maturity-onset diabetic patients
    (aged 57-81) by Doi and coworkers.  Control meals containing
    vitamins E and B12 (500 mg and 3 000 g, respectively) were
    consumed by each subject; test meals to which 3.9 g konjac flour had
    been added were consumed by subjects on an alternate day.  Vitamin
    levels in serum were determined immediately before and 1, 3, 5, 8,
    12, and 24 hr after consumption of both meals.  In normal subjects,
    the peak serum level of vitamin B12 was shifted from 3 hr to 12 hr
    when the meal was supplemented with konjac flour; however, no shift
    was seen in the 5-8 hr peak time observed in diabetics when konjac
    flour was added to the meal.  Total absorption of vitamin B12
    during 24 hr did not appear to be changed by addition of konjac
    flour to the test meal.  Intestinal absorption of vitamin B12 was
    not disturbed by konjac flour.

         Addition of konjac flour to the test meal appeared to
    significantly lower serum levels of vitamin E, especially in normal
    subjects (by 16-30% at most time points).  The results suggested
    that viscous forms of dietary fibers, such as konjac flour, may form
    a barrier around some (fat-insoluble) substances (including glucose,
    essential electrolytes and cations, and possibly vitamin B12),
    thereby delaying their absorption rather than causing malabsorption. 
    On the other hand, because konjac flour consumption may interfere
    with the absorption of bile acids, the absorption of the fat-soluble
    vitamin E, a process which is dependent on the presence of
    conjugated bile acids, may also be impaired by konjac flour
    consumption (Doi  et al., 1983).

         The effect of konjac flour (identified as glucomannan in the
    form of konjac flour) on intestinal absorption of glibenclamide, a
    sulfonylurea-type hypoglycaemic drug, was evaluated in nine healthy
    male volunteers, aged from 21 to 47 years.  All subjects
    participated, on two consecutive days, in both the control and test
    phases of the experiment.  Plasma concentrations of the drug were
    followed for six hours after subjects received the drug (2.5 g)
    alone (control phase) or with 3.9 g konjac flour (test phase); in
    both phases, subjects ate breakfast immediately after receiving the
    drug.  Compared to plasma drug levels during the control phase,
    subjects in the test phase had lower drug levels at 30 min. and this
    tendency continued for most subjects until at least 180 minutes.  In
    the control phase, mean plasma levels of glibenclamide peaked
    rapidly at 60 min. and declined more slowly; when konjac flour was
    administered with the drug, however, there was no apparent peak in
    serum drug level.  Co-administration of konjac flour with
    glibenclamide appeared to increase inter-individual variation in
    plasma drug levels.  This study also demonstrates that dietary
    intake of konjac flour can influence the pharmacokinetics of
    co-administered oral drugs (Shima  et al., 1983).

         When dry, non-expanded konjac flour was marketed as a dieting
    aid, 7 case reports of oesophageal obstruction caused by swelling of
    these tablets were made.  The case reports were reviewed in 1986:
    all patients were women taking the konjac flour tablets for weight
    loss; obstruction was complete in 5/7 cases and was caused by a
    single tablet in all but one case; obstruction was presumably caused
    by swelling of dry konjac when hydrated by body fluids; none of the
    patients died.  Expansion rates of 5 tablets (500 mg konjac flour
    each) immersed in water varied, but tablets increased their mean
    volumes from one to approximately 16 ml in 10 min; final volumes
    ranged from 12 to 17.5 ml.  It is important to note that konjac
    flour in food is not in this dry form and there are no reported
    cases of oesophageal obstruction caused by hydrated konjac gels,
    such as those traditionally eaten in Japan (Henry  et al., 1986).

         Inhalation of konjac dust ("dancing powder" in factories
    producing konnyaku, a popular food in Japan made from konjac tubers)
    in the workplace has been reported to produce allergic bronchial
    asthma (known as konnyaku asthma) in sensitized individuals
    (Nakazawa, 1983).  In Japan, konnyaku asthma occurs in workers
    exposed to konjac dust and in people who live close to konjac
    processing facilities.  An epidemiological study of Japanese plant
    employees and nearby residents conducted in 1980 found that 0.1% of
    the population (15 675 people) had konnyaku asthma.  Konnyaku asthma
    can be treated by desensitization therapy and disappears when the
    individual is no longer exposed to konjac dust.  It should be noted
    that inhalation of konjac powder has also been reported to cause
    allergic asthma in sensitized guinea pigs (Banno, 1979), although
    application of mechanically ground konjac to the skin of guinea pigs
    did not produce skin sensitization reactions (Kotkoskie, 1992).  In
    both humans and animals, konjac powder or dust has been shown to
    produce allergic bronchial asthma by a type I or anaphylactic immune
    reaction (Nakazawa, 1983).


         The Committee reviewed data from acute and short-term toxicity
    studies, as well as studies on embryotoxicity, genotoxicity,
    nutrient absorption, anti-carcinogenicity, gastrointestinal effects,
    and observations in humans.  However, the Committee was concerned
    about the lack of information on the fate of konjac flour in the gut
    and the inadequacy of the short-term toxicity studies.  The
    Committee was informed of the existence of a 13-week toxicity study
    in dogs and a 4-week toxicity study in rats that were not made
    available for review at the present meeting.

         Human studies were conducted for up to 65 days at dose levels
    of up to 8.6 g konjac flour/person/day.  Volunteers consuming
    approximately 5.2 g or more konjac flour/person/day reported
    symptoms such as loose stools, flatulence, diarrhoea and abdominal
    pain or distension.  Studies with normal and diabetic volunteers
    demonstrated that consumption of 7.2-8.6 g konjac flour/person/day
    for 17 days significantly decreased mean fasting serum glucose
    levels; in addition, a dose of 3.9-5.0 g konjac flour/person
    consumed with a single meal (or administered with glucose) was
    reported to delay the increase in serum glucose and insulin levels
    for several hours following the meal, thereby also delaying their
    return to baseline levels.  Consumption of test meals containing
    3.9 g konjac flour appeared to impair vitamin E absorption (up to
    30% decrease in peak serum levels) and influenced the
    pharmacokinetics of the co-administered drug glibenclamide.  On the
    other hand, intestinal absorption of vitamin B12 or the drug
    paracetamol were not affected by consumption of meals containing
    konjac flour (3.9 g konjac flour/person for vitamin B12 and 5 g
    konjac flour/person for paracetamol).


         On the basis of the available toxicological data, particularly
    data from human studies, the long history of use of konjac as a food
    in China and Japan, and estimates of konjac flour consumption from
    traditional and anticipated food additive uses, the Committee
    allocated a temporary ADI "not specified" for konjac flour.  The
    results of additional short-term toxicity studies, which the
    Committee was informed have been conducted in rats and dogs,
    together with adequate data on the fate of konjac flour in the gut
    are required for review in 1996.  In view of the observed impairment
    of absorption of vitamin E, information on the influence of konjac
    flour on the bioavailability of fat soluble vitamins is also
    required for review by 1996.  The Committee noted that consumption
    of dry konjac flour has been associated with oesophageal
    obstruction, and recommended that konjac flour be consumed only in
    the hydrated form.


    BANNO, H. (1979).  Cyclic GMP and cyclic AMP levels in the nasal
    mucosa, trachea, lung, and plasma of guinea pigs sensitized by the
    inhalation of konjac powder.   Allergy (Jpn.) 28, 595-601.

    BIANCARDI, G., PALMIERO, L., & GHIRARDI, P.E. (1989).  Glucomannan
    in the treatment of overweight patients with osteoarthrosis.   Curr.
     Ther. Res. 46, 908-912.

    BURGER, I.H., EARLE, K.E., & BAILIE, H. (1992).  Evaluation of
    Konjac flour for use in a commercially prepared cat food.  Submitted
    to WHO by  Waltham Center for Pet Nutrition, Waltham-on-the-Wolds,
    Leicestershire, England.

    DOI, K., MATSUURA, M., KAWARA, A., UENOYAMA, R., & BABA, S. (1982).
    Effect of glucomannan (konjac fiber) on glucose and lipid metabolism
    in normal and diabetic subjects.   Int. Cong. Ser.-Excerpta Med
     (Genet. Environ. Interact. Diabetes Mellitus). 549 , 306-312.

    DOI, K., MATSUURA, M., KAWARA, A., TANAKA, T., & BABA, S. (1983). 
    Influence of dietary fiber (konjac mannan) on absorption of vitamin
    B12 and vitamin E.   Tohoku J. Exp. Med. 141 (Suppl.) 677-681.

    EBIHARA, K., MASUHARA, R., & KIRIYAMA, S. (1981).  Effect of konjac
    mannan, a water-soluble dietary fiber, on plasma glucose and insulin
    responses in young men undergoing glucose tolerance test.   Nutr.
     Rep. Int. 23, 577-583.

    (1991).  Effect of konjac mannan on intestinal microbial metabolism
    in mice bearing human flora and in conventional F344 rats.   Fd.
     Chem. Toxic. 29, 601-606.

    ROHAN, A. (1986).  Glucomannan and risk of oesophageal obstruction. 
     British Medical Journal 292, 591-592.

    HUANG, C.-Y., ZHANG, M.-Y., PENG, S.-S., HONG, J.-R., WANG, X.,
    JIANG, H., ZHUANG, F., BAI, Y., LIANG, J., YU, Y., LUO, Z., ZHANG,
    X., & ZHAU, Z.(1990).  Effect of konjac food on blood glucose level
    in patients with diabetes.   Biomed. Environ. Sci. 3, 123-131.

    JAPAN FOOD RESEARCH LABORATORIES, Osaka, Japan for Shimizu Chemical
    Industries (1984).  On change in molecular size distribution caused
    by acid hydrolysis.  Unpublished report No. OS27060285-2.  Submitted
    to Food and Drug Administration by Hazelton Laboratories Corp.,
    Herndon, VA, USA.   Food Master File 000348 Vol. 2, 237-245.

    KIRIYAMA, S., ENISHI, A., & YURA, K. (1974).  Inhibitory effect of
    konjac mannan on bile acid transport in the everted sacs from rat
    ileum.   The J. of Nutr., 104, 69-78.

    KONISHI, F., OKU, T., & HOSOYA, N. (1984).  Hypertrophic effect of
    unavailable carbohydrate on cecum and colon in rats.   J. Nutr. Sci.
     Vitaminol., 30, 373-379.

    G.C., HARDY, C.J., & FLETCHER, M.J. (1992).  Acute toxicity studies
    with konjac flour.   Toxicolo. Letters Suppl. 281.

    MIZUTANI, T., & MITSUOKA, T. (1982).  Effect of konjac mannan on
    spontaneous liver tumorigenesis and fecal flora in C3H/He male mice. 
     Cancer Letters 17, 27-32.

    MIZUTANI, T. & MITSUOKA, T. (1983).  Effect of konjac mannan on
    1,2-dimethylhydrazine-induced intestinal carcinogenesis in Fisher
    344 rats.   Cancer Letters 19, 1-6.

    MORGAN, L.M., TREDGER, J.A., WRIGHT, J., & MARKS, V. (1990).  The
    effect of soluble- and insoluble-fibre supplementation on
    post-prandial glucose tolerance, insulin, and gastric inhibitory
    polypeptide secretion in healthy subjects.   British J. Nutr. 64,

    NAKAZAWA, T. (1983).  Studies on agriculture and asthma.   Jpn. J.
     Traumatol. Occup. Med. 32, 10-17.

    NISHINARI, K., WILLIAMS, P.A., & PHILLIPS, G.O. (1992).  Review of
    the physico-chemical characteristics and properties of konjac
    mannan.   Food Hydrocoll. 6, 199-222.

    KIRIYAMA, S. (1984).  Toxicity studies on glucomannan (1) Acute
    toxicity study in mice and rats.   Appl. Pharmacol. 27, 127-131.

    OKU, T., KONISHI, F., & HOSOYA, N. (1982).  Mechanism of inhibitory
    effect of unavailable carbohydrate on intestinal calcium absorption. 
     The J. of Nutr. 104, 410-415.

    TITTOBELLO, A. & participating physicians. (1991).  Action of
    glucomannans on complaints of inpatients affected with chronic
    constipation: a multicentric clinical evaluation.   Ital. J.
     Gastroenterol. 23, 421-425.

    REINHOLD, J.G., ISMAIL-BEIGI, F., & FARADJI, B. (1975).  Fibre vs.
    phytate as determinant of the availability of calcium, zinc, and
    iron of breadstuffs.   Nutr. Rept. Int. 12, 75-85.

    Decreased absorption of calcium, magnesium, zinc, and phosphorus by
    humans due to increased fiber and phosphorus consumption as wheat
    bread.   J. Nutr. 106, 493-503.

    KUMAHARA, Y. (1983).  Effect of dietary fiber, glucomannan, on
    absorption of sulfonylurea in man.   Horm. Metabol. Res. 15, 1-3.

    VENTER, C.S., VORSTER, H. H., & Van der NEST, D. G. (1990). 
    Comparison between physiological effects of konjac-glucomannan and
    propionate in baboons fed "Western" diets.   J. Nutr., 120(9)

    WALSH, D.E., YAGHOUBIAN, V., & BEHFOROOZ, A. (1984).  Effect of
    glucomannan on obese patients: a clinical study.   International
     Journal of Obesity 8, 289-293.

    ZHANG, M.-Y., HUANG, C.-Y., WANG, X., HONG, J.-R., & PENG, S.-S.
    (1990).  The effect of foods containing refined konjac meal on human
    lipid metabolism.   Biomedical and Environmental Sciences 3,

    See Also:
       Toxicological Abbreviations
       Konjac flour (WHO Food Additives Series 37)
       KONJAC FLOUR (JECFA Evaluation)