IPCS INCHEM Home


    INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY

    WORLD HEALTH ORGANIZATION





    SAFETY EVALUATION OF CERTAIN FOOD
    ADDITIVES AND CONTAMINANTS



    WHO FOOD ADDITIVES SERIES: 44





    Prepared by the Fifty-third meeting of the Joint FAO/WHO
    Expert Committee on Food Additives (JECFA)





    World Health Organization, Geneva, 2000
    IPCS - International Programme on Chemical Safety

    SWEETENING AGENT


    ERYTHRITOL

    First draft prepared by J.A. Eastwood and E.J. Vavasour

    Chemical Hazard Assessment Division, Bureau of Chemical Safety, Food
    Directorate, Health Protection Branch, Health Canada, Ottawa, Ontario,
    Canada

    Explanation
    Biological data
         Biochemical aspects
              Absorption, distribution, excretion, and biotransformation
         Toxicological studies
              Acute toxicity
              Short-term studies
              Long-term studies of toxicity and carcinogenicity
              Genotoxicity
              Reproductive toxicity
              Developmental toxicity
              Special studies on renal function
         Observations in humans
              Single doses
              Repeated doses
    Comments
    Evaluation
    References

    1.  EXPLANATION

         Erythritol is a four-carbon sugar alcohol
    ( meso-1,2,3,4-butanetetrol) which is 60-80% as sweet as sucrose. It
    is intended for use as a low-calorie sweetener. Erythritol is
    manufactured from glucose or sucrose by fermentation with
     Trichosporonoides megachiliensis or  Moniliella pollinis, which are
    non-pathogenic, non-toxicogenic yeasts. Erythritol also occurs
    naturally in fruits and mushrooms and is present in various fermented
    products, including wine, sake, and soya sauce, generally at a low
    concentration (700-1300 mg/kg), but in the exceptional case of a
    single species of mushroom, at 34 000 mg/kg. It is often detected in
    human and animal tissues and body fluids, including the lens,
    cerebrospinal fluid, serum, semen, and urine.

         The technical characteristics of erythritol, such as its cooling
    effect and low hygroscopicity, are more like those of xylitol than
    those of sorbitol, which represents the major market share of
    sweetening agents. If erythritol were used to replace all xylitol (20%
    of all polyol use), the projected mean intake would be 1 g/day and the
    90th percentile intake would be 4 g/day; if it were used to replace
    all polyols, the mean intake would be 4-5 g/day and the 90th
    percentile intake would be 20 g/day on the basis of estimates for
    diabetic patients.

         Erythritol has not previously been reviewed by the Committee.

    2.  BIOLOGICAL DATA

    2.1  Biochemical aspects

    2.1.1  Absorption, distribution, and excretion

         Mice

         Excretion of erythritol in urine and faeces was assessed over two
    days in male and female CD-1 mice which received 5, 10, or 20% in the
    diet for 13 weeks. In animals of each sex, most of the administered
    erythritol was recovered in urine (85-100% in males and 60-85% in
    females) and small amounts were found in the faeces (4-7% in males and
    3-5% in females). In both males and females, the percentage of the
    ingested dose excreted in urine increased with dose, while the
    percentage excreted in faeces decreased. The authors concluded that
    the erythritol that was not recovered in urine and faeces had been
    fermented in the large intestine (Til  et al., 1992, 1996).

         Rats

         Groups of three male Wistar rats were given a single oral dose of
    0.125, 0.25, 0.5, 1, or 2 g 14C-erythritol per kg bw after an 18-h
    fast, and they were fasted for an additional 8 h after dosing. Blood
    samples were collected from the tail vein 15 and 30 min and 1, 2, 4,
    6, 8, 12, 24, 48, and 72 h after dosing in order to measure
    radioactivity. The pharmacokinetics were calculated on the basis of a
    two-compartment model. At all doses, the maximum blood concentrations
    (Cmax) were found 1 h after dosing. Over the dose range 0.125-1 g/kg
    bw, there was a linear increase in Cmax and in the integrated area
    under the concentration-time curve for 0-72 h (AUC0-72), but the
    increase deviated from this relationship for the dose of 2 g/kg bw,
    suggesting that the absorption of erythritol was saturated. After the
    initial absorption, a biphasic decrease in the blood concentration of
    radiolabel was seen, with half-time values for the biphasic decrease
    of about 2 and 20 h, respectively. The biphasic decrease was also
    observed at 2 g/kg bw, but the half-time values were longer (3 and 28
    h), indicating a slower rate of elimination. The blood erythritol
    concentrations were 2-7% of the Cmax values within 24 h after dosing,
    suggesting that erythritol is generally rapidly eliminated (Nakayama,
    1990a; Noda et al., 1996).

         Three male Wistar rats were given 14C-erythritol as a single
    dose of 1 g/kg bw after an 18-h fast and were again deprived of feed
    for an additional 8 h. Blood samples were collected 30 min and 1, 4,
    8, and 24 h after dosing in order to measure radiolabel in blood and
    plasma and then calculate the ratio of distribution to erythrocytes.
    Urinary and faecal samples were collected separately over 1-8, 8-24,
    24-48, 48-72, and 72-120 h, and carbon dioxide was collected in
    expired air for determinations of radiolabel. After the final

    collection of excreta, the rats were killed and homogenized.
    14C-Erythritol was administered at a dose of 1 g/kg bw to three rats
    which had received cannulae into the common bile duct, and bile
    samples were collected for periods of 0-0.5, 0.5-1, 1-2, 2-4, 4-8,
    8-24, and 24-48 h to allow determination of radiolabel in bile.

         The mean ratio of distribution of erythritol in blood cells rose
    rapidly during the first 4 h after dosing and then more slowly to
    reach a maximum of 48% at 24 h. The mean plasma protein binding
    increased from 0.68% at 30 min to 86% at 24 h, when the plasma
    concentrations were very low. The high protein binding at low
    concentrations of erythritol suggested to the authors that a small
    portion is firmly bound to plasma protein, although the radiolabel
    could also have been derived from metabolites generated by
    gastrointestinal fermentation, and the authors indicated that the
    effect could reflect analytical uncertainties at very low
    concentrations of radiolabel. Most of the radiolabel was recovered in
    urine, urinary recovery of radiolabel accounting for 75% of the
    administered dose within 8 h and 91% within 24 h of dosing. During the
    120-h collection period, about 5% of the administered dose was
    recovered in expired air and about 1% in faeces, most being collected
    within the 8-24-h collection period. Biliary excretion during the
    first 48 h after dosing accounted for < 1% of the administered dose,
    and < 1% of the administered radiolabel was recovered in the carcass
    (Nakayama, 1990b).

         Groups of three male Wistar rats were given a single oral dose of
    1 g/kg bw of 14C-erythritol after an 18-h fast. Radiolabel was
    determined in blood, plasma, tissue samples (cerebrum, cerebellum,
    pituitary gland, Harderian gland, eyeball, thyroid gland, thymus
    gland, lung, heart, liver, spleen, pancreas, kidney, adrenal gland,
    testis, seminal vesicle, mesenteric lymph node, stomach, small
    intestine, caecum, large intestine, bladder, bone marrow, muscle,
    white fat, and brown fat), and the remaining carcass at 0.5, 1, 4, 8,
    and 24 h after dosing. An additional three rats per sacrifice were
    killed and rapidly frozen for whole-body radiography, for which thin
    sections of the whole body were prepared and contacted to highly
    sensitive X-ray films to obtain whole-body autoradiograms.

         Treatment resulted in a rapid increase in blood and plasma
    concentrations, which peaked 60 and 30 min after treatment,
    respectively. The concentration in plasma was higher than that in
    whole blood at both 30 and 60 min, suggesting that equilibrium with
    blood cells was delayed; at subsequent sacrifices, similar values were
    obtained for blood and plasma. The concentrations of radiolabel were
    highest after 30 min in kidney and liver and after 1 h in bladder and
    exceeded those in plasma. Throughout the 24 h after administration,
    the concentration of radiolabel in brain was noticeably lower than
    that in plasma. Within the gastrointestinal tract, the highest
    concentrations of radiolabel were found in the stomach and small
    intestine 30 min after treatment, and the concentrations in the

    stomach were 25 times higher than those in plasma. Thereafter, the
    concentrations in the stomach declined, while those in the caecum and
    large intestine increased. Unabsorbed erythritol quickly reached the
    caecum, as the concentrations in the caecum were higher than those in
    the small intestine by 60 min and reached a peak 4 h after
    administration. The results of autoradiography confirmed these
    observations (Nakayama, 1990c; Noda et al., 1996).

         In a study to compare the effects of dose, route of
    administration, adaptation of gastrointestinal flora, and structural
    relationship on the disposition of erythritol, groups of three or four
    male rats were fasted for 18 h before receiving: a single dose of
    0.01, 0.1, or 1 g/kg bw of 14C-erythritol by gavage; a single dose
    of 0.1 g/kg bw of 14C-glycerol by gavage; a single dose of 0.1 g/kg
    bw of 14C-erythritol intravenously into the tail vein; or a diet
    containing 10% erythritol (reported as > 8 g/kg bw per day) for two
    weeks before receiving a single dose of 0.1 g/kg bw 14C-erythritol
    by gavage. Immediately after dosing, the rats were transferred to
    individual metabolism cages and urine, faeces, and expired carbon
    dioxide were collected separately 2, 4, 6, 8, 10, 12, and 24 h after
    dosing for determination of radiolabel. Only expired carbon dioxide
    was collected from the adapted group. Pooled urine samples from the
    0-8-and 8-24-h intervals from the groups given 0.1 g/kg bw erythritol
    and glycerol by gavage were analysed by thin-layer chromatography and
    high-performance liquid chromatography (HPLC) in order to identify
    radiolabelled compounds. Blood, liver, kidney, testis, and seminal
    vesicles were collected at sacrifice from animals given erythritol by
    gavage and intravenously 24 h after dosing, and the contents of the
    stomach, small intestine, caecum, and colon were collected and pooled
    for determination of radiolabel.

         After administration of 0.1 g/kg bw 14C-erythritol by gavage,
    rats excreted 88% of the radiolabel in the urine within 24 h, about
    70% of which within the first 10 h. HPLC indicated that the radiolabel
    was associated exclusively with intact erythritol. The concentration
    of 14C-carbon dioxide in expired air increased slowly, reaching a
    peak about 8 h after dosing, and about 6% of the administered dose was
    recovered during the first 24 h after treatment. Faecal excretion of
    14C accounted for < 1% of the administered dose. In contrast, rats
    treated with 14C-glycerol excreted about 60% of the radiolabel as
    expired carbon dioxide within 24 h, and only a small proportion of the
    total was recovered in urine and faeces. Rats receiving 0.01 g/kg bw
    14C-erythritol by gavage excreted similar proportions of the dose in
    urine, expired air, and faeces as those at 0.1 g/kg bw, while rats
    receiving 1 g/kg bw 14C-erythritol by gavage excreted less of the
    dose in urine (65%) and more in expired air (16%). At 0.1 g/kg bw, the
    patterns of excretion of radiolabel were similar after intravenous or
    gavage administration, and slightly more was excreted in urine (94%
    over the 24-h period; 30% after 2 h; 50% after 4 h) and somewhat less
    in expired air (1%). While comparable amounts of the dose were

    recovered in the caecum and colon of rats dosed with 0.01 and 0.1 g/kg
    bw (< 1% of the administered dose), 4% of the dose was present in the
    caecum and colon of rats dosed with 1 g/kg bw. The proportion of the
    dose retained in the contents of the stomach and small intestine was
    10-20 times less than that in caecum and colon at all doses. The
    amounts of erythritol equivalents retained in the blood, liver, and
    kidney were increased in a higher proportion to dose in animals at the
    high dose. The erythritol equivalents in tissues exceeded those in the
    blood at 24 h at all doses, except for those in testis and seminal
    vesicles at the high dose. The concentration of erythritol equivalents
    in rats given 0.1 g/kg bw of 14C-erythritol intravenously was
    significantly lower ( p < 0.05) than that in rats given the same
    dose orally. Adapted rats expired significantly more 14C-carbon
    dioxide 6 h after administration than unadapted rats that received the
    same dose of 14C-erythritol, but the amount of expired 14C-carbon
    dioxide did not differ subsequently. Over the 24-h collection period,
    adapted rats excreted about 10% and unadapted rats 6% of the
    radiolabel as 14C-carbon dioxide.

         These studies show that orally administered 14C-erythritol is
    well absorbed and rapidly excreted in the urine. Similar proportions
    of a dose of 0.1 g/kg bw erythritol given orally and intravenously
    were excreted in the urine, but at a dose 10 times higher,
    substantially less erythritol was excreted in urine and a larger
    proportion in expired carbon dioxide. The near absence of radiolabel
    in expired carbon dioxide after intravenous administration suggests
    that fermen-tation of erythritol by the gastrointestinal flora and
    subsequent absorption of the metabolites was the source of the
    14C-carbon dioxide in expired air. A two-week adaption period prior
    to administration of 0.1 g/kg bw by gavage also resulted in a higher
    proportion in expired air but still less than that expired after the
    high dose. The larger proportions of residual radiolabel in blood,
    liver, and kidney after the high gavage dose when compared with lower
    doses and after gavage when compared with intravenous administration
    were probably due to absorption of metabolites of erythritol from the
    gut (Noda & Oku, 1990, 1992).

         Groups of 11 male Wistar rats were fed control diet or control
    diet containing 10% erythritol (added at the expense of corn starch)
    for two weeks. They were then sacrificed, the caecal contents were
    collected and pooled by group, and the contents were suspended.
    Samples of each suspension were incubated with 12 mg 14C-erythritol
    (10 µCi) for 6 h under anaerobic conditions, and the incubation
    mixture was analysed for erythritol, short-chain fatty acids, and
    carbon dioxide 1, 2, 4, and 6 h after the beginning of incubation. The
    total recoveries of radiolabel were comparable for control and treated
    groups at the end of incubation, but the proportions of all
    14C-labelled products of fermentation differed significantly
    ( p < 0.01) between the two groups: in the controls, 84% of the

    radiolabel was present as unchanged erythritol, and carbon dioxide,
    acetic acid, propionic acid, and butyric acid each accounted for < 2%
    of the radiolabel; in contrast, < 1% of the radiolabel in the caecal
    contents of treated rats was present as erythritol at the end of
    incubation, 17% of the administered dose was released as 14C-carbon
    dioxide within 2 h of incubation, and 24% of the radiolabel was
    recovered as 14C-carbon dioxide by the end of the incubation period.
    Succinic, acetic, propionic, and butyric acids were identified as
    fermentation products and accounted for about 60% of the radiolabel at
    the end of the incubation period. Succinic acid was detectable after 1
    h but not subsequently, suggesting that it was fermented to other
    products (Noda & Oku, 1990, 1992).

         Groups of five male Wistar rats were given access to diets
    containing 1, 5, or 10% erythritol for eight days  ad libitum. During
    the second four-day period, feed intake was measured and urine and
    faeces were collected for analysis of erythritol content by gas
    chromatography. Urinary recovery of erythritol accounted for
    approximately 94, 94, and 83% of the administered dose at the three
    concentrations, respectively. Irrespective of dose, approximately 1%
    of the administered dose was recovered in the faeces (Oku & Noda,
    1990).

         In a study in which Wistar rats were fed diets containing 5, 10,
    or 20% erythritol and 20% mannitol for 13 weeks, urine and faeces were
    collected over two days at weeks 5 and 10 in order to quantify
    excretion of erythritol or mannitol by these routes. Total recovery of
    erythritol in the urine and faeces was 55-76% of that ingested and was
    highest at the low dose but comparable at higher concentrations.
    Regardless of dose, only 2-7% was present in the faeces, and most of
    the ingested erythritol was excreted unchanged in the urine. Small
    amounts of erythritol were recovered in the faeces, leaving 25-45% of
    the dose unaccounted for. The authors assumed that the unrecovered
    erythritol was fermented in the large intestine, as < 3% of the
    administered mannitol was recovered unchanged in the urine and faeces
    (Til et al., 1991, 1996).

         Groups of three Wistar rats of each sex were given a single dose
    of 0.1 g/kg bw 14C-erythritol by gavage, as follows: germ-free rats
    were kept under sterile conditions until administration of commercial
    14C-erythritol; adapted conventional rats received diets containing
    weekly increases of 5, 10, and 20% erythritol for three weeks before
    administration of commercial 14C-erythritol; unadapted conventional
    rats were kept on CIVO stock diet before administration of commercial
    14C-erythritol; or germ-free rats were kept under sterile conditions
    until dosing with purified 14C-erythritol. Rats were not fasted
    before dosing. Immediately after treatment, the rats were placed in
    individual metabolism cages to allow collection of expired carbon
    dioxide, urine, and faeces over 24 h. After sacrifice, radiolabel was
    determined in urine, faeces, carbon dioxide absorption solutions,

    plasma, blood cells, the contents of the duodenum, ileum, caecum,
    colon, and the following collected tissues: empty intestinal segments,
    liver, kidneys, lungs, spleen, heart, brain, testes, prostate, uterus,
    ovaries, adrenals, stomach, and urinary bladder, and samples of
    perirenal fat, skin, and skeletal muscle as well as the residual
    carcass and excretion products adhering to fur. The chemical identity
    of the 14C-labelled urinary excretion products was determined in
    samples pooled for each sex in each group at each sampling period by
    HPLC. The presence of volatile metabolites in urinary samples was also
    assessed by lyophilization.

         More than 90% of the administered dose was recovered in urine,
    faeces, expired carbon dioxide, and the bodies of treated rats. About
    50% of the administered radiolabel was excreted in the urine within 4
    h of dosing, and within 24 h of administration the urinary recovery
    represented 65-85% of the administered dose. The greatest urinary
    recovery of radiolabel occurred in the germ-free rats given purified
    14C-erythritol by gavage and the lowest occurred in conventional
    rats adapted to erythritol. Faecal recovery of radiolabel represented
    3-16% of the administered dose. There was wide individual variation in
    the recovery from faeces, and no evidence that sex, adaptation, or
    germ-free status had any effect. Within the gastrointestinal tract,
    the radiolabel was found primarily in the caecum and colon; higher
    concentrations of residual radiolabel were observed in the lower
    gastrointestinal tract of female than male rats and in germ-free than
    conventional rats. Excretion of 14C-carbon dioxide in expired air
    was more extensive in conventional rats of either sex than in
    germ-free rats, accounting for 6-13% of the administered dose in
    conventional rats and < 1% in germ-free rats. A smaller difference
    was seen between adapted and unadapted conventional rats. Conventional
    rats expired considerably more 14C-carbon dioxide, presumably
    because a fraction of the ingested erythritol was fermented by the
    intestinal microflora to metabolites which were absorbed and further
    metabolized; the germ-free rats excreted more unchanged erythritol in
    the urine and faeces. The radiolabel retained in the rats after 24 h
    represented < 3% of the administered dose, and the majority was found
    in the carcass after the skin and organs had been removed. The highest
    concentrations of radiolabel found in the organs were in the livers of
    rats dosed with commercial 14C-erythritol, and most of the urinary
    radiolabel was identified as erythritol. In rats that received
    commercial 14C-erythritol, 2.5-2.9% of the radiolabel was erythrose
    and 0.2-0.35% was glucose. In germ-free rats that received purified
    14C-erythritol, less radiolabel was associated with erythrose. No
    volatile radioactive components were identified by lyophili-zation of
    the urine samples (van Ommen & de Bie, 1990; van Ommen et al., 1996).

         Excretion of erythritol in the urine and faeces of groups of 20
    male rats receiving 2, 5, or 10% erythritol or 10% mannitol in the
    diet was determined over two days at weeks 26, 50, 78, and 102 of
    feeding and at week 42 for the group given 10% of either compound. The
    average recovery of erythritol in the urine represented 76, 63, and

    61% of the administered dose, with 59-118% for rats at 2% erythritol,
    58-73% in those at 5%, and 52-66% in those at 10%. Faecal recovery
    accounted for 2.6% (1.1-4.1), 1.7% (0.3-3.7), and 1.8% (1.4-2.8) of
    the erythritol consumed by males fed the three concentrations,
    respectively. In rats fed 10% mannitol, 3-4% of the dose was recovered
    in urine and < 1% in faeces (Lina et al. 1994, 1996).

         Dogs

         Three beagle dogs were given a single oral dose of 1 g/kg bw
    14C-erythritol after an 18-h fast and were then housed individually
    in metabolism cages and fasted for an additional 8 h. Blood samples
    were collected from each dog 1, 5, 15, and 30 min and 1, 2, 3, 4, 5,
    6, 8, 24, 48, 72, 96, and 120 h after dosing for determination of
    radiolabel in blood and plasma and for calculation of the ratio of
    distribution to erythrocytes. Urine and faeces were collected
    separately for periods of 0-8, 8-24, 24-48, 48-72, and 72-120 h after
    dosing to allow determination of radiolabel. Radiolabel in expired air
    was determined in 7% of the total expired air [sic], collected by
    applying masks. The concentration of erythritol in blood and plasma
    reached a peak 30 min after dosing and decreased in a biphasic manner
    with half-times of 0.07 and 1.7 h in blood and 0.5 and 5.4 h in
    plasma. The portion of radiolabel distributed to blood cells rose from
    23% at 30 min after dosing to about 33% at 24 h. The plasma protein
    binding ratio was 2.3% at 30 min and increased to 83% 24 h after
    administration, when the plasma concentrations were very low. The
    primary route of excretion of radiolabel was in the urine, with about
    44% of the administered dose being excreted in the first 8 h and 94%
    in the first 24 h after dosing; over the next four days, an additional
    2% of the administered dose was recovered in urine. Faecal excretion
    accounted for < 1% of the administered radiolabel, while about 1% was
    recovered in expired air (Nakayama, 1990d; Noda et al., 1996).

         Three beagle dogs were given a single oral dose of 1 g/kg bw
    14C-erythritol after an 18-h fast, housed individually in metabolism
    cages, and fasted for an additional 8 h. Urine was collected up to 24
    h after dosing to allow measurement of radiolabel and identification
    of metabolites. HPLC analysis of urine excreted during the first 24 h
    after treatment indicated that all the radiolabel was associated with
    unchanged erythritol. No additional peaks were observed that indicated
    the presence of metabolites (Noda, 1994).

         Groups of four beagles of each sex were offered diets containing
    2, 5, or 10% erythritol (equal to 0.7, 1.7, and 3.8 g/kg bw per day)
    for 53 weeks. Excretion in the urine and faeces was studied over 24-h
    periods at 13, 26, 38, and 52 weeks. The calculated average urinary
    recovery of erythritol was higher in males (83, 57, and 57%) than
    females (66, 49, and 45% of the administered dose at 2, 5, and 10%).
    Faecal recovery of erythritol accounted for 13, 7.4, and 5.2% of the
    administered dose in males and 9.7, 5.5, and 3.5% in females at the

    three doses, respectively. There was wide individual variation in the
    24-h excretion of erythritol in animals of each sex at each dose at
    week 13 and in females throughout the study, especially in females at
    the high dose in which the urinary recovery ranged from 0.38 to 38
    g/24 h during week 38. Excretion of erythritol as carbon dioxide in
    expired air was not determined (Dean & Jackson, 1992; Dean et al.,
    1996).

         Sheep

         The development of the blood-brain and blood-cerebrospinal fluid
    barriers to lipid-insoluble substances of various sizes was studied in
    fetal sheep early (60 days) and late (125 days) in gestation, with
    radiolabelled erythritol, sucrose, inulin, and albumin. The results
    for erythritol, the smallest molecule tested, differed from those for
    the other markers. The concentrations of erythritol achieved by 6 h at
    both fetal ages were similar, especially in brain, whereas the
    steady-state concentrations of the other molecules decreased between
    60 and 125 days. The concentrations of erythritol were substantially
    higher than would be expected if it had entered the compartments by
    passive diffusion by the same route as the other markers. In addition,
    erythritol took longer than the other markers to reach a steady state.
    The investigators concluded that erythritol probably penetrates cells
    directly across the membranes of brain cells, there is trans-ependymal
    exchange of erythritol from the cerebrospinal fluid to the brain, and
    erythritol penetrates directly across choroid plexus epithelial cell
    membranes into the cerebrospinal fluid and across brain endothelial
    cell membranes into the brain extracellular fluid (Dziegielewska
    et al., 1979).

         Humans

         The absorption and excretion of a single oral dose of erythritol
    was studied in groups of five healthy male volunteers who consumed a
    single dose of 10 or 20 g erythritol as a 20% aqueous solution
    followed by an equal volume of tap water as a wash, for a total of 100
    ml at the low dose and 200 ml at the high dose. Urine samples were
    collected over 0-3, 3-8, and 8-24 h after treatment with both doses
    and over 24-48 h after treatment with 20 g. The urine volume was
    measured and the samples were analysed for erythritol by gas
    chromatography. The volunteers were asked not to consume alcoholic
    beverages, drugs, coffee, or foods containing sweeteners during the 24
    h before dosing or during the urine collection period. No symptoms of
    discomfort were reported by subjects consuming 10 g erythritol, but
    some (number not specified) at the high dose reported a diuretic
    effect as a subjective symptom, and the urine volumes of these men
    were higher than those of men receiving the lower dose during the
    first 24 h after treatment. In both groups, urinary excretion was
    highest during the first 3 h after dosing. The urinary volumes of men
    given the high dose were similar for the 0-24-h and 24-48-h periods.

    The authors concluded that erythritol had no diuretic effect at either
    dose because the urine volumes were within the normal range for adult
    males (1-1.6 L/day) and the urine volumes of men at the high dose were
    comparable during the first and second 24-h periods. In the absence of
    baseline data on the urine volumes of these subjects or on their fluid
    intakes after treatment, and since the men at the high dose received a
    larger volume of fluid with the dose of erythritol, the possible
    diuretic effect of erythritol could not be assessed. At both doses,
    erythritol was rapidly excreted: about 40% was recovered in the urine
    within 3 h and about 90% by 24 h. With 20 g, about 4% of the dose was
    recovered 24-48 h after treatment. If it is assumed that the 10% of
    the erythritol that was not recovered was fermented by the intestinal
    flora, the energy value for erythritol can be estimated to be 0.4
    kcal/g (Noda et al., 1988).

         In studies designed to investigate the metabolism of erythritol
     in vivo in healthy volunteers and to compare the fermentation of
    erythritol by human faecal flora  in vitro with that of glucose and
    lactitol, four male and two female volunteers aged 21-25 undertook an
    overnight fast and were then chosen at random to receive a single dose
    of 25 g 13C-erythritol, 13C-glucose, and 13C-lactitol in 250 ml of
    water with at least three days between each treatment. Breath samples
    were taken for analysis of 13C-carbon dioxide and H2 before
    treatment and at 30-min intervals up to 6 h after treatment. The ratio
    of 13C:12C-carbon dioxide was measured by isotope-ratio mass
    spectrometry. Urine samples were collected over 0-6 and 6-24 h after
    treatment, and the erythritol and lactitol concentrations in urine
    were measured by HPLC. In order to maintain a constant metabolic rate,
    the subjects remained at rest during the study. For the assay of
    fermentation  in vitro, faecal samples were collected from six
    healthy volunteers (sex and age not specified) who ate a normal
    western diet. None of the subjects complained of gastrointestinal
    symptomsand none had used antibiotics in the past six months. The
    samples were incubated under anaerobic conditions for 6 h, and then
    the H2 concentration was measured in the head-space of the incubation
    vials.

         During the first 6 and 24 h after dosing, 52 and 84%,
    respectively, of the administered erythritol was recovered in the
    urine. No increase in expired 13C-carbon dioxide or H2 was observed,
    indicating that no fermentation had occurred in the gut. In contrast,
    there was a rapid increase in expired 13C-carbon dioxide after
    consumption of glucose and a more gradual rise after ingestion of
    lactitol. Excretion of H2 in expired air was also increased after
    treatment with lactitol. Neither lactitol nor glucose was detected in
    significant amounts in the urine. After a 6-h incubation with
    erythritol, the amount of H2 formed by the faecal flora was
    comparable to that in control vials, but significantly ( p < 0.001)
    more H2 was produced in the glucose and lactitol vials than in either
    control or erythritol vials (Hiele et al., 1993).

         After a 12-h fast, five men received a single oral dose of 0.3
    g/kg bw erythritol as a 20% aqueous solution. Urine samples were
    collected over 48 h and blood samples over 24 h to allow determination
    of erythritol. The compound appeared to be readily absorbed, the
    plasma concentration peaking at 430 g/ml 30 min after treatment, with
    a half-time of 3.4 h. Erythritol was most rapidly excreted in the
    urine during the first 3 h after treatment, and about 90% of the
    administered dose was recovered unchanged in the urine within 48 h.
    Since at most only about 10% of the erythritol was available for
    respiration, the authors concluded that the available energy of
    erythritol was < 10% that of digestible carbohydrates: 0.4 kcal/g or
    1.7 kJ/g (Noda et al., 1994).

         The kinetics of erythritol in plasma and urine were investigated
    in three men and three women after an overnight fast. Each subject
    ingested a single oral dose of 1 g/kg bw dissolved in 250 ml of water,
    and blood samples were taken 5, 10, 15, 30, 45, 60, 90, 120, 180, and
    240 min after dosing for determination of plasma concentrations of
    erythritol. Plasma creatinine concentrations were determined in a
    blood sample taken before treatment. Urine was collected over 0-30
    min, 30-60 min, 1-2 h, 2-3 h, and 3-24 h after treatment for
    determination of the volume and of the erythritol and creatinine
    concentrations. Erythritol was detected in the plasma 10 min after
    dosing, and the mean plasma concentrations increased steadily from 15
    min after treatment to a peak of 2.2 mg/ml after 90 min. The urine
    volume and erythritol concentration reached a maximum during 1-2 h
    after ingestion, at about the same time that the plasma concentration
    of erythritol peaked. Urinary recovery of erythritol over the 24-h
    collection period accounted for 78% of the administered dose, with 30%
    collected after 3 h. During 1 and 2 h after administration, the
    clearance of erythritol was about half that of creatinine, indicating
    tubular reabsorption of erythritol by the kidney (Bornet et al.,
    1996a).

         After 12 male and 12 female volunteers received a dose of 0.4 or
    0.8 g/kg bw erythritol in a chocolate snack, the plasma erythritol
    concentrations increased rapidly, reaching peaks of 3 and 5 mmol/L 1
    and 2 h after dosing with 0.4 and 0.8 g/kg bw, respectively. Starting
    2 h after treatment, the plasma erythritol concentrations were
    significantly ( p < 0.05) higher in the group given 0.8 g/kg bw than
    in that given 0.4 g/kg bw. At both doses, erythritol appeared in the
    urine within 2 h of dosing, the largest quantities being collected
    between 2 and 4 h after administration. Erythritol was still present
    in urine 22 h after treatment. The concentration of erythritol in the
    urine of individuals given 0.8 g/kg bw was about twice and
    significantly ( p < 0.05) greater than that in the urine of people
    given 0.4 g/kg bw. On average, 61 and 62% of the administered
    erythritol was recovered in the urine after the 0.4 and 0.8 g/kg bw
    doses, respectively, within 22 h (Bornet et al., 1996b).

         The peak serum concentration of erythritol in five
    non-insulin-dependent diabetic patients (sex not indicated) who
    consumed a single dose of 20 g erythritol in solution occurred 1 h
    after administration and was 650 ± 37 µg/ml. On average, 82, 88, and
    88% of the administered erythritol was recovered in the urine 24, 48,
    and 72 h after dosing, respectively (Ishikawa et al., 1992, 1996).

         In 12 male subjects who consumed 1 g/kg bw per day erythritol in
    a variety of foods during a five-day test period under controlled
    conditions, the mean urinary excretion was 61-88% of the nominal
    ingested dose, with an average of 78% (Tetzloff et al., 1996).

    2.2  Toxicological studies

    2.2.1  Acute toxicity

         Studies of the toxicity of single doses of erythritol are
    summarized in Table 1.

        Table 1. Studies of the acute toxicity of erythritol
                                                                                                  

    Species      Sex         Route                    LD50             Reference
                                                      (mg/kg bw)
                                                                                                  

    Mouse        NR          Intraperitoneal            7000-9000      Beck et al. (1936)
    Mouse        NR          Intraperitoneal    approx. 8000-9000      Beck et al. (1938)
    Rat          NR          Oral                     > 18 000         Beck et al. (1938)
    Rat          Male        Intravenous                 6 600         Yamamoto et al. (1987)
                 Female                                  9 600
    Rat          Male        Subcutaneous             > 16 000         Yamamoto et al. (1987)
                 Female                               > 16 000
    Rat          Male        Oral                       13 100         Yamamoto et al. (1987)
                 Female                                 13 500
    Dog          Male        Oral                      > 5 000         Ozeki et al. (1988)
                                                                                                  

    NR, not reported
    
    2.2.2  Short-term studies of toxicity

         Mice

         Groups of 10 CD-1 mice of each sex were fed diets containing 0,
    5, 10, or 20% erythritol (purity, 99.9%; equivalent to 0, 7.5, 15, and
    30 g/kg bw per day)  ad libitum for up to 13 weeks. In order to adapt
    the mice to the high dose of erythritol, the dose was increased from
    5% on days 1-3 to 10% on days 4-7 and finally to 20% from day 7. The

    mice were observed at least once daily for clinical signs of toxicity,
    and individual body weights and feed consumption were measured weekly
    for each cage throughout the study. Water consumption for each cage
    was measured over four-day periods during weeks 4, 8, and 12. During
    weeks 10 and 11, respectively, male and female mice were transferred
    to individual metabolism cages for collection of urine and faeces over
    two days. The urinary volume and pH were determined after
    centrifugation, and urine samples were analysed for density,
    osmolality, protein, gamma-glutamyl-transferase (GGT),  N-acetyl
    glucosaminidase (NAG), sodium, potassium, calcium, phosphate, citrate,
    creatinine, and erythritol. The 48-h faecal samples were analysed for
    unabsorbed erythritol. Blood samples were obtained from all mice on
    day 91 to allow determination of the glucose concentration and
    haematological parameters. Before sacrifice on day 95 for males and
    day 96 for females, blood samples were collected for determination of
    clinical biochemical parameters. After sacrifice, the mice were
    examined grossly for pathological changes, and the weights of the
    brain, caecum, heart, kidneys, liver, spleen, and testes were
    recorded. Tissues from the livers and kidneys of all animals and
    tissues of the adrenal, caecum, and spleen of all control and
    high-dose mice were subjected to detailed histopathological
    examination.

         There were no unscheduled deaths during the study. Even at the
    concentration of 20% of the diet, diarrhoea was not observed as an
    effect of treatment. The mean body weight of males fed 20% erythritol
    was significantly ( p < 0.01) lower than that of controls from day
    14 to the end of the study. Although the body weights of the other
    treated animals tended to be lower than those of controls, the
    difference was not statistically significant. The feed intake of mice
    receiving erythritol was comparable to or higher than that of
    controls, but when the treated diets were corrected for their
    erythritol content the intakes of treated mice tended to be comparable
    to or lower than those of controls. Both males and females fed
    erythritol showed a dose-related increase in water intake, the
    increase being greater in males and reaching statistical significance
    ( p < 0.01) at the 10 and 20% doses throughout the study. The most
    notable increases in water intake by females were seen at the same
    doses, but the sample size was too small to permit statistical
    analysis because of group caging. There were no treatment-related
    changes in erythrocyte or leukocyte parameters, clotting potential, or
    clinical biochemistry. There was a significant ( p < 0.01)
    dose-related increase in the urine volume of mice of each sex at 10
    and 20%. The density and osmolality of the urine of males and females
    fed 5% erythritol were greater than those of controls but lower than
    those of controls at the 20% dietary concentration. There were
    dose-related increases in the 24-h excretion of protein, GGT and NAG
    activities, and sodium, potassium, calcium, phosphate, citrate, and
    creatinine concentrations in animals of each sex, which were
    statistically significant at the high dose. The increased excretion of
    protein and GGT activity were also significant in animals of each sex

    at the intermediate dose, and the increase in phosphate excretion was
    significant in females at this dose. The authors speculated that the
    increased 24-h excretion of calcium was a result of increased calcium
    absorption, an effect of fermentation in the large intestine reported
    for polyols. Both males and females had dose-related increases in the
    absolute and relative weights of the kidney which was significant at
    20%. [The enlarged kidneys may represent an adaptive response to the
    diuretic properties of erythritol.] Dose-related increases were seen
    in the absolute and relative weights of the full and empty caecum,
    which was statistically significant at the 20% dose for relative
    weights in males and for both relative and absolute weights in
    females. The incidences of macroscopic and microscopic findings did
    not differ among groups. 

         The effects of erythritol in this study were considered to be
    increased water intake, diuresis, and increased urinary excretion of
    protein and marker enzymes in males and females at 10 and 20% in the
    diet. The concentration of 20% was also associated with increased
    kidney weights in females, decreased body weights of males, and
    increased caecal weights in animals of each sex. Since these effects
    were physiological or adaptive, they were not considered to be toxic
    effects. Consequently, the NOEL was 5% erythritol, equivalent to 7.5
    g/kg bw per day (Til et al., 1992, 1996).

         Rats

         Groups of six male Wistar rats were given access to diets
    containing 0, 5, or 10% erythritol (purity not specified; equivalent
    to 0, 5, and 10 g/kg bw per day), 10% glycerol, or 10% sorbitol
     ad libitum for 28 days. Body weights were recorded every two days.
    Before the end of the study, all rats were transferred to individual
    metabolism cages to allow measurement of feed and water intake, feed
    efficiency, and urinary and faecal excretion over the next four days.
    They were sacrificed after an 18-h fast, and a blood sample was
    collected for determination of total protein, glucose, total
    cholesterol, and triglycerides in serum. The liver, kidney, small
    intestine, caecum, and colon were weighed, and a homogenate was
    prepared from the intestinal mucosa for determination of the
    activities of the digestive enzymes sucrase, maltase, isomaltase,
    alkaline phosphatase, and leucine aminopeptidase.

         Rats fed 10% erythritol had diarrhoea during the first three days
    of the study but not thereafter, but three rats fed 10% sorbitol had
    to be removed from the study because of severe diarrhoea and 'very
    poor body weight gain' and the other three rats in this group were
    reported to have diarrhoea throughout the first two weeks of the
    study. After 28 days on test, the mean body weights of rats at 5 and
    10% erythritol and 10% glycerol tended to be higher than those of
    controls but the difference was not statistically significant, while
    the mean body weight of the rats at 10% sorbitol was lower than that

    of controls. Feed intake and feed efficiency during the last four days
    of the study were comparable in all groups. The mean water intake of
    the group given 10% erythritol was significantly ( p < 0.01) greater
    than that of controls. There was a dose-related, statistically
    significant ( p < 0.05) increase in the urine volume of rats
    receiving 5 and 10% dietary erythritol, with volumes of 3.1, 7.8, and  
    18 ml/day in groups receiving 0, 5, and 10% erythritol, respectively.
    The urine volume was not affected by 10% dietary glycerol or sorbitol.
    Daily faecal excretion was comparable in all groups. Only the relative
    organ weights were reported; those of the liver, kidney, small
    intestine, and colon were comparable at all doses of erythritol, while
    the caecal weight of rats at 10% erythritol was significantly
    increased over that of controls. Dietary sorbitol significantly
    increased the relative weights of the small intestine, caecum, and
    colon, while glycerol had no effect. Erythritol had no effect on
    intestinal enzyme activity, whereas sorbitol significantly
    ( p < 0.05) decreased the activities of sucrase, isomaltase,
    alkaline phosphatase, and leucine amino peptidase. A dose-related
    decrease in serum triglycerides was associated with erythritol
    treatment which was statistically significant at the 10% dose, with
    concentrations of 190, 170, and 140 mg/dL at 0, 5, and 10% erythritol.
    This parameter was not affected by treatment with 10% glycerol or
    sorbitol.

         Diuresis was observed at both 5 and 10% dietary erythritol, and
    the 10% concentration was also associated with transient diarrhoea,
    increased water intake, increased caecal weight, and decreased serum
    triglyceride concen-tration. The LOEL was 5% of the diet, equivalent
    to 5 g/kg bw per day (Oku & Noda, 1990).

         Groups of 10 Wistar rats of each sex were given access to diets
    containing 0, 5, or 10% erythritol (purity, 98.5%), equal to 5.4 and
    11 g/kg bw per day in males and 5 and 9.9 g/kg bw per day in females,
    respectively,  ad libitum for 28 days. The animals were observed at
    least once daily for signs of toxicity, and feed and water consumption
    per cage and individual body weights were measured weekly.
    Ophthalmoscopic examinations were made on all rats before the start
    and at the end of the study. On day 24, all rats were placed in
    individual metabolism cages and deprived of water for 24 h and of feed
    for the final 16 h of this period, during which time urine was
    collected from individual animals. Blood samples were collected at the
    end of the fasting period and analysed for glucose and haematological
    end-points. All males were killed on day 28 and all females on day 29,
    and blood was collected for clinical chemistry. The animals were then
    examined macroscopically for pathological changes, and the weights of
    the adrenals, brain, caecum, heart, kidneys, liver, lungs, ovaries,
    pituitary, prostate, seminal vesicles, spleen, sublingual salivary
    glands, submaxillary salivary glands, testes, thymus, thyroid, and
    uterus were recorded. Samples of these and 25 tissues and organs were
    processed for histopatho-logical examination.

         Diarrhoea and/or soft stools were reported in all cages
    containing males receiving 10% erythritol, one cage of females
    receiving 5% erythritol, and all cages of females receiving 10%
    erythritol during the first 11 days of the study. There were no
    unscheduled deaths. The feed consumption of males at the high dose
    tended to be lower than that of controls throughout the study, but the
    difference was significant only during the first week. Feed efficiency
    was not significantly affected by erythritol treatment, and the
    reduction in body weight was probably secondary to the lower energy
    density of this diet and the reduced feed consumption of these males.
    The water consumption of animals of each sex at the high dose tended
    to be greater than that of controls throughout the study. Dietary
    erythritol resulted in a dose-related increase in plasma alkaline
    phosphatase activity, the increase reaching statistical significance
    at 10%. There were no other toxicologically significant changes in
    haematological or clinical chemical parameters. No impairment of renal
    function (as measured from urine volume and density) was seen in
    erythritol-treated rats. Although the concentrations of urinary
    electrolytes were lower in erythritol-treated females, significantly
    for sodium and potassium, the total excretion of electrolytes over the
    16-h collection period was not significantly affected. In males,
    erythritol treatment resulted in a dose-related increase in the
    absolute and relative kidney weights which was statistically
    significant at both doses. The same trend was observed in females but
    the differences were not statistically significant. Both males and
    females showed a dose-related increase in the absolute and relative
    weights of the empty or full caecum, which was significant in males at
    the high dose and in females at both doses. Female rats had
    dose-related increases in absolute and relative spleen weights that
    were statistically significant ( p < 0.05) at 10%: 0.37, 0.42, and
    0.43 g/kg bw and 2.1, 2.3, and 2.4 g/kg bw at 0, 5, and 10%,
    respectively. No treatment-related effects were observed during
    macroscopic examination of tissues and organs. Microscopic examination
    revealed no histological changes that could be attributed to
    erythritol, and no histopathological findings were associated with the
    increased kidney, caecum, or spleen weights. 

         In this study, increased kidney weights in males and increased
    caecal and spleen weights in females were seen at both the 5 and 10%
    doses. At 10%, the kidney weights of females, the caecal weights of
    males, and plasma alkaline phosphatase activity in both males and
    females were increased. The LOEL was 5% erythritol, equal to 5 g/kg bw
    per day (Til & Wijnands, 1991; Til & Modderman, 1996).

         Erythritol (purity, 99.7%) was administered to groups of 12
    Slr:Wistar rats of each sex by gavage at doses of 0, 1, 2, 4, or 8
    g/kg bw per day for 13 weeks. Additional groups of six rats of each
    sex received erythritol by gavage at doses of 0, 4, or 8 g/kg bw per
    day for 13 weeks and were then untreated for four weeks. The animals
    were observed for clinical signs and deaths twice daily during the

    treatment period and once daily during the recovery period. Body
    weights were recorded weekly during both periods. Daily feed and water
    intakes were calculated during each week of the study. Sensory
    function, including pupil reflex, corneal reflex, pain response,
    auditory response, and righting reflex, were assessed, and
    ophthalmoscopic examinations were conducted before treatment and on
    the day before autopsy. The animals were transferred to metabolism
    cages the day before scheduled sacrifice and fasted for about 18 h.
    Urine samples were collected during this period, and urine volume,
    colour, specific gravity, osmotic pressure, pH, protein, glucose,
    ketone bodies, bilirubin, occult blood, nitrite, urobilinogen, sodium,
    potassium, chlorine, and sediments, but not calcium, were determined.
    Blood samples were collected before necropsy for haematological and
    clinical biochemical examinations. The weights of the following organs
    were recorded: salivary glands, thymus, lungs and bronchi, heart,
    thyroid and parathyroid, liver, spleen, kidneys, adrenals, seminal
    vesicles, testes, prostate, ovaries, uterus and vagina, brain, and
    pituitary, and a further 23 organs and tissues were preserved for
    microscopic examination. Small sections of the liver and kidneys were
    prepared for electron microscopy.

         Diarrhoea and soft stools were noted in almost all animals at 4
    g/kg bw per day during the first week of the study and in those at 8
    g/kg bw per day during the first two weeks of treatment. In addition,
    both males and females receiving 8 g/kg bw per day were reported to
    have reduced spontaneous movement. These symptoms were observed
    sporadically in these two groups for the remainder of the treatment,
    and none were reported during the recovery period. No
    treatment-related deaths occurred. Erythritol had no significant
    effect on the body weights of female rats, but the body weights of
    males at 8 g/kg bw were significantly ( p < 0.01) reduced from week
    7 to the end of the treatment period and remained significantly
    reduced until the end of the first week of the recovery period.
    Throughout the study, male and female rats receiving 8 g/kg bw per day
    tended to consume less feed than controls, although feed consumption
    was comparable for all groups during the recovery period. The water
    intakes of male and female rats at 8 g/kg bw per day were greater than
    that of controls, and the difference was statistically significant
    throughout the treatment period for females and for some of the
    treatment period for males. The water intakes returned to control
    levels during the recovery period. No treatment-related effects were
    observed in tests of sensory function, in the eyes, or in
    haematological parameters. A small dose-related decrease in plasma
    sodium concentration was noted, which reached statistical significance
    at 4 and 8 g/kg bw per day in males and at 8 g/kg bw per day in
    females. There was also a small dose-related decrease in plasma
    chlorine concentration, but only that of males receiving 8 g/kg bw per
    day differed significantly from that of controls. A dose-related
    increase in serum alkaline phosphatase activity was observed in
    animals of each sex, without statistical significance. Both males and
    females showed a dose-related increase in plasma blood urea nitrogen,

    the increase reaching statistical significance at 4 and 8 g/kg bw per
    day in females and at 8 g/kg bw per day in males. The proportion of
    total protein present as alpha-1 globulin was significantly lower in
    males and females receiving 8 g/kg bw per day. At the end of the
    four-week recovery period, all the clinical biochemical parameters in
    erythritol-treated rats were comparable to those of controls.
    Erythritol caused a dose-related increase in urine volume, which
    reached statistical significance at 8 g/kg bw per day. The specific
    gravity and osmotic pressure of the urine were increased in
    erythritol-treated rats when compared with controls, but the
    differences were statistically significant only in males, with the
    greatest increases in rats at the intermediate dose and decreasing at
    the higher dose, suggesting competing processes (increased solute
    excretion and increased diuresis). A dose-related increase in urinary
    excretion of sodium and chlorine reached statistical significance in
    rats of each sex at 4 and 8 g/kg bw per day. Urinary potassium
    excretion by male and female rats receiving 8 g/kg bw per day was
    significantly higher than in controls. Urinary parameters were similar
    in all groups after the four-week recovery period. Gross examination
    revealed hypertrophy of the caecum in 7/12 males and 6/11 females at 8
    g/kg bw per day but in none of the other groups. Both males and
    females had a dose-related increase in relative kidney weight that
    reached statistical significance in males at 4 and 8 g/kg bw per day
    and in females at 8 g/kg bw per day; the absolute kidney weights
    tended to be higher in erythritol-treated rats although none of the
    differences was statistically significant. Calcium deposits were
    observed in the kidneys of a few treated females at all doses and in
    males at 4 and 8 g/kg bw per day. No calcium deposits were observed in
    control rats of either sex. Dilatation of the renal tubules occurred
    more frequently at 8 g/kg bw per day, affecting more than 50% of the
    animals. Rats of each sex showed dose-related increases in absolute
    and relative adrenal weights which were significant in males at 8 g/kg
    bw per day and in females for relative weights only. Males at 8 g/kg
    bw per day also had significantly lower absolute and relative thymus
    weights than controls. A similar trend was observed in female rats at
    the high dose, but the differences were not statistically significant.
    The investigators attributed the increased adrenal and decreased
    thymus weights to treatment-associated stress. At the end of the
    treatment period, an increased incidence of sinusoid dilatation in the
    adrenal cortex was seen in males at 4 and 8 g/kg bw per day but not in
    females. At the end of the four-week recovery period, only the
    increased relative kidney weights of males at 8 g/kg bw per day
    remained. The incidences of sinusoid dilatation of the adrenals and
    dilatation of the renal tubules were comparable in all groups.

         The NOEL was 2 g/kg bw per day on the basis of diarrhoea and/or
    soft stools and increased urinary excretion of sodium and chlorine in
    animals of each sex, increased blood urea nitrogen in females, and
    decreased plasma sodium, increased kidney weights, and increased
    sinusoid dilatation of adrenal glands in males at 4 g/kg bw per day.
    The dose of 8 g/kg bw per day was associated with diarrhoea and

    decreased spontaneous movement, increased water consumption, decreased
    plasma sodium and chlorine concentrations, increased blood urea
    nitrogen, increased urine volume and urinary excretion of sodium,
    potassium, and chlorine, increased kidney and adrenal weights, and
    dilatation of the renal tubules. All of these effects except the
    increased kidney weights in males were reversible within the recovery
    period (Yamamoto et al., 1989).

         Groups of 15 male Wistar rats were given access to diets
    containing 0, 5, 10, or 20% erythritol or 20% mannitol (purity not
    indicated for either compound; equivalent to 0, 2.5, 5, and 10 g/kg bw
    per day)  ad libitum for 13 weeks. An additional group of 15 rats was
    given access to the 20% erythritol diet for only 6 h each day. The
    rats were observed for clinical signs of toxicity and deaths at least
    once daily. Individual body weights were recorded twice a week
    throughout the study; feed consumption was determined weekly for each
    cage, and the daily intake of each rat was calculated. During weeks 5
    and 10, all the surviving rats were transferred to individual
    metabolism cages, and urine and faeces were collected separately for
    two days. Urine samples were assessed for volume, pH, density,
    osmolality, protein, low-molecular-mass proteins, GGT and NAG
    activity, sodium, potassium, calcium, phosphate, citrate, creatinine,
    erythritol, and mannitol. The 48-h faecal samples were assessed for
    unabsorbed erythritol. On day 79, all surviving rats given 0, 20% ad
    libitum, and restricted diet were kept in metabolism cages for
    separate collection of urine and faeces and were deprived of water for
    24 h and of feed for the last 16 h of this period, when urine was
    collected for determination of volume and density. Blood samples were
    obtained from surviving rats on day 86 for determination of
    haematological parameters. At sacrifice on days 91, 92, and 93, blood
    samples were collected from surviving rats for clinical chemistry
    (excluding blood urea nitrogen). The rats were then examined grossly
    for pathological changes and the weights of the adrenals, brain,
    caecum (full and empty), femur, heart, kidneys, liver, pituitary,
    spleen, testes, thymus, and thyroid were recorded. A detailed
    histopathological examination was made of the liver, kidneys, urinary
    bladder, duodenum, caecum, and brain of all rats in the control, 20%
    erythritol, and 20% mannitol groups. In addition, tissues from the
    spleen and stomach of all rats were examined microscopically because
    changes were observed in these organs at the 20% dose.

         Soft stools and sometimes diarrhoea were observed in rats fed 20%
    erythritol and more frequently in those fed 20% mannitol. There were
    no unscheduled deaths due to treatment. The mean body weights of rats
    given 20%  ad libitum, restricted diet, and 20% mannitol were
    significantly ( p < 0.05) lower than those of controls from day 14
    and throughout the rest of the study. Rats given the restricted diet
    had lower body weights than rats fed the same diet  ad libitum. At a
    concentration of 20%, erythritol had a more severe effect on body
    weight than mannitol. Rats receiving dietary erythritol consumed
    significantly ( p < 0.05) more feed than controls on a per-cage

    basis throughout the study, but nutrient intake was comparable for
    those fed  ad libitum and lower for those given restricted diet when
    the intakes were corrected for erythritol content. As the feed
    consumption of rats given 20% mannitol was comparable to that of
    controls throughout the study, whether caged by group or individually,
    they consumed fewer nutrients. Erythritol caused a dose-related
    increase in the mean water intake of rats fed  ad libitum which was
    significant ( p < 0.05) at the doses of 10 and 20%. Rats on the
    restricted diet consumed more water than controls but less than the
    group given 20% erythritol  ad libitum. Rats fed 20% mannitol tended to
    consume more water than the controls but the difference was not
    statistically significant. Some minor changes were noted in a few
    haematological parameters which were not considered to be
    toxicologically significant.

         At the end of the study, the plasma alkaline phosphatase
    activities were higher in the groups fed 10 or 20% erythritol or 20%
    mannitol, but only those of animals on the restricted diet or 20%
    mannitol differed significantly ( p < 0.05) from those of controls.
    Erythritol treatment was consistently associated with a dose-related
    increase in urine volume that became statistically significant at the
    10% dose. Both urine density and osmolality tended first to increase,
    then decrease with increasing dietary erythritol, although
    significantly only for animals on the restricted diet. Although the
    urine volume of rats given 20% mannitol was not increased, the
    osmolality tended to be lower. The urinary pH was comparable in all
    groups during week 6 and slightly but significantly higher in all
    erythritol-treated rats during week 10. All the values for density,
    osmolality, and pH lay within the intervals reported in the
    literature, suggesting that the effects observed in the current study
    were physiological responses and not toxicologically significant.
    Urinary calcium excretion was increased with increasing doses of
    dietary erythritol, and the concentrations in the urine of rats fed
    20%  ad libitum and restricted diet were significantly higher than
    those of controls. The 24-h urinary excretion of GGT, citrate, sodium,
    and phosphate of rats at 5% erythritol tended to be higher than that
    of controls and then decreased with increasing concentrations of
    dietary erythritol to concentrations comparable to or lower than those
    of controls. A similar trend was observed for urinary creatinine and
    potassium excretion during week 10. Rats fed erythritol  ad libitum
    showed a dose-related increase in NAG activity which was statistically
    significant only during week 10. Feeding of 20% mannitol was
    associated with increased citrate excretion and decreased sodium and
    phosphate excretion but not with increased calcium excretion. In the
    urine concentration test, conducted on day 80, the urine volume was
    increased from 2.8 ml/rat per 16 h in controls to 4.9 ml/rat given 20%
     ad libitum and 4.5 ml/rat given restricted diet. Although the urine
    density tended to be lower in the groups given 20% erythritol, the
    differences were not statistically significant.

         Dietary administration of erythritol resulted in a dose-related
    increase in the absolute and relative weights of the full or empty
    caecum, which reached statistical significance at 10% and 20%
    erythritol. Feeding of 20% mannitol resulted in greater increases in
    caecal weights than those observed with a comparable amount of
    erythritol. The pH of the caecal contents decreased with increasing
    dietary concentration of erythritol, the increase reaching statistical
    significance at concentrations of 10% and greater. The lowest caecal
    pH was observed in the rats fed 20% mannitol. During gross
    examination, dilatation of the caecum was observed in 1/14 rats given
    20%  ad libitum and in 9/15 given 20% mannitol. The relative kidney
    weights were significantly higher than those of controls in the groups
    receiving 10 or 20% erythritol or 20% mannitol, and the absolute
    kidney weights tended to be higher in rats receiving 5 or 10% but not
    20% erythritol, without statistical significance. The absolute and
    relative weights of the bladders of erythritol-treated rats tended to
    increase, but only the relative weights of the kidneys of the rats on
    restricted diet differed significantly from those of controls. The
    effects on the weights of kidney and bladder were minimal. When the
    kidneys were examined grossly, bilateral hydronephrosis was observed
    in one rat given 20%  ad libitum, and unilateral hydronephrosis was
    observed in one rat given 10%  ad libitum and in one on restricted
    diet. Microscopic examination confirmed the hydronephrosis in the one
    rat on the restricted diet, and pelvic dilatation without accompanying
    nephrotic changes was diagnosed in the one rat given 20%  ad libitum.
    The kidneys of rats at 10% erthyritol were not examined
    microscopically.

         The NOEL was 5% erythritol, equivalent to 2.5 g/kg bw per day, on
    the basis of diuresis and increased water intake, urinary NAG and
    serum alkaline phosphatase activity, and caecal weights at 10%
    erythritol (Til et al., 1991, 1996).

         Groups of 22 Slc:Wistar rats of each sex received erythritol
    (19.7% solution; purity not reported) at doses of 0, 1, 1.73, or 3
    g/kg bw per day by intravenous injection for 180 days. Six additional
    rats in each group except that receiving 1 g/kg bw per day were
    allowed to recover without treatment for one month. Control animals
    received physiological saline at a volume corresponding to that
    received by animals at the high dose. The rats were observed for
    deaths and clinical signs of toxicity at least once a day during both
    treatment and recovery. Individual body weights and feed and water
    consumption were recorded weekly for each cage throughout the
    treatment and recovery periods. Ophthalmoscopic examinations were
    performed on six rats of each sex per group after three and six months
    of treatment and on all rats during the last week of recovery. After
    three and six months of treatment, urine was collected from six rats
    of each sex per group in metabolism cages. The frequency of urination
    was recorded during the first hour after dosing, and samples collected
    during the first 3 h and during 3-24 h after dosing were analysed; the
    latter sample was analysed for specific gravity, volume, and sodium,
    chlorine, and potassium concentrations. Before necropsy at the end of

    treatment or recovery, blood samples were obtained for assessment of
    haematological and biochemi-cal parameters. All animals were examined
    externally and then sacrificed, their organs and tissues were examined
    grossly, and the weights of the brain, pituitary, submaxillary and
    sublingual salivary glands, thymus, thyroids, heart, lungs, liver,
    kidneys, spleen, adrenals, testes, epididymides, seminal vesicles,
    prostate, ovaries, and uterus were recorded. Another 27 organs or
    tissues were preserved for histopathological examination. The livers
    and kidneys of two rats of each sex per group were processed and
    stained for electron microscopy. 

         No unusual clinical observations or deaths related to treatment
    were found during the study. The mean body weights of males at the
    high dose were lower than those of controls during most of the
    treatment period through to the end of recovery, and at the end of
    treatment the difference from controls was approximately 6%. The mean
    body weights of males at the intermediate dose and all treated females
    were comparable to those of controls throughout treatment and
    recovery, while males at 1 g/kg bw per day weighed significantly more
    than controls during the last two weeks of treatment. Average feed
    consumption during treatment was unaffected, but the treatment
    resulted in dose-related increases in the water intake of animals of
    each sex at all doses, which were more often statistically significant
    ( p < 0.05) in males than in females and increased with dose. As the
    water intakes of controls tended to be higher during recovery than
    during treatment, the water intake of treated rats did not change
    during the recovery period. Erythritol treatment resulted in a
    dose-related decrease in the erythrocte and haemoglobin counts and in
    the haematocrits of animals of each sex, which were statistically
    significant in males at all doses and in females at the intermediate
    and high doses, although these parameters remained within the
    reference ranges reported in the literature. The number of
    reticulocytes was reported to be significantly higher in females
    receiving 1.73 or 3 g/kg bw per day and in males receiving 3 g/kg bw
    per day than in controls. At the end of the recovery period, the
    haemato-logical parameters in treated animals were comparable to those
    in their respective controls. Females showed a dose-related increase
    in serum blood urea nitrogen that was significant at the high dose.
    Significantly lower serum sodium and calcium concentrations were seen
    in all treated males than in controls, and males at the high dose also
    had significantly higher serum potassium concentrations. The lowered
    serum sodium may simply reflect increases in that of controls
    resulting from the infusion of physiological saline. At three and six
    months, the urine volumes of animals of each sex at 1.73 and 3 g/kg bw
    per day were significantly higher than in controls, the increase in
    urination being evident in rats at the high dose within the first hour
    of treatment. The urine density decreased with increasing dose of
    erythritol but at neither time were the decreases statistically
    significant. Excretion of sodium and chlorine at 21 h was
    significantly lower than in controls in erythritol-treated males at
    three and six months except for males at the low dose at three months,
    but this observation was considered not to be toxicologically

    significant since the controls excreted increased amounts of sodium
    and chlorine during the treatment period in response to the injection
    of physiological saline. Furthermore, erythritol had no significant
    effect on urinary potassium excretion by males or on sodium,
    potassium, or chlorine excretion by females during exposure. The
    urinary pH of females at the high dose was significantly decreased
    ( p < 0.05), although the values remained within the reference
    ranges. After the one-month recovery period, urinary parameters in
    treated males were comparable to those in controls, but
    erythritol-treated females continued to excrete significantly more
    urine of a significantly lower density. Urinary electrolyte excretion
    was comparable in all female groups at the end of the recovery period.
    Treated rats of each sex had dose-related increases in the absolute
    and relative weights of the kidneys, which were statistically
    significant at 1.73 g/kg bw per day. No difference in absolute or
    relative kidney weights was observed at the end of recovery. The
    absolute and relative weights of the thymus were lower in males at the
    high dose but only the absolute weight differed significantly
    ( p < 0.05) from that of controls. The absolute and relative weights
    of the left and right adrenal glands were higher in rats of each sex
    at the high dose than in their respective controls, but only the
    relative weights differed significantly from those of controls.
    Consistent with the known mode of action of erythritol, no increase in
    caecal weights was reported in this study. No abnormalities were
    observed during histopathological or electron microscopic examination
    conducted at the end of the treatment and recovery periods.

         The NOEL was 1 g/kg bw per day on the basis of increased water
    intake, urine production, reticulocyte count, and adrenal and kidney
    weights in animals of each sex, decreased body weight and serum
    potassium in males, and increased blood urea nitrogen in females at
    higher doses. At the end of the one-month recovery period, the body
    weights of males at the high dose remained depressed, and diuresis was
    still present in females at the intermediate and high doses (Kamata,
    1990a).

         Dogs

         Groups of three beagles of each sex were given erythritol
    (purity, 99.7%) by gavage at doses of 0, 1.25, 2.5, or 5 g/kg bw per
    day, on seven days per week for 13 weeks. An additional two animals of
    each sex receiving the 0, 2.5, or 5 g/kg bw per day were treated for
    13 weeks and then allowed to recover without treatment for four weeks.
    The animals were observed for general condition 1, 4, and 24 h after
    dosing and during treatment and once daily during recovery. The
    respiratory rate, heart rate, pupil size, body temperature, body
    weights, and feed and water consumption were measured weekly
    throughout treatment and recovery. Electrocardiograms and
    ophthalmoscopic examinations were conducted on all animals two weeks
    before the beginning of the study, before the first dose, and during
    weeks 7 and 13 of treatment and week 4 of the recovery period. Blood
    samples were obtained and tests for liver and renal function with

    indocyanine green and phenolsulfonphthalein were performed after
    fasting on all animals three weeks before the beginning of the study,
    before the first dose, and during weeks 7 and 13 of treatment and week
    4 of the recovery period. Samples of 24-h urine were collected from
    non-fasted animals during weeks 3 and 1 before treatment, during weeks
    7 and 13 of treatment immediately after dosing, and at week 4 after
    treatment. Samples of urine collected during the first 4 h and during
    4-24 h were analysed for sodium, potassium, and chlorine but not
    calcium. Faecal samples were collected at the same time as the 4-h
    urine samples and tested for the presence of occult blood. At the end
    of treatment and recovery, the animals were killed and subjected to
    gross necropsy. The following organs were removed and weighed:
    submaxillary glands, thymus, lungs and bronchi, heart, thyroid and
    parathyroid, liver and gall-bladder, pancreas, spleen, kidneys,
    adrenals, prostate, testes, ovaries, uterus, brain, and pituitary. An
    additional 26 organs and tissues were removed and prepared for
    microscopic examination. Sections from the kidney of one animal of
    each sex per dose were processed at the end of treatment and recovery
    for electron microscopy, and additional sections from the kidneys of
    one animal of each sex at 0, 2.5, and 5 g/kg bw per day at the end of
    treatment and of one animal of each sex at 5 g/kg bw per day at the
    end of the recovery period were processed for electron microscopy by
    an alternative technique.

         Changes in general condition that were associated with treatment
    were seen in animals of each sex at the intermediate and high doses
    and to a lesser extent in those at the low dose. These included
    increased water intake and frequency of urination, increased incidence
    of soft stools and diarrhoea, vomiting, salivation at dosing and for
    1-1.5 h after dosing, reddening of the mucosa about 30 min after
    dosing and epidermal redness in animals at the high dose, wet oral
    mucous membranes 30 min after dosing, and dryness of the mucosa and
    nose. These were transient effects which were resolved within 24 h and
    may have been due to increased plasma osmotic pressure. None of the
    effects on mucous membranes, salivation, urination, and water
    consumption observed during treatment were present during the recovery
    period. Erythritol had no effect on body temperature, respiratory
    rate, heart rate, or pupil diameter and no significant effect on body
    weight or feed consumption. It caused dose-related increases in water
    consumption in animals of each sex during treatment, but because of
    large individual variations in water intake, consistent, statistically
    significant increases were observed only in dogs at the intermediate
    and high doses and bitches at the high dose. Animals of each sex at
    2.5 and 5 g/kg bw per day showed a gradual increase in water intake
    during treatment. During the recovery period, the water consumption of
    animals at the intermediate dose was higher than that of controls, and
    that of animals at the high dose was comparable. Erythritol had no
    significant effect on any haematological or clinical biochemical
    parameter and had no effect on liver or renal function. The urine
    volumes of animals at the high dose were significantly higher than
    those of controls during treatment but returned to pre-treatment
    concentrations during the recovery period. The urine of bitches at the
    low and intermediate doses had a significantly higher specific gravity

    and osmotic pressure than that of controls during treatment; although
    the specific gravity of the urine of bitches at the high dose was also
    higher than that of controls, it was significantly so only at week 13.
    These differences probably reflected the urinary erythritol
    concentrations. The urine became paler with the increased volume
    associated with erythritol treatment.

         During weeks 7 and 13 of treatment, diarrhoea was observed more
    often in animals at the high dose than in controls, and occult blood
    was observed in the faeces of three of four dogs at this dose that had
    diarrhoea at the end of treatment. Occult blood was a more frequent
    finding in dogs at the high dose because it was often observed in
    association with diarrhoea. During the recovery period, occult blood
    was observed only in the faeces of the one bitch at the intermediate
    dose that was affected during treatment. At the end of the 13-week
    treatment, non-significant decreases in absolute and relative thymus
    weights were seen in dogs at the high dose and bitches at the
    intermediate and high doses, which remained lower in dogs at the high
    dose at the end of the recovery period. The reduced weights were
    associated with atrophy of the thymus in all bitches at the
    intermediate and high doses, one of three dogs at the intermediate
    dose, and all dogs at the high dose. Red discolouration of the small
    intestinal mucosa was observed in two of three dogs at the
    intermediate dose and one of three at the high dose and in one of
    three bitches at the intermediate and high doses, but microscopic
    examination of the small intestine revealed no abnormalities. A
    variety of histopathological changes were seen in the renal tubules of
    two dogs at the high dose at the end of treatment, which included
    desquamation, eosinophilic degeneration, hydropic degeneration,
    pyknosis, slight dilatation, and slight necrosis. At the end of the
    recovery period, hyaline casts were observed in the kidneys of one dog
    at the intermediate dose and eosinophilic degeneration and pyknosis of
    the tubules of one dog at the high dose.

         The dose of 1.25 g/kg bw per day caused a marginal increase in
    the incidence of diarrhoea and/or loose stools in dogs, but since
    diarrhoea in these animals was not associated with changes in body
    weight, haematological or clinical biochemical end-points, or
    intestinal lesions, it was considered not to be toxicologically
    significant. The NOEL was 1.25 g/kg bw per day on the basis of
    increased incidences of vomiting, diarrhoea and/or loose stools,
    increased water consumption, and atrophy of the thymus in animals of
    each sex at higher doses. The dose of 5 g/kg bw per day was associated
    with increased urine volume, the presence of occult blood in urine,
    and histopathological changes in the kidneys of dogs. While the
    clinical signs and clinical biochemical and urinary parameters had
    disappeared by the end of the four-week recovery period, thymic
    atrophy and renal abnormalities were still apparent (Yamaguchi et al.,
    1990).

         Groups of four beagles of each sex were given 0, 1, 2.2, or 5
    g/kg bw per day of erythritol (solution of 19.7 or 20.1%; purity not
    stated) by intravenous injection for 180 days. Controls received
    physiological saline at the same volume as that received by the
    high-dose group. An additional two animals of each sex received 0,
    2.2, or 5 g/kg bw per day by the same schedule and were allowed to
    recover without treatment for one month after treatment. As two
    bitches at the high dose had to be killed after dosing errors, a
    supplementary study was initiated in which groups of two bitches were
    given erythritol at 0 or 5 g/kg bw per day by the same protocol as
    used in the main study. The data from the two studies were combined
    such that there were 8, 4, 6, and 6 bitches in the treatment phase and
    4, 0, 2, and 2 bitches in the recovery phase at 0, 1, 2.2, and 5 g/kg
    bw per day, respectively. All animals were observed at least twice a
    day during treatment and recovery for deaths and clinical signs of
    toxicity, and body weights and feed and water consumption were
    measured weekly throughout the study. Blood samples and 24-h urine
    samples were collected from each animal before dosing, at three and
    six months of treatment, and at the end of the recovery period. Urine
    samples collected during the first 3 h and over 24 h were analysed,
    and the 24-h urine volumes were measured. The blood samples were
    anlaysed for haematological and clinical biochemical end-points.
    Ophthalmoscopic examination, electrocardiography, and liver and renal
    function tests (clearance of indocyanine green and
    phenolsulfonphthalein, respectively) were performed before dosing, at
    three and six months of treatment and at the end of recovery. After a
    16-h fast on the last day of dosing or at the end of the recovery
    period, all surviving animals were killed and necropsied. After gross
    examination, the liver, kidneys, spleen, heart, lungs, cerebrum,
    cerebellum, pituitary, adrenals, thyroids, thymus, pancreas, salivary
    glands, sublingual glands, testes, prostate, ovaries, and uterus were
    removed and weighed, and 52 organs and tissues from animals in all
    groups were preserved for microscopic examination. Samples from the
    livers and kidneys of two animals of each sex at each dose at the end
    of treatment, all animals at the end of recovery, and any animal found
    dead during the study were fixed and prepared for examination by
    electron microscopy.

         A number of effects on general condition were noted in treated
    animals which included hypoactivity on the first day of dosing,
    increased frequencies of urination and vomiting during the first hour
    after dosing, and the appearance of red urine early in the study,
    particularly in animals at the higher dose. Salivation was also
    reported in bitches at the high dose. The incidences of diarrhoea and
    soft stools were not increased by treatment. During the recovery
    period, the incidences of vomiting and salivation were comparable in
    all groups. Erythritol had no effect on body weight or feed
    consumption in animals of either sex during treatment or recovery, but
    a dose-related increase in water consumption was seen during treatment
    that reached statistical significance ( p < 0.05) in dogs at all
    doses and in bitches at the two higher doses; no statistically

    significant difference in water consumption was seen during the
    recovery period. Some minor changes in electrocardiographic parameters
    were observed in treated animals, but the responses were inconsistent
    and the values of the affected parameters were within normal ranges
    and the changes were considered not to be toxicologically significant.
    Treated bitches showed a dose-related increase in blood urea nitrogen
    at three and six months which was statistically significant at the
    intermediate and high doses at three months and at all doses at six
    months. Since there was no evidence of impaired renal function and no
    histopathological evidence of renal damage, the toxicological
    significance of the increased blood urea nitrogen is questionable.
    Treatment resulted in a dose-related increase in urine volume which
    was statistically significant at 2.2 and 5 g/kg bw per day in dogs and
    at 5 g/kg bw per day in bitches, and a dose-related decrease in the
    specific gravity of the urine which was statistically significant ( p
    < 0.05) in dogs at the intermediate and high doses. There were no
    differences in urine volume or specific gravity at the end of the
    recovery period. Controls, given physiological saline, excreted more
    urinary sodium and chlorine during treatment than before
    administration or during recovery, while potassium excretion was
    unaffected. In comparisons with pre-treatment urinary excretion of
    sodium, conducted to circumvent the problem introduced by use of
    physiological saline in the controls, erythritolwas found to increase
    excretion in animals of each sex at the intermediate and high doses.
    These doses also tended to increase the chlorine excretion over that
    before treatment, most markedly in males. Urinary excretion of
    potassium was decreased in dogs at both three and six months and in
    bitches at the high dose at three months. Urinary excretion of sodium,
    chlorine, and potassium was comparable in all groups during the
    recovery period. There were no treatment-related differences in the
    results of liver and renal function tests. Haemorrhage was observed in
    the urinary bladders of animals that had reddish urine during
    treatment. but there were no other gross or histopathological findings
    which could be related to treatment.

         There was no NOEL in this study. The LOEL was 1 g/kg bw per day
    on the basis of increased blood urea nitrogen and vomiting in bitches
    and increased water consumption by dogs at this dose. Higher doses
    were associated with vomiting, increased water consumption, increased
    urine volume, haemorrhage in the urinary bladder resulting in reddish
    urine, increased urinary sodium excretion in animals of each sex, and
    increased urinary chlorine excretion, decreased urinary potassium
    excretion, and atrophy of the acinar epithelium and foliate
    interstitial fibrosis of the prostate in dogs (Kamata, 1990b).

         Groups of four beagles of each sex were offered 400 g/day of a
    diet containing 0, 2, or 5% or 450 g/day of a diet containing 10%
    erythritol (purity not stated) for up to 53 weeks. Animals given 5 or
    10% were introduced to the formulated diets gradually over one week.
    The intakes of erythritol were estimated from data on feed consumption

    to be approximately 0, 0.7, 1.7, and 3.8 g/kg bw per day for the four
    groups, respectively. All animals were observed at regular intervals
    each day for signs of ill health or reaction to the test material.
    Body weights and feed consumption were recorded weekly from two weeks
    before the start of treatment until the end of the study. Water
    consumption was measured over one day once before treatment and once
    each month for the first five months of treatment and thereafter over
    three days of each month. Once before treatment and during weeks 13,
    26, 38, and 52, both eyes of all animals were examined by indirect
    ophthalmoscopy, and haematological, clinical chemical, and urinary
    parameters were measured in blood samples and 24-h urine samples.
    Urinary excretion of sodium, potassium, and chlorine was measured, but
    that of calcium was not. Faecal samples were collected at the same
    time, analysed for erythritol content, and weighed. Water consumption
    was determined during urine collection. All animals were killed after
    53 weeks of treatment and subjected to gross necropsy. The adrenals,
    brain, caecum, heart, kidneys, liver and gall-bladder, lungs, spleen,
    pancreas, pituitary, prostate or uterus, submaxillary salivary gland,
    testes or ovaries, thymus, and thyroid with parathyroids were weighed,
    and 19 tissues or organs were examined histopathologically.

         Erythritol at concentrations up to 10% of the diet appeared to be
    well tolerated, as there were no reports of vomiting or diarrhoea and
    no significant effects on feed consumption. Treatment did not
    adversely affect the body weights or body-weight gain of animals of
    either sex. The diet containing 10% erythritol was associated with
    increased water consumption although there were large inter-and
    intra-individual variations in water consumption during the study,
    particularly in animals at the high dose. No effects on haematological
    or clinical biochemical parameters were noted. The urinary parameters
    were reported only as median values because of the large individual
    variations within each group. No mean data were given, and no
    statistical analysis was conducted. There were no apparent differences
    in any of the parameters between groups before treatment. At the first
    determination during week 13, dogs and bitches at the intermediate and
    high doses produced 30-40% less urine than before treatment and in
    comparison with their respective control groups, whereas throughout
    the remainder of the study animals of each sex at 10% erythritol
    produced more urine than controls. During week 13, the 24-h urinary
    excretion of sodium, calcium, magnesium, and GGT in animals of each
    sex at the intermediate and high doses and of potassium and phosphorus
    in bitches at the intermediate and high doses was lower than that of
    their controls either as an absolute measurement or when expressed
    relative to excretion of creatinine. There was no difference between
    treated and control animals in the urinary excretion of electrolytes
    or GGT at any subsequent determination. Slightly higher absolute and
    relative kidney weights were observed in animals at the high dose,
    which were not statistically significant, and no unusual
    histopathological observations were made. The absolute weights of the
    prostate were significantly higher in dogs at the high dose than in
    controls, and the relative weights followed the same trend. The
    increased prostate weights were associated with an increased incidence

    of cystic acini. As this is a common finding in dogs and occurs with
    wide inter-individual variation, it was considered to be irrelevant
    toxicologically. Treatment had no effect on caecal diameter.

         The NOEL was 5% erythritol in the diet, equal to 1.7 g/kg bw per
    day, on the basis of increased water consumption and urine production
    in animals of each sex at 10% erythritol in the diet (Dean & Jackson,
    1992; Dean et al., 1996).

    2.2.3  Long-term studies of toxicity and carcinogenicity

         Rats

         Groups of 20 Wistar rats of each sex were given access to diets
    containing 0, 1, 3, or 10% erythritol (purity, 98.5%)  ad libitum,
    equal to 0, 0.46, 1.4, and 5 g/kg bw per day, respectively, for males
    and 0, 0. 54, 1.7, and 7.5 g/kg bw per day, respectively, for females,
    for 78 weeks. Each animal was observed at least once daily for deaths
    and clinical signs of toxicity and palpated weekly to detect masses.
    Ophthalmoscopic examinations were made on all rats before the start of
    the study and on all controls and rats at the high dose during weeks
    13, 26, 52, and 77. Individual body weights and feed and water
    consumption were recorded for each cage weekly for the first 13 weeks
    and then every four weeks for the remainder. Blood samples were
    collected from 10 rats of each sex per group at 13, 26, 52, and 78
    weeks for determination of standard haematological and clinical
    biochemical parameters, and tests for renal concentration were at 12,
    25, 52, and 77 weeks. The rats were deprived of water for 24 h and of
    feed during the last 16 h of this 24-h period, when all rats were
    housed individually in metabolism cages to allow collection of urine
    and faeces. At the end of the collection period, a blood sample was
    collected for determination of glucose, and individual urine samples
    were analysed, calcium being the only electrolyte measured. During
    week 79, all surviving rats were killed and examined grossly for
    pathological changes. The weights of the adrenals, brain, caecum (full
    and empty), heart, kidneys, liver, ovaries, pituitary, spleen, testes,
    and thyroid were recorded, and 28 organs and tissues were preserved
    for histopathological examination, including the colon, rectum, small
    intestine, and urinary bladder. All of these tissues and organs were
    examined microscopically for control and rats at the high dose, the
    kidneys, liver, and lungs of all rats at the low and intermediate
    doses, all gross lesions in all rats, and all the tissues and organs
    of rats that died or were killed when moribund during the study.

         The only treatment-related clinical observation was the
    occurrence of soft faeces in all rats at the high dose at some time
    during the first two weeks of the study. Erythritol had no effect on
    survival: two male controls and six at the high dose and one female
    control and three at the low dose died or were killed in a moribund
    condition during the study, but these deaths were considered not to be
    related to treatment. The body weights of males at the high dose were
    statistically significantly lower than those of controls by 5-9%; the
    body weights of females at this dose also tended to be lower (3-8%),

    but the differences were not statistically significant. Treated rats
    of each sex tended to consume more feed than controls but the
    differences were not significant. When the intakes of
    erythritol-containing diets were adjusted for their reduced energy
    content, the intakes were comparable to or lower than those of
    controls. Treated rats of each sex had a dose-related increase in
    water consumption, and males and females at the high dose consumed
    significantly ( p < 0.05) more water than controls during most of
    the collection periods, whereas those at the low and intermediate
    doses consumed significantly more water during only a few of the
    collection periods. None of the differences in haematological
    parameters could be related to treatment. Rats of each sex at the high
    dose had increased plasma alkaline phosphatase activity throughout the
    study, although the differences were significant only in females at
    week 13 and in males at weeks 26 and 78. In the tests for renal
    concentration, rats at 10% erythritol tended to have higher urine
    volumes, which were statistically significant ( p < 0.05) only
    during weeks 12 and 25. The development of severe nephrosis or
    pyelonephrosis in male rats at the end of the study may have
    contributed to the variation in the data. Urine density was comparable
    in treated and control animals of each sex except at week 12 when it
    was significantly lower in males at the high dose as compared with
    controls. Treatment with 10% erythritol was associated with increased
    16-h urinary calcium excretion, which was statistically significant in
    all tests in females and in all but that conducted during week 77 in
    males. The absolute and relative weights of the full and empty caeca
    showed dose-related increases in animals of each sex which were
    statistically significant at concentrations of 3 and 10% erythritol
    for the full caecum and at 10% for the empty caecum. The absolute and
    relative kidney weights tended to be increased in treated rats, the
    highest values being observed in those at the intermediate dose, but
    only the absolute kidney weight of females at the intermediate dose
    differed significantly from that of controls. Histopathological
    examination of the kidneys revealed no treatment-related effect on the
    incidence of hydronephrosis or any other lesion in females. The
    incidences of all other neoplastic and non-neoplastic lesions were
    comparable in control and treated groups, including pelvic
    nephrocalcinosis in the kidneys of female rats. 

         The NOEL was 3% erythritol in the diet, equal to 1.4 g/kg bw per
    day, on the basis of increased water consumption, increased plasma
    alkaline phosphatase activity, diuresis, and increased urinary calcium
    excretion at the high dose (Til & van Nesselrooij, 1994).

         Groups of 50 Wistar rats of each sex were given access to diets
    containing 0, 2, 5, or 10% erythritol (purity, > 99.9%) (equal to 0,
    0. 86, 2.2, or 4.6 g/kg bw per day for males and 0, 1, 2.6, or 5.4
    g/kg bw per day for females) or 10% mannitol (purity, 98%) (equal to
    4.4 and 5.2 g/kg bw per day in males and females, respectively)
     ad libitum for up to 104 weeks. Satellite groups of 20 male rats
    were fed the same diets for 52 and 78 weeks. In order to adapt the

    rats to the high doses, the dietary concentrations of erythritol and
    mannitol were increased gradually over three weeks. Each animal was
    observed at least once daily for clinical signs of toxicity and
    deaths. Individual body weights were recorded weekly during the first
    13 weeks of the study and once every four weeks thereafter. Weekly
    feed consumption was measured for each cage during the first 26 weeks
    of the study and then during one week each month for the remainder of
    the study. Water consumption was measured weekly during the first four
    weeks of the study and once every four weeks thereafter.
    Ophthalmoscopic observations were made before the beginning of the
    study and during weeks 12, 25, 51, 76, and 102 in all controls and
    groups at the high dose and during week 102 in those at the low and
    intermediate doses. Blood samples for haematological investigations
    were collected from 10 rats of each sex per main study group during
    weeks 26, 51, 78, and 103. For clinical biochemistry, blood was
    collected from the 20 males in the satellite groups during week 26,
    from all surviving males in these groups sacrificed during weeks 53
    and 79, and from 10 rats of each sex per group at terminal sacrifice.
    Urine and faeces were collected from 20 male rats in each group in the
    satellite study during weeks 26, 50, and 78 and from controls and rats
    at 10% erythritol and 10% mannitol during week 42 and, in the main
    study, from 20 male rats at each dose during week 102. Rats were
    transferred to individual metabolism cages for collection of 48-h
    urine and faeces, and feed and water consumption was measured during
    the sampling period. The faecal samples were extracted to allow
    assessment of unabsorbed sugar alcohols. After gross examination at
    weeks 53, 79, and 105-107, the adrenals, brain, caecum (full and
    empty), heart, kidneys, liver, ovaries, pituitary, spleen, testes, and
    thyroid were removed from each animal and weighed. In addition, 27
    organs were retained for histopathological examination, including the
    colon, rectum, small intestine, and urinary bladder. In the satellite
    and main studies, all of these tissues and organs from controls and
    from rats at 10% erythritol and 10% mannitol and all those that died
    before the end of the study were examined histologically. Selected
    tissues were examined, including the kidneys, from rats at the low and
    intermediate doses in the satellite study and from surviving rats at
    the low and intermediate doses in the main study. All gross lesions
    observed in rats in any group were examined microscopically. 

         There were no treatment-related effects on the mortality rate or
    on the general condition or behaviour of the animals. The body weights
    of male rats at the intermediate and high doses and females at the
    high dose were significantly lower than those of controls during most
    of the study. At the end of the study, surviving males at 2%
    erythritol also weighed significantly ( p < 0.05) less than
    controls. Feeding of 10% mannitol also resulted in reduced body
    weights in both males and females. Except for males fed 10%
    erythritol, the body weights remained within 10% of those of controls,
    and survival was not adversely affected even at the 10% dose. The feed
    consumption of males at 10% erythritol was higher than that of

    controls, while their feed efficiency tended to be lower, although the
    differences were not always statistically significant. When the feed
    consumption was adjusted for the energy density of the 10% erythritol
    diet, the intakes were comparable or slightly greater than that of
    controls. Erythritol had no effect on feed intake or feed efficiency
    in female rats or in rats of either sex fed 10% mannitol. Males and
    females at 5 and 10% erythritol consumed more water than controls, and
    this effect was consistently statistically significant ( p < 0.05)
    for males at the high dose. Males and females fed 10% mannitol also
    had a small increase in water consumption, which was occasionally
    statistically significantly different from that of controls. No
    treatment-related effects on haematological or clinical biochemical
    parameters were noted with either erythritol or mannitol. Feeding 5 or
    10% erythritol or 10% mannitol resulted in increased 24-h urine
    volumes which were statistically significant at 26 weeks for rats at
    5% erythritol and at all times except 102 weeks for those given 10%
    erythritol or mannitol. While erythritol had no significant effect on
    urine density, mannitol statistically significantly decreased the
    density at all times except 102 weeks. The osmolarity of the urine was
    lower throughout the study in males fed 10% erythritol or 10% mannitol
    than in controls and the difference was statistically significant ( p
    < 0.05) during the first year of the study. Urinary excretion of
    total and low-molecular-mass protein, the enzymes GGT and NAG, and of
    sodium, potassium, and phosphate was increased at all times in males
    receiving 10% erythritol except in week 102. These increases were
    consistently statistically significant ( p < 0.05) except for
    urinary excretion of potassium. Significant increases in excretion of
    the enzymes were also noted in rats at 5% erythritol during the first
    year of the study. Urinary excretion of calcium and citrate was
    increased at all times in males receiving diets containing 10%
    erythritol or 10% mannitol, although the increase was consistently
    significant only for excretion of calcium. Necropsy of male rats at
    52, 78, and 104 weeks and of females at 104 weeks revealed
    significantly ( p < 0.01) higher absolute and relative weights of
    the full or empty caecum in rats given 10% erythritol or mannitol when
    compared with controls, except for the relative full or empty caecal
    weights of females fed mannitol. The caecal weights were also
    increased in males fed 5% erythritol, although the differences were
    not always statistically significant, and the empty caecal weights
    were significantly increased in females fed 5% erythritol. No
    histopathological changes were associated with the increased weights.
    At 104 weeks, the absolute and relative weights of the adrenals in
    females given 10% erythritol were significantly ( p < 0.05) higher
    than those of controls. No evidence of phaeochromocytomas or other
    pathological changes were associated with the increased weights. Males
    fed 10% mannitol also had a higher incidence of benign
    phaeochromocytoma and of medullary hyperplastic foci than controls,
    which was not statistically significant. At 52 and 78 weeks, but not
    at terminal sacrifice at week 104, males fed 10% erythritol had higher
    absolute and relative kidney weights than controls, but only the
    relative weights were statistically significantly ( p < 0.05)
    different. Marginal effects on the kidney weights of female rats at 5
    and 10% erythritol were seen, which reached statistical significance

    only for the relative weights of animals at 10%. Mannitol had no
    significant effect on the absolute or relative kidney weights.
    Nephrosis occurred more frequently and was more severe in males fed 5
    or 10% erythritol at 78 weeks but not at study termination, suggesting
    that erythritol might affect the rate at which nephrosis develops in
    male rats but not the overall incidence of this lesion. At week 104,
    there was a higher incidence ( p < 0.05) of pelvic nephrocalcinosis
    in the kidneys of all treated females than in controls, with
    incidences of 8/49, 20/50, 35/49, 26/49, and 39/50 at 0, 2, 5, and 10%
    erythritol and 10% mannitol, respectively. Females fed 10% mannitol
    also had a higher incidence of renal pelvic transitional-cell
    hyperplasia than controls, and males showed a significant ( p <
    0.01) increase in the incidence of pelvic nephrocalcinosis at weeks 78
    (0/20 and 8/20) and 104 (8/49 and 31/48). The incidence of
    nephrocalcinosis was not increased in males at any dose, and the
    effect in females was considered not to be toxicologically significant
    since this is a common age-associated lesion in female rats and the
    incidence in controls was considered to be low in the context of
    background rates. In a 78-week study conducted concurrently in the
    same laboratory with the same strain of rats, the control incidence of
    nephrocalcinosis was 60% (12/20), which is comparable to that in
    treated female rats in this study. Other effects seen with mannitol
    were a significantly higher incidence of mononuclear-cell infiltrates
    in the Harderian glands of female rats than in controls and a
    significantly higher incidence of C-cell adenoma of the thyroid in
    males.

         Erythritol was thus not carcinogenic at doses up to 10% of the
    diet, equal to 4600 mg/kg bw per day. The NOEL was 2% in the diet,
    equal to 8.6 g/kg bw per day, on the basis of increased water intake
    and caecal weights in animals of each sex and diuresis in male rats at
    5% and higher. At 10% in the diet, erythritol was associated with
    increased urine volume and urinary excretion of calcium, enzymes, and
    other electrolytes in males, increased kidney weights in animals of
    each sex, and increased adrenal weights in females (Lina et al., 1994,
    1996).

    2.2.4  Genotoxicity

         Studies of the genotoxicity of erythritol are summarized 
    in Table 2.

    2.2.5  Reproductive toxicity

         Mice

         Groups of 24 male Crj:CD-1 (ICR) mice were given 0, 1, 2, 4, or 8
    g/kg bw per day of erythritol (purity, 99.6%) by gavage from six weeks
    of age for nine weeks before mating, throughout mating, and until
    copulation was confirmed by the presence of a vaginal plug in their
    female mates. Groups of 24 female Crj:CD-1 (ICR) mice were treated in
    the same way from nine weeks of age for 15 days before mating,

    throughout the mating period, and until day 6 of gestation; the
    presence of a vaginal plug was considered evidence of copulation. The
    mice were observed daily for clinical signs of toxicity and deaths
    before and after dosing, and pregnant females were observed daily for
    clinical signs of toxicity and deaths through the remainder of
    gestation. The body weights of males were recorded weekly during
    treatment. Females were weighed before mating on days 0, 3, 6, 9, 12,
    and 15 of dosing, and pregnant females were weighed daily on days 0-18
    of gestation. The feed and water intakes of males were measured each
    week during the nine-week treatment period before mating, and those of
    females were measured over three-day periods during the 15-day
    pre-mating period and throughout gestation. Daily feed and water
    consumption were calculated from these data. Pregnant females were
    killed on the morning of gestation day 18 and examined
    macroscopically, and their ovaries, uterus, and vagina were removed.
    The ovaries and placenta were weighed and the number of corpora lutea,
    the number and placement of implantation sites, the numbers of early
    (implantation scars and resorbed embryos) and late (macerated fetuses
    and dead fetuses at term) fetal deaths, and the number of live fetuses
    were recorded. All live fetuses were weighed, sexed, and examined
    externally, and about half of the fetuses from each litter were fixed
    in 95% ethanol and their visceral organs and tissues examined
    macroscopically.

         Diarrhoea was observed sporadically in animals of each sex
    throughout administration of 4 and 8 g/kg bw per day but was more
    severe and occurred more frequently early in the treatment period at 8
    g/kg bw per day. No treatment-related effects on survival were noted.
    Dietary administration of erythritol had no effect on the body weights
    of male mice during treatment, on the body weights of females before
    mating or during gestation, or on feed consumption of males or females
    before mating or of females during gestation. Before mating, the
    average daily water consumption of treated males was higher than in
    controls, and statistical analysis of the weekly data indicated that
    the increase in water intake was significant ( p < 0.05) in all
    treated groups during week 1, for rats at the intermediate and high
    doses during weeks 1 and 2, and for rats at the high dose during weeks
    3-6, inclusive. During the 15-day treatment period before mating,
    females receiving erythritol consumed more water than controls, the
    increases being significant ( p < 0.05) for rats at 4 and 8 g/kg bw
    per day in three out of five and four out of five of the collection
    periods, respectively. Erythritol had no significant effect on the
    water consumption of females during gestation, the length of the
    estrus cycle, the copulation rate, the pregnancy rate, the number of
    corpora lutea or resorptions, or the average numbers of implantations
    per dam or the implantation ratio (number of implants per corpora
    luteum), the numbers of live and dead fetuses per dam, or the fetal
    sex ratio. The weights of the ovary and placenta were comparable in
    all treated groups. Two of 24 males at the high dose had dilatation of
    the renal tubules at terminal necropsy, which the investigators
    attributed to the diuretic activity of erythritol. There were no

    macroscopic or microscopic findings in female mice that could be
    related to treatment, and no external, skeletal, or visceral changes
    were detected in the fetuses.

         Erythritol at doses up to 8 g/kg bw per day thus had no effect on
    the reproductive performance of a single generation of male and female
    mice. In animals of each sex, doses of 4 g/kg bw per day and higher
    were associated with diarrhoea, mainly early in the dosing period, and
    increased water consumption. A very low incidence of renal tubular
    dilatation was noted in males receiving 8 g/kg bw per day for more
    than nine weeks (Tateishi et al., 1989).

         A study of similar design was conducted in mice treated
    intravenously. Groups of 24 male Crj:CD-1 (ICR) mice received 0, 1,
    1.73, or 3 g/kg bw per day erythritol (purity, 99.8%) by intravenous
    injection (site not specified) from six weeks of age for nine weeks
    before mating, throughout mating, and until copulation was confirmed
    by a vaginal plug in their female mates. Groups of 24 female Crj:CD-1
    (ICR) mice received the same doses of erythritol by intravenous
    injection from nine weeks of age, for 15 days before mating,
    throughout the mating period, and until day 6 of gestation. Mice were
    observed daily for clinical signs of toxicity and deaths before and
    after dosing, and pregnant females were observed daily for clinical
    signs of toxicity and deaths throughout the remainder of gestation.
    The body weights of males were recorded weekly during treatment.
    Females were weighed before mating on days 0, 3, 6, 9, 12, and 15 of
    dosing. Those that did not become pregnant were weighed every three
    days, and those that were pregnant were weighed daily on days 0-18 of
    gestation. The feed and water intakes of males were measured each week
    during the nine-week treatment period before mating, whereas those of
    females were measured over three-day periods during the 15-day
    pre-mating period and throughout gestation. Daily feed and water
    consumption was calculated from these data. Pregnant rats were killed
    on gestation day 18 and examined macroscopically, and their ovaries,
    uterus, and vagina were removed. The ovaries and placenta were weighed
    and the number of corpora lutea, number and placement of implantation
    sites, numbers of early (implantation scars and resorbed embryos) and
    late (macerated fetuses and dead fetuses at term) fetal deaths, and
    number of live fetuses were recorded. All live fetuses were weighed,
    sexed, and examined externally and then the visceral organs and
    tissues from about half the fetuses in each litter were examined
    macroscopically. The skeletons of these fetuses were examined for
    abnormalities and ossification. The heads of the remaining fetuses and
    their visceral organs were examined for abnormalities.


        Table 2. Results of assays for the genotoxicity of erithritol
                                                                                                   

    End-point      Test system                 Concentration      Results      Reference
                                                                                                   

    Reverse        S. typhimurium TA98,        370-30 000         Negative     Blijleven
    mutationa      TA100, TA1535,              µg/plate                        (1990)
                   TA1537, TA1538

    Reverse        S. typhimurium TA98,        15.8-5000          Negative     Kawamura
    mutationb      TA100, TA1537;              µg/plate                         et al. (1996)
                   E. coli WP2 uvrA

    Chromosomal    Chinese hamster             1.25-10 mmol/L     Negative     Nakatsuru
    aberrationb    fibro-blast line CHL/IU                                     et al. (1988);
                                                                               Kawamura et al.
                                                                               (1996)
                                                                                                   

    a In the presence and absence of a metabolic activation system from the 9000 × g fraction of
      the livers of Aroclor 1254-induced male rats
    b In the presence and absence of a metabolic activation system of unspecified origin
    

         Two males and one female at the high dose died during the study.
    Administration of erythritol was not associated with diarrhoea and had
    no significant effect on the body weights of males or females before
    mating or on those of females during gestation. The feed consumption
    of males before mating and of females before mating and during
    gestation was not affected by treatment. Males at 1.73 and 3 g/kg bw
    per day tended to consume more water than controls, although the
    increases did not achieve statistical significance, whereas females at
    3 g/kg bw per day consumed significantly ( p < 0.05) more water than
    controls before mating (5.6, 5.6, 6.2, and 7 g/kg bw per day in mice
    at 0, 1, 1.73, and 3 g/kg bw per day, respectively, and
    non-significantly more during gestation. Erythritol had no effect on
    reproductive or fetal parameters, the length of the estrus cycle, the
    copulation rate, the pregnancy rate, or the numbers of corpora lutea,
    implantations, or resorptions per dam. The findings at necropsy were
    discussed in the text of the report but no data were included, and the
    total number of animals per group that were examined microscopically
    was not specified. The weights of the ovaries and placenta were
    comparable in all treated groups. Histopathological examination
    revealed two females at the high dose that had slight dilatation of
    the renal tubules and one male at the high dose which had dilatation
    of the renal tubules and Bowman capsule. The investigators attributed
    these findings to the diuretic effect of erythritol. None of the other
    histopathological changes could be attributed to treatment. No
    external, skeletal, or visceral changes were observed in the fetuses.

         Thus, intravenous doses of erythritol at up to 3 g/kg bw per day
    had no effect on the reproductive performance of male or female mice
    over a single generation. Doses of 3 g/kg bw per day were associated
    with a low incidence of deaths, increased water consumption, and
    dilatation of the renal tubules in mice of each sex (Tateishi et al.,
    1992).

         Rats

         Randomly assigned groups of 24 Wistar rats of each sex were fed
    diets containing 0, 2.5, 5, or 10% erythritol (purity, 99.9%) (equal
    to 1.5, 3.1, and 6.5 g/kg bw per day for males and 1.7, 3.3, and 7.1
    g/kg bw per day for females before mating and during gestation and
    3.6, 7.5, and 16 g/kg bw per day during lactation) for two consecutive
    generations with one litter per generation. The F0 and F1 parents
    were fed their respective diets for 10 weeks before mating and were
    then mated within dose groups at a ratio of one male to one female.
    Parental generations were observed at least once daily for clinical
    signs of toxicity and mortality. The individual body weights of all
    parental males were recorded weekly throughout the study, whereas
    parental females were weighed weekly before mating, on days 0, 7, 14,
    and 21 of gestation, and on days 1, 7, 14, and 21 of lactation. Feed
    consumption before mating was measured weekly per cage for animals of
    each sex, and individual feed consumption was measured for each dam

    weekly during gestation and lactation. Water consumption was not
    measured. Reproductive performance was assessed for each generation by
    recording the mating index, female fertility index, duration of
    gestation, fecundity index, and gestation index. F1 and F2 litters
    were examined one day after birth to determine the numbers of live and
    dead pups, and the pups were weighed, sexed, and examined for external
    abnormalities. On day 4 of lactation, the litters were culled to eight
    pups, consisting when possible of four pups of each sex. On days 4, 7,
    14, and 21  post partum, the litters were examined to determine pup
    viability and body weights and to detect external abnormalities. F1
    pups were weaned on day 21  post partum, and 24 pups of each sex at
    each dose were selected at random from as many litters as possible to
    rear the F2 generation. After the F1 pups had been weaned, all F0
    parents were killed for necropsy. All pups of the F2 litters were
    examined externally for abnormalities at weaning and subsequently
    discarded. The uterus of each dam was examined to determine the number
    of implantation sites, and samples of the ovaries, uterus including
    cervix, vagina, testes, epididymides, seminal vesicles, prostate,
    coagulating glands, pituitary gland, and any tissue or organ with
    macroscopic lesions were preserved, sectioned, and stained for
    microscopy. All of these tissues from controls and rats at the high
    dose were examined microscopically.

         Rats of each sex in the F0 generation but none of the F1
    parental rats had diarrhoea during the first few days of treatment
    with 10% erythritol. The body weights of F0 and F1 males at 10%
    erythritol were lower than those of controls, but those of females
    were only slightly affected. After an initial decrease in the group
    given 10% erythritol, feed consumption was increased in all treated
    rats. The increase was dose-related and was probably secondary to the
    lower caloric value of the erythritol diets. Treatment did not affect
    the reproductive performance or fertility of the F0 or F1 parental
    rats, as indicated by a lack of effect on the indices for mating,
    female fertility, fecundity, and gestation. Precoital times and the
    duration of gestation were comparable in all treated groups in both
    generations. No erythritol-related macroscopic or microscopic findings
    were made at necropsy of parental animals. The live litter size on day
    1  post partum and the sex ratio were comparable in all F1 and F2
    litters. While the viability index (survival to day 4) was slightly
    but significantly reduced for the F1 pups, a contrasting trend was
    observed for the F2 pups. Treatment had no effect on the lactation
    index (days 4-21 of lactation) of the F1 and F2 litters. The body
    weights of F1 pups were comparable in all groups on day 1  post
     partum, but those at the high dose gained less weight during
    lactation, and their mean body weights were 8 and 14% lower than those
    of controls on days 14 and 21 of lactation, respectively. The decrease
    in mean body weight was statistically significant ( p < 0.001) on
    day 21 of lactation. Erythritol had no effect on the body weights of

    F2 pups. A higher incidence of small pups (< 75% of the mean body
    weight of control pups) was seen in the F1 generation of dams fed 10%
    erythritol from day 14 of lactation but not in the litters of dams at
    the low and intermediate doses. None of the other external
    abnormalities in F1 or F2 pups was related to treatment.

         The NOEL for reproductive toxicity was the highest dose tested,
    equal to 3.1 g/kg bw per day. Although decreased body weights were
    seen in F1 pups at this dose during lactation and throughout the
    remainder of the study, the F2 litters were not affected (Smits-van
    Prooije et al., 1996a; Waalkens-Berendsen et al., 1996).

    2.2.6  Developmental toxicity

         Mice

         Erythritol (purity not stated) was administered as a 20.1%
    solution by intravenous injection to groups of 42 pregnant female SPF
    Crj:CD-1 (ICR) mice at doses of 0, 1, 2, or 4 g/kg bw per day on days
    6-15 of gestation, and 27 dams per group were scheduled for sacrifice
    on day 18 of gestation while the remaining 15 dams per group were
    allowed to give birth naturally and to nurse their pups to weaning at
    21 days. Dams were observed twice daily for deaths and clinical signs
    of toxicity during treatment and once daily at other periods. Body
    weights were recorded daily on days 0-18 of gestation and on days 0,
    4, 7, 14, and 21 of lactation for dams allowed to deliver naturally.
    Feed and water intake were measured daily throughout gestation and
    over days 0-4, 4-7, 12-14, and 19-21 of lactation. In dams sacrificed
    on day 18 of gestation, the number of corpora lutea, number and sites
    of implantations, numbers of early (unformed fetuses) and late (formed
    fetuses but indistinct extremities) resorptions, and numbers of
    macerated, dead, and live fetuses were recorded, and viable fetuses
    were examined for external abnormalities, weighed, and sexed. About
    two-thirds of the viable fetuses from each dam were fixed, cleaned,
    and stained to allow examination of skeletal abnormalities,
    variations, and ossification. The remaining viable fetuses and those
    with external abnormalities were fixed and their visceral organs
    examined and preserved.

         In dams allowed to deliver naturally, the numbers of live and
    dead pups were counted and the pups were sexed and examined for
    external abnormalities. Throughout lactation, the dams and pups were
    observed daily for deaths, general clinical signs, and nursing
    behaviour. On day 4 of lactation, the litters were culled to eight
    pups consisting of four pups of each sex whenever possible. The pups
    were weighed on days 0, 4, 7, 14, and 21 of lactation and examined for
    separation of auricles, hair growth on the back, and righting reflex
    on day 4, for eruption of the incisors and separation of the eyelids
    on day 11, for walking reflex on day 14, and for free-fall reflex and
    auditory function on day 21. Pups that showed delayed differentiation
    were observed daily up to and after weaning, until development was

    complete. After weaning of pups on day 21, the dams were killed and
    necropsied. The number of implantation sites was counted, and any
    organs or tissues with abnormalities were preserved. F1 pups were
    kept up to 70 days of age, during which time individual body weights
    were recorded weekly. From 25 days after birth, the pups were examined
    daily for the descent of the testes or opening of the vagina. At
    weaning, one pup of each sex per litter was administered various tests
    of behavioural function  Motor coordination was tested by the rotor
    rod method and emotional behaviour (e.g. ambulation, rearing,
    preening, and grooming) on an open field in four-week-old pups, and
    learning ability was assessed in pups at six to seven weeks of age in
    a shuttle-box avoidance test for up to four days or until positive
    memory was retained. The pups used in the learning ability test were
    killed at seven weeks of age and their skeletons were examined by
    X-ray. Mice that were not not used in the behavioural function tests
    were killed at 56 days of age and necropsied, and their brain, lungs,
    heart, liver, spleen, kidneys, testes, and ovaries were removed and
    weighed. Any organs or tissues that showed abnormalities were fixed
    and preserved. The two pups of each sex per litter used in the tests
    of motor coordination and emotional behaviour were used to assess
    reproductive performance. At 10 weeks of age, offspring within the
    same treatment group were mated one-to-one for up to 14 days, but not
    with siblings. Copulation was confirmed by the presence of a vaginal
    plug, and the day of copulation was established as day 0 of gestation.
    Mated females were weighed individually on days 0, 7, 14, and 18 of
    gestation, and the dams were killed and necropsied on day 18.
    Pregnancy was confirmed by the observation of corpora lutea and
    implantations. The number of corpora lutea, number and sites of
    implantations, numbers of early (unformed fetuses) and late (formed
    fetuses but indistinct extremities) resorptions, and numbers of
    macerated, dead, and live fetuses were recorded. Live fetuses were
    weighed, sexed, examined for external abnormalities, and then fixed
    and preserved.

         Dams at 4 g/kg bw per day erythritol showed signs of decreased
    spontaneous movement and staggering gait immediately after dosing on
    days 6-12 of gestation, and two of the females at the high dose, only
    one of which was pregnant, died during gestation. Intravenous
    administration of erythritol during days 6-15 of gestation had no
    effect on the body weights of dams during gestation or lactation. Dams
    at the high dose consumed less feed and more water than controls on
    some days of treatment. Gross examination revealed cleft palate in two
    fetuses in one litter and three fetuses in another litter of dams at
    the high dose, but only two of these fetuses were found on microscopic
    examination to be missing the palatine bone. Exencephaly was observed
    in two fetuses in a third litter from a dam at the high dose and in
    one fetus from a dam at the intermediate dose. Skeletal abnormalities
    were observed in fetuses from one, one, three, and seven litters of
    dams at 0, 1, 2, and 4 g/kg bw per day, respectively. Although the
    incidences of the individual anomalies did not differ significantly

    between the groups, the total incidence was significantly ( p <
    0.05) higher with 4 g/kg bw per day than in controls, due mostly to
    wavy ribs in seven fetuses from one litter, fused sternebrae in four
    fetuses from three litters, and cleft palate in two fetuses from two
    litters at the high dose. There were no treatment-related effects on
    ossification. The incidence of abnormalities observed during visceral
    examination of the fetuses was unaffected by treatment.

         In dams allowed to deliver naturally, erythritol had no effect on
    the length of gestation, litter size, number of live or dead fetuses,
    or sex ratio. Viability to day 4 tended to be lower in the
    intermediate and high-dose groups, but the remaining pups all survived
    to day 21 of lactation and there were no unscheduled deaths after
    weaning. The body weights of male and female pups were comparable in
    all groups during lactation and after weaning. Erythritol had no
    effect on the age at which auricle separation, hair growth, incisor
    eruption, eyelid separation, and descent of the testes or opening of
    the vagina occurred, and there were no abnormal findings in either the
    reflex or auditory function tests. The responses of the F1 mice in
    the motor coordination, behavioural, and learning ability tests were
    comparable in all treated groups. Skeletal examination of seven-week
    old pups by X-ray revealed no treatment-related effects. A slight
    effect on the kidney weights of males at the intermediate and high
    doses was noted, with no evidence of a dose-response relationship. The
    reproductive performance of the F1 generation (pregnancy ratio, body
    weights during gestation, numbers of early and late resorptions,
    numbers of live and dead fetuses, sex ratios, fetal body weights) was
    unaffected by erythritol. and there were no gross abnormalities in the
    F2 offspring that could be attributed to treatment.

         The NOEL for maternal and developmental toxicity was 2 g/kg bw
    per day. Erythritol at 4 g/kg bw per day was toxic to the dams, as
    evidenced by effects on movement and slight effects on feed
    consumption during treatment. Developmental toxicity, characterized by
    increased incidences of cleft palate, wavy ribs, and fused sternebrae,
    was observed in fetuses removed surgically from dams at the maternally
    toxic dose of 4 g/kg bw per day. Erythritol had no effect on the
    morphological or skeletal development of naturally delivered pups and
    no effect on their reproductive performance (Ota et al., 1990).

         Rats

         Groups of 32 pregnant Wistar rats were fed diets containing 0,
    2.5, 5, or 10% erythritol (purity, 99.9%)  ad libitum (equal to 0,
    1.7, 3.3, and 6.6 g/kg bw per day) on days 0-21 of gestation. The
    initial study was conducted on groups of 24 mated females, but because
    an insufficient number became pregnant an additional study was
    conducted with groups of eight mated females, and the data from the
    two studies were combined and presented together in the report. The
    animals were observed for general condition and behaviour twice daily
    during working days and once daily on weekends. Body weights were

    recorded on days 0, 7, 14, and 21 of gestation, and feed consumption
    was measured over days 0-7, 7-14, and 14-21 of gestation and daily
    intake calculated. On day 21 of gestation, the dams were killed and
    subjected to gross necropsy. The uterus and ovaries were then removed
    and the weights of ovaries, full and empty uterus, and placenta were
    recorded. The numbers of corpora lutea, implantation sites, early and
    late resorptions, live and dead fetuses, and grossly visible malformed
    fetuses were determined, and all fetuses were weighed, sexed, measured
    for length, and examined. One-half of the fetuses from each litter
    with their placentas were sectioned for examination of visceral
    alterations. The remaining fetuses were stained with Alizarin Red S
    for skeletal examination. Visceral and skeletal examinations were
    performed on high-dose and control fetuses.

         No clinical signs of toxicity were reported in treated dams, and
    no unscheduled deaths occurred during the study. Dams at 10%
    erythritol gained less weight during gestation, particularly during
    the second week when the difference from controls was significant
    ( p < 0.05). Feed consumption was comparable for all groups
    throughout gestation. Erythritol had no significant effect on the
    numbers of early and late resorptions and live and dead fetuses, sex
    ratios, fetal crown-rump length, body weight, the mean weight of the
    gravid and empty uterus, or the mean ovary weight. No unusual changes
    attributable to erythritol were observed during external, visceral, or
    skeletal examinations.

         Erythritol was thus not embryotoxic, fetotoxic, or teratogenic at
    doses up to 10% of the diet, equal to 6.6 g/kg bw per day. The NOEL
    for maternal toxicity was 5% in the diet, equal to 3.3 g/kg bw per
    day, on the basis of decreased body weights and body-weight gains in
    dams receiving 10% erythritol (Smits-van Prooije et al., 1996b).

         Rabbits

         Groups of 17 mated rabbits were injected with a 20% solution of
    erythritol (purity not stated) into the auricular vein to provide
    doses of 0, 1, 2.2, or 5 g/kg bw per day on days 6-18 of gestation and
    were observed twice daily for deaths and clinical signs of toxicity
    during treatment and once daily at other periods. The body weights of
    the dams were recorded on day 0 of gestation and then every second day
    on days 6-28 of gestation. Feed and water intakes were measured on day
    1, every odd day from day 7 to 27 and on day 28 of gestation. On day
    28 of gestation, the dams were killed and necropsied to confirm
    pregnancy. The numbers of corpora lutea, implantation sites, early and
    late resorptions, and dead and live fetuses were counted, and the
    lungs, heart, liver, spleen, kidneys, adrenals, and ovaries of the
    dams were removed and weighed. Live fetuses were examined for external
    abnormalities, weighed, and sexed. About half of the viable fetuses
    from each litter, including those with external abnormalities, were
    randomly selected for skeletal examination. The remaining viable
    fetuses were subjected to examination of visceral organs.

         One dam at the high dose died on day 20 of gestation, and gross
    necropsy revealed congestion of the lungs and liver. Treatment with
    erythritol at 5 g/kg bw per day was associated with auricular oedema
    and occasional polyuria and lethargy. The body weights of all groups
    were comparable throughout the study. Dams at the high dose consumed
    10-20% less feed than controls during treatment, although the
    difference from controls was not statistically significant, but feed
    consumption remained depressed in this group until day 27 of
    gestation. The water intakes were increased by 40-50% between days 7
    and 19 of gestation in all treated dams as compared with controls, and
    the difference was statistically significant ( p < 0.05) on days
    7-13 of gestation at the low and intermediate doses and on days 7-17
    of gestation at the high dose. The absolute and relative organ weights
    were comparable in all groups. Erythritol did not affect the numbers
    of resorptions, live or dead fetuses, or sex ratios. The body weights
    of fetuses at the high dose tended to be lower than those of their
    respective controls, but the difference was not statistically
    significant. None of the fetuses had external anomalies, and the only
    skeletal anomaly observed was fused sternebrae in one fetus at the
    high dose. The total incidence of skeletal variations was higher at
    the intermediate and high doses (46, 49, 55, and 62% in order of
    ascending dose) but was not statistically significant. Although more
    fetuses at the high dose had variable numbers of presacral vertebrae
    and thirteenth ribs than control fetuses, the differences were not
    statistically significant. Furthermore, the number of litters with
    affected fetuses was similar in all groups. Erythritol had no effect
    on skeletal ossification. Visceral examinations revealed no
    treatment-related effect.

         The NOEL for maternal toxicity was 2.2 g/kg bw per day on the
    basis of the persistent reduction in feed consumption. There was no
    evidence that erythritol was teratogenic at doses up to that which was
    toxic to the does (5 g/kg bw per day). The decrease in body weight and
    increased numbers of fetuses with skeletal variations at the high dose
    provide limited evidence of fetotoxicity at the maternally toxic dose
    of 5 g/kg bw per day (Hashima Laboratory, 1989; Shimizu et al., 1996).

    2.2.7  Special studies on renal function

         In a study to examine the effect of repeated doses of erythritol
    on serum blood urea nitrogen and urinary excretion of electrolytes
    added to the drinking-water, groups of 12 female Wistar rats were
    given erythritol at 8 g/kg bw per day (purity, 99.7%) by gavage and
    given access to tap water containing sodium hypochlorite, a
    low-electrolyte solution (35 mmol/L sodium; 35 mmol/L potassium; 80
    mmol/L chloride; 5 mmol/L calcium) or a high-electrolyte solution (70
    mmol/L sodium; 70 mmol/L potassium; 160 mmol/L chlorine; 10 mmol/L
    calcium)  ad libitum for four weeks. Female rats were used because
    they had greater increases in blood urea nitrogen than males in
    preliminary studies. The animals were observed for clinical signs of
    toxicity before and during the 3-5 h after treatment; body weights

    were recorded on days 0, 6, 12, 18, 24, and 27 of the study; and the
    feed and water intakes were measured over three-day intervals
    throughout the study. On day 27, the animals were transferred to
    metabolism cages and urine was collected during an 18-h fast before
    necropsy, when blood was collected. All organs and tissues were
    examined macroscopically, and the kidneys were weighed and examined
    histopathologically.

         Almost all the treated rats had soft, muddy, or watery stools on
    the first day of treatment, regardless of electrolyte intake, and
    diarrhoea was observed in some rats for the remainder of the study. No
    individual data were reported, and the number of animals with
    diarrhoea was not specified, nor is mention made of whether the
    electrolyte intake influenced the incidence of diarrhoea during the
    latter part of the study. No animals died as a result of treatment.
    None of the treatments significantly affected body weights, and
    electrolyte consumption alone had no effect on daily feed or water
    consumption. Erythritol resulted in a slight decrease in feed
    consumption, which became more pronounced when the rats were also
    exposed to electrolytes. During the first six days of the study,
    erythritol caused increased water intake, which was further increased
    in groups exposed to electrolytes. During the remainder of the study,
    the water intakes of animals given erythritol in tap water were
    comparable to those of controls given tap water, while those in rats
    given erythritol plus low-electrolyte water or erythritol plus
    high-electrolyte water remained higher than those of groups receiving
    erythritol or electrolytes alone. This difference in drinking-water
    intake between erythritol-treated and control groups receiving
    supplementary electrolytes resulted in a higher net intake of
    electrolytes by the erythritol-treated groups. The haematocrit was
    unaffected by erythritol and/or electrolytes. Erythritol alone
    resulted in significant ( p < 0.05) decreases in serum sodium and
    chloride concentrations, while co-administration of electrolytes
    significantly increased the concentrations to the those observed in
    control rats. Erythritol caused a decrease in osmotic pressure that
    was unaffected by co-administration of electrolytes. Although exposure
    to erythritol alone resulted in slightly increased blood urea
    nitrogen, the difference was not statistically significant. When
    erythritol was administered with high-electrolyte concentrations, the
    blood urea nitrogen was reduced to concentrations significantly lower
    than those of animals receiving erythritol plus tap water or tap water
    alone. The total concentration of serum protein in animals given
    erythritol plus tap water was comparable to that of animals given tap
    water alone, but when erythritol was combined with electrolytes, the
    total serum protein concentrations decreased in a dose-related manner.
    The ratio of albumin:globulin was increased by erythritol and further
    increased by co-administered electrolytes. Exposure to electrolytes in
    the drinking-water increased the specific gravity and osmotic pressure
    of the urine and increased the excretion of sodium, potassium,
    chlorine, and urea nitrogen. Although the urine volume was not

    affected by erythritol alone, it increased the specific gravity and
    osmotic pressure over those of controls receiving tap water. The urine
    volume was increased by additional exposure to electrolytes, with no
    further effects on specific gravity or osmotic pressure. Animals
    treated with erythritol plus tap water had greater excretion of
    sodium, potassium, and chlorine, and exposure to electrolytes further
    increased the urinary excretion in proportion to their increased
    intake, irrespective of erythritol treatment. In the absence of
    erythritol, electrolyte supplementation resulted in a small,
    dose-related increase in the absolute and relative kidney weights,
    which differed statistically significantly from those of controls only
    for the relative weights in rats given high-electrolyte water.
    Erythritol treatment alone increased both the absolute and relative
    kidney weights, although only the relative weight differed
    significantly from that of controls receiving tap water.
    Co-administration of electrolytes with erythritol did not affect
    kidney weights when compared with erythritol alone. Independent of
    electrolyte concentrations, dilatation of the renal tubules was
    reported more frequently in erythritol-treated rats. The authors
    concluded that the diuretic action of erythritol leads to
    hyponatraemia and the increase in blood urea nitrogen results from
    compensatory homeostatic mechanisms, since this increase was inhibited
    by addition of electrolytes to the drinking-water (Shibata et al.,
    1991).

         In a study to evaluate the effects of erythritol in rats in which
    renal function was reduced by nephrectomy, groups of six
    nephrectomized or sham-operated male rats were fed diets containing 0,
    2, or 5% erythritol (purity, 99.9%)  ad libitum for four weeks, equal
    to 0, 1.1, and 2.7 g/kg bw per day for the nephrectomized rats and 0,
    1.1, and 2.9 g/kg bw per day for the sham-operated group. Rats were
    observed twice daily for general signs of toxicity, body weights were
    recorded twice weekly, and feed and water consumption was measured
    over three-or four-day periods throughout the study. On day 27, the
    rats were transferred to metabolism cages and urine was collected
    during an 18-h fast prior to necropsy on day 28. Blood was collected
    to allow measurement of prothrombin time and activated partial
    thromboplastin time, in addition to haematological and clinical
    biochemical parameters. The animals were then killed and the brain,
    pituitary gland, thyroid glands, salivary glands, thymus, lungs,
    bronchi, heart, liver, spleen, adrenals, kidneys, testes, seminal
    vesicles, and prostate were weighed. A complete macroscopic
    examination was performed, and the tissues and organs listed above
    plus 23 additional ones were prepared for microscopic examination.

         There were no unscheduled deaths. A number of effects were
    associated with the nephrectomy which were independent of erythritol
    treatment and included lower feed consumption and body weights than in
    sham-operated rats. Irrespective of erythritol intake, nephrectomized
    rats had greater water consumption than the sham-operated group and an
    increase in urine volume, decreased specific gravity and osmotic
    pressure, and increased excretion of sodium, potassium, and chlorine.

    Marked amounts of protein were present in the urine of all
    nephrectomized rats. These rats had lower erythrocyte and platelet
    counts and higher mean corpuscular volume and reticulocyte and
    leukocyte counts than sham-operated rats. Nephrectomized rats also had
    shorter prothrombin and activated partial thromboplastin times,
    despite the reduced platelet count. They also had higher GGT and
    alpha-amylase activities, and their serum concentrations of total
    cholesterol, triglycerides, phospholipids, bilirubin, blood urea
    nitrogen, inorganic phosphate and alpha-globulin were all higher and 
    their albumin concentrations lower than those of sham-operated rats. 
    The absolute and relative weights of the heart, kidney, and adrenals 
    were significantly higher and the absolute and relative weights of the
    liver and seminal vesicle were significantly lower in nephrectomized
    than sham-operated rats.

         There were no clinical signs of toxicity, and no significant
    effects on body weights, body-weight gain, or feed consumption were
    associated with treatment in either the nephrectomized or
    sham-operated groups. The mean water intake was reported to be
    significantly higher ( p value not specified) in nephrectomized and
    sham-operated rats given 5% erythritol than in concurrent controls,
    and dietary erythritol resulted in increased erythrocyte, leukocyte,
    and platelet counts and in haemoglobin and haematocrit than in their
    respective controls, but only the differences in the sham-operated
    group reached statistical significance. The prothrombin and activated
    partial thromboplastin times tended to increase with increasing dose
    of erythritol in the sham-operated but not the nephrectomized rats.
    Administration of erythritol was associated with a decrease in serum
    triglyceride concentration in both operated groups, while total
    cholesterol and phospholipids were unaffected. A dose-related decrease
    in albumin and an increase in alpha-globulin were observed in both
    groups of erythritol-treated rats, resulting in a dose-related
    decrease in the albumin: globulin ratio. In the sham-operated group,
    chlorine excretion was significantly higher in rats given 5%
    erythritol than in controls, reaching the values for this parameter in
    nephrectomized rats. Erythritol had no significant effect on absolute
    or relative organ weights in either operated group. There were no
    histopathological changes attributable to erythritol treatment.

         The NOEL for both sham-operated and nephrectomized rats was 2% in
    the diet, equal to 1 g/kg bw per day, on the basis of increased water
    consump-tion in both groups, increased erythrocyte and platelet
    counts, increased haemaglobin and haematocrit (sham-operated),
    decreased serum albumin and albumin:globulin ratio, and increased
    serum alpha-1 globulin (nephrectomized) at 5% erythritol (Kanai et
    al., 1992).

    2.3  Observations in humans

    2.3.1  Single doses

         In a study to establish the minimum single dose that causes the
    transient diarrhoea seen after ingestion of erythritol, six male
    volunteers aged 26-46 years took single doses of sucrose (60 g),
    erythritol (30, 40, 50, and 60 g), and sorbitol (10 g) on separate
    days. Each substance was mixed into about 180 ml of water or weak
    coffee and ingested within 10 min. Sucrose was taken first, followed
    one to two days later by 30 g of erythritol; the dose of erythritol
    was then increased at one-to two-day intervals until diarrhoea was
    observed. Three to four days after diarrhoea was observed, 10 g of
    sorbitol were tested as a positive control. A questionnaire was
    completed before the beginning of the study to establish each
    subject's normal gastrointestinal function and excretion pattern. The
    day before consumption of each test substance, the participants were
    requested to avoid foods and drinks that induce diarrhoea. Before each
    test, the participants completed a questionnaire eliciting information
    on the previous day's food consumption and excretion pattern.
    Questionnaires were also used after ingestion of each substance to
    follow gastrointestinal symptoms and bowel movements for the next
    24 h.

         All subjects completed the study, although all complained of the
    sweetness of both sucrose and erythritol. The average erythritol
    intakes were calculated on the basis of body weight to be 0.46, 0.62,
    0.77, and 0.92 g/kg bw of the 30-, 40-, 50-, and 60-g doses
    respectively. The sucrose and sorbitol doses corresponded to intakes
    of 0.92 g/kg bw and 0.15 g/kg bw, respectively. None of the
    participants reported having diarrhoea after ingestion of 60 g of
    sucrose or 30 g of erythritol, but ingestion of 40, 50, and 60 g of
    erythritol induced diarrhoea in two participants in each group. All
    subjects experienced diarrhoea after consuming 10 g of sorbitol. The
    symptoms associated with the diarrhoea, including abdominal pain
    and/or borborygmus, were similar after erythritol and sorbitol
    treatment. Diarrhoea occurred 2-12 h after ingestion, and the subjects
    recovered within 24 h. Erythritol was also associated with thirst and
    headache, although the report does not indicate the dose at which
    these symptoms occurred. The NOEL for the diarrhoetic effect of
    erythritol was 30 g or 0.46 g/kg bw (Umeki, 1992).

         The study was repeated with 12 healthy volunteers (eight men and
    four women), at doses of 30, 40, and 50 g. Three men did not complete
    the sugar test because of pain [ sic] associated with the sweetness
    of the ingestion mixture. In addition, one man did not complete the
    test with 50 g erythritol and one man and three women did not complete
    the test with sorbitol, for unexplained reasons. The mean intakes of
    erythritol were 0.47, 0.62, and 0.78 g/kg bw for men and 0.57, 0.76,
    and 0.94 g/kg bw for women consuming 30, 40, and 50 g of erythritol,
    respectively. The mean intakes of sugar and sorbitol were 0.94 and
    0.16 g/kg bw for men and 1.1 and 0.19 g/kg bw for women, respectively.

    No diarrhoea occurred in the five men and four women who consumed
    sugar or in the 12 subjects who consumed 30 g of erythritol. One of
    eight men reported diarrhoea after taking 40 g of erythritol, and two
    men and one woman reported diarrhoea after taking 50 g. Some subjects
    reported abdominal symptoms after consumption of 40 or 50 g of
    erythritol, but the number of affected subjects and the nature of the
    symptoms were not specified. Three of seven men experienced diarrhoea
    and abdominal symptoms within 3 h of taking sorbitol, but the symptoms
    had resolved within 24 h. The NOEL for the diarrhoetic effect of an
    acute dose of erythritol was 30 g or 0.47 g/kg bw (Takahashi, 1992a).

         The effect of a single oral dose of erythritol on serum glucose
    and insulin concentrations was investigated in five healthy male
    volunteers aged 45-58 years and weighing 54-65 kg. After a 12-h fast,
    the subjects received a single oral dose of 0.3 g/kg bw of erythritol
    as a 20% aqueous solution. One week later, they received the same dose
    of a 20% glucose solution, again after a 12-h fast. The treatments
    were administered at 9:30 h, and after 12:30 h the subjects were
    permitted to eat or drink freely except for wine and fermented foods
    containing erythritol. Urine samples were collected 0-3, 3-8, 8-24,
    and 24-48 h after dosing for determination of osmotic pressure,
    sodium, potassium, and chlorine. Blood samples were collected 0.5, 1,
    2, 3, 8, and 24 h after dosing to allow determination of the serum
    concentrations of insulin, glucose, total cholesterol, triglycerides,
    free fatty acids, sodium, potassium, and chlorine. Erythritol had no
    significant effect on serum glucose or insulin concentrations. After
    glucose administration, the serum glucose and insulin concentrations
    both rose rapidly up to 30 min after dosing, and the peak glucose and
    insulin concentrations after ingestion of glucose were significantly
    higher than those after erythritol ingestion. The serum concentrations
    of total cholesterol and triglycerides after glucose and erythritol
    ingestion did not differ significantly during the subsequent 24 h. The
    serum free fatty acid concentrations differed after the two treatments
    throughout the study but were within normal limits, whereas those of
    sodium, potassium, and chlorine and urine volumes were comparable
    after erythritol and glucose ingestion. The urinary osmotic pressure
    was higher 3 h after administration of erythritol than after glucose,
    but the difference was not statistically significant. The urinary
    excretion of sodium, potassium, and chlorine over 48 h after
    administration was comparable in the two groups (Noda et al., 1994).

         A study of the metabolic effect of a single oral dose of
    erythritol on plasma glucose and insulin concentrations involved three
    healthy male and three healthy female volunteers aged 24-43 years with
    an average body mass index of 22 ± 1.65. After an overnight fast, all
    subjects ingested a single oral dose of erythritol at 1 g/kg bw
    dissolved in 250 ml of water. Blood samples were collected for
    analysis of plasma glucose and insulin concentrations 15 min before
    administration and every 30 min up to 3 h after administration, and a
    medical examination was conducted 24 h after administration. Three
    women and one man reported gastrointestinal symptoms after consumption
    of erythritol. While two of the women reported diarrhoea, the others

    reported abdominal cramping, discomfort, and flatulence. Neither
    plasma glucose nor plasma insulin was affected (Bornet et al., 1996a).

         In a further study, 24 healthy volunteers (12 men and 12 women),
    aged 20-46 years, were assigned, irrespective of sex, to one of four
    treatment groups so that the groups were homogeneous with respect to
    body mass index. All of them received a set breakfast at 8:00 h and a
    lunch at 12:45 h. Groups of six subjects received a mid-morning snack
    at 10:00 h which consisted of chocolate containing erythritol at 0.4
    or 0.8 g/kg bw or sucrose at 0.8 g/kg bw. A further group received no
    snack. Throughout the test, subjects were given drinking-water
     ad libitum, and the total volume ingested was recorded. Before the
    first meal was taken, an indwelling venous catheter was fitted into
    each subject and left in place until 18:00 h, and blood samples were
    collected hourly between 8:00 and 18:00 h for analysis of glucose,
    insulin, electrolytes, creatinine, osmolarity, haematocrit, albumin,
    and erythritol. Urine was collected at 2-h intervals between 8:00 and
    18:00 h and then between 18:00 and 8:00 h the following morning. The
    urine samples were analysed for creatinine, electrolytes, osmolarity,
    erythritol, and NAG. Satiety was evaluated before each meal on an
    arbitrary scale upon which the subjects recorded their sensation of
    hunger, ranging from none (0) to extreme (100). The occurrence of
    digestive complaints was evaluated on the day after the test from the
    responses to a questionnaire.

         The perception of hunger was comparable in all groups that
    received the mid-morning snack, but before lunch the perception of
    hunger was significantly ( p < 0.005) greater in the group that did
    not receive a snack than in those that did. Those given erythritol
    reported more gastrointestinal effects (number of subjects not
    specified), but none of the differences from controls was
    statistically significant. The water consumption and urine volumes
    during the 22 h after treatment were comparable to those of negative
    controls, whereas the urine volumes of subjects given sucrose were
    significantly greater than those of the negative controls and those
    receiving the high dose of erythritol ( p < 0.05) 8-22 h after
    administration, with no increase in water consumption. The plasma
    glucose and insulin concentrations of subjects who received erythritol
    were comparable to those of subjects who received no mid-morning
    snack, while the insulin concentrations in people given sucrose were
    significantly ( p < 0.05) higher than those of negative controls 1
    and 2 h after eating the snack, and the plasma glucose concentrations
    remained within normal ranges. Erythritol had no effect on plasma
    osmolarity or calcium concentration. Urinary excretion of sodium and
    chloride was increased in the groups given erythritol and sucrose over
    that in the negative control group during the first 6 and 4 h after
    treatment, respectively, but the total excretion of sodium and
    chlorine over the collection period was comparable in the two groups,
    and both had significantly higher baseline sodium ion excretion than
    negative controls. Urinary osmolar excretion was significantly
    ( p < 0.005) higher in the group given the high dose of erythritol

    between 2 and 8 h after administration than in the groups given no
    treatment or sucrose. Urinary osmolar excretion in the group given the
    low dose of erythritol also tended to be increased, but differed
    significantly ( p < 0.005) from that of volunteers given sucrose
    only during the 4-6-h collection period. Urinary excretion of NAG
    during the study was not significantly affected by erythritol (Bornet
    et al., 1996b).

         A study to investigate the effect of a single oral dose of
    erythritol on glucose tolerance and to examine the concentrations of
    erythritol in the blood and urine of patients with
    non-insulin-dependent diabetes mellitus (NIDDM) involved five patients
    (sex not indicated), of an average age of 52 ± 19 years. All of the
    patients relied only on diet as treatment, and none showed signs of
    hepatic dysfunction. After a fast, the patients drank a solution of 20
    g of erythritol in 100 ml of water at 9:00 h. Blood samples were taken
    before and 0.5, 1, 1.5, 2, 3, 4, 6, 8, and 24 h after erythritol
    administration. The patients were allowed to eat food 3 h after
    dosing. Urine was collected 24 h before and from 0-24, 24-48, and
    48-72 h after consumption of erythritol. The blood samples were
    analysed for glucose, insulin, free fatty acids, 3-hydroxybutyric
    acid, and erythritol, and the urine was analysed for erythritol.

         No adverse effects were reported after consumption of erythritol.
    No data on urine volumes were reported, but the blood glucose and
    insulin concentrations were comparable to those before administration
    up to 3 h after ingestion of erythritol. The concentrations of free
    fatty acids and 3-hydroxybutyric acid in blood increased after
    erythritol treatment but decreased rapidly after ingestion of food 3 h
    later. The increases seen after treatment probably occurred because
    erythritol provides no energy and the subjects remained in a fasting
    metabolic state (Ishikawa  et al. 1996).

         A study designed to estimate the maximum ineffective dose of
    erythritol and the median effective dose (ED50) for diarrhoea
    involved a group of seven men who ingested increasing amounts of
    erythritol in jelly at doses of 25, 50, and 75 g (men) and 12 women
    who ingested 25, 37.5, and 62.5 g at one-to two-day intervals until
    diarrhoea was induced or the highest dose was reached. When their
    gastrointestinal condition had normalized, the participants consumed
    increasing amounts of sorbitol at doses of 12.5 and 25 g in jelly as a
    positive control. Sucrose, used as a negative control, was
    administered in jelly at the dose at which erythritol induced
    diarrhoea. Six women also participated in a supplementary study
    designed to determine if the delivery vehicle affected the incidence
    of diarrhoea in which each consumed 50 and 75 g of erythritol in jelly
    and in an aqueous solution. On the day before each test, the
    participants were requested to avoid food and beverages that might
    induce diarrhoea and to complete a questionnaire on their food intake.
    On the day of administration, the jelly containing the test material
    was given with water 2-3 h after lunch or breakfast. The participants

    were then permitted to consume any food or drink except wine and
    fermented foods containing erythritol or sorbitol for about 2 h after
    consuming the test material. A questionnaire was used during the day
    of administration to record the time of onset of abdominal symptoms
    and diarrhoea, the type of abdominal symptoms, the frequency of
    defaecation, and the macroscopic appearance of the stools. Blood
    samples were collected from all participants before the beginning of
    the study and two to three days after completion for clinical
    chemistry and haematology. All the women participated in a food intake
    survey on the day of administration to allow determination of nutrient
    intakes during each test.

         The abdominal symptoms reported most frequently by female
    participants who ingested erythritol were nausea, borborygmus, thirst,
    and flatulence, while the men most frequently reported borborygmus and
    thirst. After sorbitol ingestion, the women reported symptoms of
    flatulence and borborygmus, while most of the men reported no
    abdominal symptoms. Table 3 shows results of determinations of the
    no-effect concentrations and ED50 values for the diarrhoetic effects
    of erythritol and sorbitol in men and women. Doses of sucrose up to
    1.2 g/kg bw did not induce diarrhoea in any of the participants, and
    the delivery vehicle had no effect on the incidence of diarrhoea.

         The results indicate that while women more frequently reported
    gastro-intestinal symptoms after ingesting erythritol, they were less
    sensitive to its diarrhoetic effect. Although a lower dose of sorbitol
    than erythritol was required to induce diarrhoea in both men and
    women, abdominal symptoms were more frequently associated with
    erythritol (Oku & Okazaki, 1996a,b). 

    Table 3. No-observed-effect levels and median
    effective concentrations (ED50) of erythritol
    in volunteers
                                                    

    Concentration            Men            Women
                                                    

    NOEL (g/kg bw per day)

      Erythritol             0.66           0.80
      Sorbitol               0.17           0.24

    ED50 (g/kg bw per day)

      Erythritol             1.1            1.6
      Sorbitol               0.21           0.52
                                                    

    2.3.2  Repeated doses

         In a study of the gastrointestinal effects of erythritol in
    healthy adults during short-term exposure, eight man and two women
    ingested 20 g of erythritol dissolved in 150-180 ml of cold water
    twice daily (40 g/day) 2-3 h after breakfast and lunch for five days.
    The men were 27-59 years old (mean, 45 years) and weighed 56-71 kg
    (mean, 63 kg), and the women were 48 and 63 years old and both weighed
    54 kg. The participants were requested to avoid foods and drinks that
    were likely to cause diarrhoea the day before the beginning of the
    study, but there were no restrictions on food or beverage consumption
    during the study, and the participants were encouraged to maintain
    their normal food intakes and daily activities. After each dose, the
    volunteers completed a questionnaire about their abdominal condition,
    including pain, discomfort, gas, distension, and the time and
    frequency of evacuation and/or diarrhoea.

         No laxative effects were observed, but the response of one
    participant suggests that people who are hypersensitive to sugar
    alcohols could have diarrhoea at this dose. One man complained of
    upper gastrointestinal pain and diarrhoea on the second day of the
    study, but no other symptoms were reported (Takahashi, 1992b).

         The gastrointestinal tolerance of healthy adults to repeated
    ingestion of a coffee drink containing erythritol was evaluated in six
    male volunteers aged 30-53 years and with a mean body weight of 74 kg.
    The men drank five tins of coffee drink, one after each meal and two
    between meals, for three consecutive days, for a total quantity of
    erythritol ingested of 68 g and an average intake of 0.91 g/kg bw per
    day. Each man completed a questionnaire on his general condition,
    including stool consistency and abdominal condition after ingestion.
    Effects on urine volume were not evaluated. No changes in stool
    consistency or on the frequency of defaecation were reported, and no
    abdominal symptoms were observed. Three of the men reported dryness or
    irritation of the throat and/or stomach, which was resolved by
    drinking water. The author suggested that these symptoms were due to
    the high osmotic activity of the coffee drink (Hamada, 1996).

         In a study of the gastrointestinal tolerance of healthy adults to
    repeated ingestion of a tea drink containing erythritol, eight male
    volunteers aged 30-53 years, with a mean body weight of 70 kg,
    consumed five tins of tea drink, one after each meal and two between
    meals, for three consecutive days, for a total of 60 g erythritol and
    an average intake of 0.86 g/kg bw per day. Each subject completed a
    questionnaire on his general condition, including stool consistency
    and abdominal condition after ingestion. Effects on urine volume were
    not evaluated. No changes in stool consistency or on the frequency of
    defaecation were reported, and no adverse effects on gastrointestinal
    condition or general health were reported (Masuyama, 1996).

         The gastrointestinal tolerance and diuretic response to repeated
    doses of erythritol were examined in a double-blind, two-way
    cross-over study in which healthy men consumed erythritol and, for
    comparison, sucrose. Twelve men aged 22-46 and weighing 65-98 kg
    consumed each test material for seven days, comprising a two-day
    adaption period at home and a five-day test period under supervision.
    The men consumed 0.3 g/kg bw erythritol or 0.6 g/kg bw sucrose in test
    foods during the first and second day of adaptation, respectively, and
    1 g/kg bw per day of erythritol or sucrose in yoghurt, biscuits, soft
    drinks, or chocolate under supervision. In order to attain the correct
    dose, fixed amounts were given in test foods at each meal except the
    evening meal, when individual doses of the test compounds were
    administered in order to reach the nominal dose of 1 g/kg bw for each
    subject. Beverages such as mineral water and fruit juice were allowed
     ad libitum, but consumption of caffeine-containing drinks was
    limited to four cups per day. The men were asked to abstain from
    excessive alcohol intake during the adaptation periods and to abstain
    completely during the test period. During the test period, the
    following data were recorded each day: sensory perception of test
    foods; overall well-being, feelings of hunger and thirst, desire for
    sweet or salty foods, and subjective perception of regularity,
    consistency, and quantity of stools, frequency and quantity of urine,
    gastrointestinal intolerance, and other side-effects; fluid intake;
    food intake; body weight; and blood pressure while seated. The entire
    urine volume was collected at five 3-h intervals during the day and
    one 9-h interval overnight throughout the test period, and sodium,
    potassium, chlorine, calcium, phosphorus, citrate, GGT and NAG
    activities, ß2u-globulin, urea, and creatinine were determined. The
    concentration of erythritol was determined in a separate aliquot.

         The daily dose 1 g/kg bw per day erythritol was well tolerated,
    with no increase in the reported incidence of gastrointestinal
    symptoms such as flatulence, abdominal cramps, and diarrhoea. One man
    reported thirst during treatment with erythritol, but there were no
    differences in the subjective judgements of the frequency and quantity
    of urine production. Fluid intake varied considerably among the study
    participants, from 800 to 6600 ml/man per day, and the averages for
    both treatment periods were reported to be high in comparison to usual
    intakes, but there was no significant difference between the two
    periods. Urine production was about 7% higher during erythritol
    treatment, but the increase was not statistically significant. Urinary
    excretion of erythritol resulted in significantly ( p < 0.001)
    increased urine osmolarity and hourly output of osmotically active
    solutes, but urinary pH and the excretion of creatinine, urea,
    citrate, sodium, potassium, and chlorine were not affected. Marginal
    but statistically significant increases in calcium concentration,
    µ-albumin, ß2u-globulin, and NAG activity were noted consistently
    over the five-day period after ingestion of erythritol, although the
    values for these parameters remained within reference intervals and

    below values that would be considered clinically relevant. The data on
    GGT activity were considered unreliable and were therefore not
    reported.

         Thus, 1 g/kg bw of erythritol was consumed as part of the regular
    diet over five days without adverse gastrointestinal symptoms. Urine
    volume and urinary electrolyte and protein excretion were not
    significantly affected at this dose, suggesting the absence of a
    diuretic effect, although the high fluid consumption of the
    participants and the consumption of coffee and/or tea (up to four cups
    per day) may have prevented the detection of any diuretic effect of
    erythritol (Tetzloff et al., 1996).

         In order to investigate the effects of repeated doses of
    erythritol on blood glucose control and renal function in patients
    with NIDDM, three male out-patients (mean age, 65 ± 6 years) and eight
    female out-patients (mean age, 50 ± 14 years) consumed 20 g of
    erythritol in a solution throughout the day daily for 14 days with
    their usual diet, but without specific restriction on timing or
    division of the daily dose. Food intake was monitored for three days
    before the test and three days before the end of the test. Body
    weights before and after administration and blood glucose and
    haemoglobin A1c after fasting were determined in all participants as
    indices of control of diabetes. In four or five of the subjects, blood
    urea nitrogen, creatinine, ß2u-globulin, and urinary proteins (not
    specified) were measured as indices of renal function before and after
    erythritol treatment.

         None of the participants reported diarrhoea or any other
    subjective symptoms during treatment. The blood glucose concentrations
    after fasting, reported for nine subjects (sex not specified),
    decreased from 181 ± 60 mg/dL before administration to 165 ± 57 mg/dL
    after administration, which was not significant. The haemoglobin A1c
    concentrations after fasting, reported for all 11 participants, were
    the same as those before treatment for four, decreased in six, and
    increased in one subject after erythritol treatment. The large
    decreases in two subjects resulted in a decrease in the group mean
    value after treatment, to 7.5 ± 1.6% from 8.5 ± 1.5% before
    administration. Blood urea nitrogen, creatinine, and ß2u-globulin
    values were reported for only five or fewer subjects, but erythritol
    had no effect on these parameters. Urinary proteins, as measured
    colorimetrically for an unspecified number of participants, were
    reported not to be affected by erythritol treatment. The study is of
    limited value since data were not collected and/or reported for all
    study participants. Although the authors concluded that administration
    of erythritol daily for 14 days did not adversely affect blood glucose
    concentrations or renal function in patients with NIDDM, no conclusion
    can be drawn (Miyashita et al., 1993; Ishikawa et al., 1996).

    3.  COMMENTS

         Studies of the disposition of erythritol in mice, rats, dogs, and
    humans showed that ingested erythritol was rapidly and extensively
    absorbed and rapidly excreted in the urine, without undergoing
    metabolism. Excretion in the faeces was a minor route after dietary
    administration to mice, rats, and dogs; no data were available for
    humans. The small but significant proportion of the administered dose
    recovered in expired carbon dioxide after oral administration was
    probably the result of fermentation of erythritol in the lower
    gastrointestinal tract, and the amount increased with increasing dose.
    In contrast, the major route of excretion of orally administered
    glycerol, lactitol, and mannitol was expired carbon dioxide,
    negligible amounts being excreted unchanged in the urine and faeces.
    These findings indicate the importance of gastrointestinal
    fermentation in the disposition of these polyols. There was no
    evidence of fermentation of erythritol by gastrointestinal flora in
    people who had not been exposed to it previously.

         Erythritol had little toxicity when administered orally to mice,
    rats, and dogs as a single dose. The range of symptoms observed in
    animals that subsequently died were considered to be non-specific
    effects of large amounts of a hypertonic solution.

         The toxicity of erythritol was studied in mice given the compound
    in the diet for 13 weeks, in rats treated for 28 days in the diet (two
    studies) or for 13 weeks in the diet or by gavage, and in dogs treated
    by gavage for 13 weeks or in the diet for one year. In all of these
    studies, doses representing up to 20% of the diet were used. In
    rodents of each sex, administration of erythritol was accompanied by
    dose-related increases in water consumption and urine volume. Urine
    density and osmotic pressure were increased at the lower doses and
    decreased at the higher doses, reflecting the competing factors of
    high concentrations of erythritol and its effects on diuresis. Urinary
    excretion of electrolytes, particularly sodium, potassium, and calcium
    (measured only with dietary administration), and of protein was also
    increased in rats and mice. Increased kidney weights were observed in
    rats but not mice. In a study in which erythritol was administered for
    four weeks in the diet of rats that had undergone partial nephrectomy,
    no difference in response was seen between sham-operated and
    nephrectomized animals. Other effects related to diuresis were seen in
    response to erythritol only in the 13-week study in rats treated by
    gavage; these included increased blood urea nitrogen concentration,
    decreased sodium and chloride ion content of serum, an increased
    incidence of slight dilatation of the renal tubules, and increased
    adrenal weights accompanied by dilatation of the sinusoids of the
    adrenal cortex. These effects were no longer seen after a four-week
    recovery period. The results of a supplementary study suggested that
    the increase in blood urea nitrogen concentration was a compensatory
    homeostatic response to serum hyponatraemia. The more extensive

    effects noted after gavage were probably due to achievement of higher
    maximum plasma concentrations of erythritol after a bolus dose than
    after gradual intake in the diet.

         Gastrointestinal effects were seen in all of the studies in which
    erythritol was administered orally. These included transient laxation
    and soft stools in rats and increased caecal weights in both rats and
    mice. The decrease in caecal pH in the 13-week study in rats treated
    in the diet would have promoted increased absorption of calcium and
    might therefore account for the increased urinary excretion of calcium
    ion. Serum alkaline phosphatase activity was increased by treatment in
    these studies. Since the main source of circulating alkaline
    phosphatase in rats is the intestine, the increase in plasma activity
    may have resulted from the intestinal effects of erythritol.

         In dogs, transient clinical effects (salivation, vomiting,
    reddening of the epidermal and mucous membranes, laxation, and soft
    stools) were seen after treatment by gavage but not after dietary
    administration. With the exception of the effect on faecal
    consistency, these were attributed to increased plasma osmolality. As
    in the rodents, water consumption and urine volume were increased in
    both studies, with increased osmotic pressure and specific gravity of
    the urine observed at the lower doses and decreases in these
    parameters at the higher doses. Renal function, as assessed by
    clearance of phenolsulfonphthalein after treatment by gavage, was not
    affected. In the one-year study, some histopathological changes were
    seen in the kidneys of two of the three dogs at the high dose, which
    regressed during the four-week recovery period and were considered to
    be a transient functional osmotic response. The changes in the
    prostate were not considered to be toxicologically relevant.

         All of the effects seen in these short-term studies in rodents
    and dogs were considered to be physiological or adaptive responses to
    the osmotic diuretic effects of absorbed erythritol, or, in rodents,
    the effect of gastrointestinal fermentation of unabsorbed erythritol.
    All of these effects were reversed when feeding of erythritol ceased.
    Intravenous administration resulted in physiological changes that were
    qualitatively similar to those observed after dietary or gavage
    administration but were more marked. The NOELs in the studies of
    dietary administration were 5% of the diet, equivalent to 7.5 g/kg bw
    per day in the 13-week study in mice, 2.5 g/kg bw per day in the
    13-week study in rats, and 1.7 g/kg bw per day in the 53-week study in
    dogs. The NOELs in the studies of administration by gavage for 13
    weeks were 2 g/kg bw per day for rats and 1.2 g/kg bw day for dogs.

         Erythritol was not carcinogenic in rats treated in the diet; no
    long-term study of toxicity and carcinogenicity in mice was available.
    Effects similar to those seen in the short-term studies were observed,
    with the addition of earlier onset of nephrosis in males at the high
    dose. The NOELs for the physiological responses to erythritol were 3%
    and 2% of the diet, equal to 1.4 and 0.9 g/kg bw per day, in the
    78-and 104-week studies, respectively.

         Erythritol did not exhibit mutagenic or clastogenic activity
     in vitro.

         No reproductive or developmental toxicity was observed at doses
    up to 8 g/kg bw per day in mice treated by gavage or at doses
    representing up to 10% of the diet of rats.

         The gastrointestinal and renal effects of erythritol and its
    effects on glucose control in humans have been studied in volunteers.
    When single doses of 30-75 g of erythritol were administered in
    solution or jelly to healthy adults in three studies, the NOEL for
    induction of laxation was 0.46-0.66 g/kg bw in men and 0.76-0.80 g/kg
    bw in women. The dose that induced laxation in 50% of the subjects was
    estimated to be 1.1 g/kg bw in men and 1.6 g/kg bw in women. Although
    women appeared to be less sensitive to erythritol-induced laxation,
    they more frequently reported gastrointestinal symptoms, including
    nausea, than the men. When a divided dose of 40 g of erythritol in
    solution was ingested by healthy individuals for five days or divided
    doses of 60-68 g/day were ingested in tea or coffee for three days, no
    laxation occurred at doses up to 0.91 g/kg bw per day in men and 0.74
    g/kg bw per day in women. Single doses of 0.3 and 1 g/kg bw erythritol
    in aqueous solution given to healthy adults had no effect on plasma
    glucose or insulin concentrations, and neither urine volume nor
    urinary excretion of sodium, chloride, or potassium ions was affected
    by the lower dose.

         No gastrointestinal symptoms were seen when single doses of 0.4
    or 0.8 g/kg bw given as a snack or repeated doses of 1 g/kg bw per day
    in a variety of foods spread over the day for five days were ingested
    by healthy individuals. Urine osmolarity was increased with erythritol
    treatment, proportionally to dose, in both of these studies. After
    single doses of erythritol, no changes in measured plasma osmolarity
    were observed, and there was no effect on water consumption, urine
    volume, or 24-h excretion of electrolytes or
     N-acetyl-glucosaminidase. Glucose control was not affected.

         In patients with non-insulin-dependent diabetes, a 20-g dose of
    erythritol in solution consumed on one day or for 14 consecutive days
    did not induce laxation and had no effect on glucose control.

    4.  EVALUATION

         The NOELs for physiological responses to orally administered
    erythritol in animals were mostly between 1 and 2 g/kg bw per day.
    Since the observed effects of erythritol in animals are a
    physiological response to an osmotically active substance, application
    of a safety factor to the NOELs observed in studies in experimental
    animals was considered inappropriate. In humans, a dose of 1 g/kg bw
    per day consumed in a variety of foods for five days was without

    effect, although the same and lower doses consumed in aqueous solution
    as a bolus dose after fasting resulted in laxation. The Committee
    established an ADI 'not specified'1 for erythritol for use as a
    sweetening agent.

    5. REFERENCES

    Beck, F.F., Carr, C.J. & Krantz, J.C., Jr (1936) Acute toxicity of
    certain sugar alcohols and their anhydrides.  Proc. Soc. Exp. Biol.
     Med., 35, 98-99.

    Beck, F.F., Carr, C.J. & Krantz, C.J., Jr (1938) The fate of
    erythritol and erythritan in the animal body.  Q. J. Pharm.
     Pharmacol., 11, 234-239.

    Blijleven, W.G.J. (1990) Examination of erythritol for mutagenic
    activity in the Ames test (TNO Report No. V90.306), TNO-CIVO
    Industries, Zeist, Netherlands Organization for Applied Scientific
    Research.

    Bornet, F.R.J., Blayo, A., Dauchy, F. & Slama, G. (1996a) Plasma and
    urine kinetics of erythritol after oral ingestion by healthy humans.
     Regul. Toxicol. Pharmacol., 24, S280-S285.

    Bornet, F.R.J., Blayo, A., Dauchy, F. & Slama, G. (1996b)
    Gastrointestinal response and plasma and urine determinations in human
    subjects given erythritol.  Regul. Toxicol. Pharmacol., 24, 
    S296-S302.

    Dean, I. & Jackson, F. (1992) Erythritol: One year oral (dietary)
    toxicity study in dogs (IRI Report No. 7900), Tranent, Scotland,
    Inveresk Research International.

    Dean, I., Jackson, F. & Greenough, R.J. (1996) Chronic (1-year) oral
    toxicity study of erythritol in dogs.  Regul. Toxicol. Pharmacol.,
    24, S254-S260.

    Dziegielewska, K.M., Evans, C.A.N., Malinowska, D.H., Mollgård, K.,
    Reynolds, J.M., Reynolds, M.L. & Saunders, N.R. (1979) Studies of the
    development of brain barrier systems to lipid insoluble molecules in
    fetal sheep.  J. Physiol., 292, 207-231.

                    

    1 ADI 'not specified' is a term applicable to a food component of
    very low toxicity which, on the basis of the available chemical,
    biological, toxicological, and other data, the total dietary intake of
    the substance arising from its use at the levels necessary to achieve
    the desired effect and from its acceptable background in food, does
    not, in the opinion of the Committee, represent a hazard to health.
    For this reason and for those stated in the evaluation, the
    establishment of an ADI expressed in numerical form is deemed
    unnecessary.

    Hamada, A. (1996) Effects of repetitive ingestion of coffee drink
    containing erythritol with laxative effect as the measure. Unpublished
    report from Nikken Chemicals Co., Ltd.

    Hashima Laboratory (1989) Teratology study of NIK-242 in rabbits
    (intravenous dosing), seg. II, Nihon Bioresearch Inc.

    Hiele, M., Ghoos, Y., Rutgeerts, P. & Vantrappen, G. (1993) Metabolism
    of erythritol in humans: Comparison with glucose and lactitol. 
     Br. J. Nutr., 69, 169-176.

    Ishikawa, M., Hirose, C., Tsujino, D., Miyashita, M., Kawashima, Y. &
    Nakamura, T. (1992) The effect of erythritol on glucose tolerance in
    diabetes patients. Unpublished report from Yokohama-shi Seibu
    Hospital, St Marianna University School of Medicine, Department of
    Metabolic Endocrinology and Department of Nutrition.

    Ishikawa, M., Miyashita, M., Kawashima, Y., Nakamura, T., Saito, N. &
    Modderman, J. (1996) Effects of oral administration of erythritol on
    patients with diabetes.  Regul. Toxicol. Pharmacol., 24, S303-S308.

    Kamata, S. (1990a) A 6 month intravenous chronic toxicity study of
    NIK-242 in rats with 1 month recovery period: Final report.
    Unpublished report from Safety Research Institute for Chemical
    Compounds Co., Ltd.

    Kamata, S. (1990b) A 6 month intravenous chronic toxicity study of
    NIK-242 in beagle dogs with 1 month recovery period. Unpublished
    report from Safety Research Institute for Chemical Compounds Co., Ltd.

    Kanai, M., Yamamoto, H., Takahashi, T., Onishi, T. & Shigeki, Y.
    (1992) A 4-week feeding toxicity study on erythritol in rats with
    reduced renal function. Unpublished report from Division of
    Toxicology, Ohmiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Kawamura, Y., Saito, Y., Imamura, M. & Modderman, J.P. (1996)
    Mutagenicity studies on erythritol in bacterial reversion assay
    systems and in Chinese hamster fibroblast cells.  Regul. Toxicol.
     Pharmacol., 24, S261-S263.

    Lina, B.A.R., Bos-Kuijpers, M.H.M. & Til, H.P. (1994) Chronic (2-year)
    oral toxicity and carcinogenicity study with erythritol in rats (TNO
    Report No. V93.059). Unpublished report from TNO-CIVO Industries,
    Netherlands Organization for Applied Scientific Research, Zeist.

    Lina, B.A.R., Bos-Kuijpers, M.H.M., Til, H.P. & Bär, A. (1996) Chronic
    toxicity and carcinogenicity study of erythritol in rats.  Regul.
     Toxicol. Pharmacol., 24,

    Masuyama, I. (1996) Effect of repetitive ingestion of tea drink
    containing erythritol with laxative effect as a measure. Unpublished
    report from Nikken Chemicals Co., Ltd.

    Matsuzawa, T., Nomura, M. & Unno, T. (1993) Clinical pathology
    reference ranges of laboratory animals.  J. Vet. Med. Sci., 55,
    351-362.

    Miyashita, M., Kawashima, Y., Nakamura, T., Ishikawa, M., Tsujino, D.
    & Someya, K. (1993) The effect of continuous administration of the
    sweetener erythritol on diabetes patients. Unpublished report from St
    Marianna University School of Medicine.

    Nakatsuru, S., Akiyama, S. & Saitou, S.A. (1988) Screening for
    mutagenicity of NIK-242 on reverse mutation test in bacteria and
    chromosome aberration test in cultured cells. Unpublished report from
    Division of Toxicology, Omiya Research Laboratory, Nikken Chemicals
    Co., Ltd.

    Nakayama, K. (1990a) Blood radioactivity levels after single oral
    administration of 14C-NIK-242 in rats. Unpublished report from
    Division of Metabolism, Omiya Research Laboratory, Nikken Chemicals
    Co.

    Nakayama, K. (1990b) Blood cell incorporation, protein binding and
    excretions in urine, feces, expired air and bile after single oral
    administration of 14C-NIK-242 in rats. Unpublished report from
    Division of Metabolism, Omiya Research Laboratory, Nikken Chemicals
    Co., Ltd.

    Nakayama, K. (1990c) Radioactivity distribution after single oral
    administration of 14C-NIK-242 in rats. Unpublished report from
    Division of Metabolism, Omiya Research Laboratory, Nikken Chemicals
    Co., Ltd.

    Nakayama, K. (1990d) Pharmacokinetics after single oral administration
    of NIK-242 to dogs. Unpublished report from Division of Metabolism,
    Omiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Noda, K. (1994) Analysis of metabolites in urine after single oral
    administration of NIK-242 to dogs. Unpublished report from Division of
    Metabolism, Omiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Noda, K. & Oku, T. (1990) The fate and availability of erythritol in
    rats. In: Hosoya, N., ed.,  International Symposium on Caloric
     Evaluation of Carbohydrates, Tokyo, Research Foundation for Sugar
    Metabolism, pp. 51-63.

    Noda, K. & Oku, T. (1992) Metabolism and disposition of erythritol
    after oral administration in rats.  J. Nutr., 122, 1266-1272.

    Noda, K., Nakayama, K. & Inoue, Y. (1988) Study on excretion of
    erythritol after oral administration to human subjects. Unpublished
    report from Division of Metabolism, Omiya Research Laboratory, Nikken
    Chemicals Co., Ltd.

    Noda, K., Nakayama, K. & Oku, T. (1994) Serum glucose and insulin
    levels and erythritol balance after oral administration of erythritol
    in healthy subjects.  Eur. J. Clin. Nutr., 48, 286-292.

    Noda, K., Nakayama, K. & Modderman, J. (1996) Fate of erythritol after
    single oral administration to rats and dogs.  Regul. Toxicol.
     Pharmacol., 24, S206-S213.

    Oku, T. & Noda, K. (1990) Influence of chronic ingestion of newly
    developed sweetener, erythritol on growth and gastrointestinal
    function of the rats.  Nutr. Res., 10, 987-996.

    Oku, T. & Okazaki, M. (1996a) Laxative threshold of sugar alcohol
    erythritol in human subjects.  Nutr. Res., 16, 577-589.

    Oku, T. and Okazaki, M. (1996b) Estimation of maximum non-effective
    dose for diarrhea induced by low energy bulking sweetener erythritol
    in normal human subjects. Unpublished report from Department of
    Nutrition, Faculty of Medicine, University of Tokyo, Tokyo.

    van Ommen, B. & de Bie, A.T.H.J. (1990) Disposition study with
    14C-erythritol in rats (TNO Report No. V90.307). Unpublished report
    from TNO-CIVO Industries, Netherlands Organization for Applied
    Scientific Research, Zeist.

    van Ommen, B., de Bie, B. & Bär, A. (1996) Disposition of
    14C-erythritol in germfree and convention rats.  Regul. Toxicol.
     Pharmacol., 24, S198-S205.

    Ota, T., Kato, M. & Nakagawa, K. (1990) Teratology study of NIK-242 in
    mice (intravenous dosing): Final report. Unpublished report from
    Hashima Research Laboratory, Nihon Bioresearch Center Inc., Division
    of Toxicology, Omiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Ozeki, M., Hirao, A., Araki, N., Yamaguchi, T., Kimura, H., Ito, K.,
    Shintani, S., Morita, K. & Kitamura, S. (1988) Acute oral toxicity
    study of NIK-242 in dogs. Unpublished report from Division of
    Toxicology, Omiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Shibata, M., Yamamoto, S., Takahashi, K., Kitamura, S. & Ichikawa, N.
    (1991) Study on increased BUN caused by repeated administration of
    erythritol in rats. Unpublished report from Division of Toxicology,
    Omiya Research Laboratory, Nikken Chemicals Co., Ltd.

    Shimizu, M., Katoh, M., Imamura, M. & Modderman, J. (1996) Teratology
    study of erythritol in rabbits.  Regul. Toxicol. Pharmacol., 24,
    S247-S253.

    Smits-van Prooije, A.E., Waalkens-Berendsen, D.H. & Wijnands, M.V.W.
    (1996a) Dietary two-generation reproduction study with erythritol in
    rats (TNO Report No. V92.552). Unpublished report from TNO-CIVO
    Industries, Netherlands Organization for Applied Scientific Research,
    Zeist.

    Smits-Van Prooije, A.E., Waalkens-Berendsen, D.H. & Bär, A. (1996b)
    Embryotoxicity and teratogenicity study with erythritol in rats.
     Regul. Toxicol. Pharmacol., 24, S232-S236.

    Takahashi, C. (1992a) Study on the maximum no-effect level of
    erythritol using transient diarrheal action as index. Unpublished
    report from Nikken Chemicals Co., Ltd.

    Takahashi, C. (1992b) The effect of continuous ingestion of erythritol
    with laxative action serving as index. Unpublished report from Nikken
    Chemicals Co., Ltd.

    Tateishi, T., Yamamoto, S., Kagawa, M., Kosuge, M., Takahashi, K.,
    Nakano, S. & Kasai, Y. (1989) Oral reproduction study of erythritol
    (NIK-242) with mice prior to and in the early stages of pregnancy.
    Unpublished report from Omiya Research Lab, Nikken Chemicals Co., Ltd.

    Tateishi, T., Yamamoto, S., Kagawa, M., Mizutani, M., Kosuge, M.,
    Takahashi, K., Itoh, K. & Kasai, Y. (1992) Fertility study of NIK-242
    in ICR strain mice (intravenous dosing). Unpublished report from
    Toxicological Laboratory, Ohmiya Research Laboratory, Nikken Chemicals
    Co., Ltd.

    Tetzloff, W., Dauchy, F., Medimagh, S., Carr, D. & Bär, A. (1996)
    Tolerance to subchronic, high-dose ingestion of erythritol in human
    volunteers.  Regul. Toxicol. Pharmacol., 24, S286-S295.

    Til, H.P. & Modderman, J. (1996) Four-week oral toxicity study with
    erythritol in rats. Regul. Toxicol. Pharmacol., 24, S214-S220.

    Til, H.P. & van Nesselrooij, J.H.J. (1994) Chronic (78-week) oral
    toxicity study with erythritol in rats (Final report) (TNO Report No.
    V93.367). Unpublished report from TNO-CIVO Industries, Netherlands
    Organization for Applied Scientific Research, Zeist.

    Til, H.P. & Wijnands, M.V.W. (1991) Sub-acute (28-day) oral toxicity
    study with erythritol in rats (Final report) (TNO Report No. V91.050).
    Unpublished report from TNO-CIVO Industries, Netherlands Organization
    for Applied Scientific Research, Zeist.

    Til, H.P., Falke, H.E. & Kuiper, C.F. (1991) An exploratory,
    subchronic feeding study with erythritol in adult male rats (Final
    report) (TNO Report No. V90.003). Unpublished report from TNO-CIVO
    Industries, Netherlands Organization for Applied Scientific Research,
    Zeist.

    Til, H.P., Kuiper, C.F. & Bruyntjes, J.P. (1992) Sub-chronic (13-week)
    feeding study with erythritol in mice (Final report) (TNO Report No.
    V90.421). Unpublished report from TNO-CIVO Industries, Netherlands
    Organization for Applied Scientific Research, Zeist.

    Til, H.P., Kuiper, C.F., Falke, H.E. & Bär, A. (1996) Subchronic oral
    toxicity studies with erythritol in mice and rats.  Regul. Toxicol.
     Pharmacol., 24, S221-S231.

    Umeki, Y. (1992) Study concerning transient diarrhea. Unpublished
    report from Food Business Department, Mitsubishi Kasei Corp.

    Waalkens-Berendsen, D.H., Smits-Van Prooije, A.E., Wijnands, M.V.M. &
    Bär, A. (1996) Two-generation reproduction study of erythritol in
    rats.  Regul. Toxicol. Pharmacol., 24, S237-S246.

    Yamaguchi, T., Kimura, H., Ito, K., Shintani, S., Morita, K.,
    Kitamura, S., Kanai, M., Motegi, S., Hirai, A. & Kagawa, A. (1990) A
    13-week oral subacute toxicity study of NIK-242 in dogs with 4-week
    recovery period (PRL/35). Unpublished report from Omiya Research
    Laboratory, Nikken Chemicals Co., Ltd.

    Yamamoto, H., Tateishi, T., Sadamasu, K., Kosuge, M., Takahashi, K.,
    Nakano, S. & Kasai, Y. (1987) Acute intraveneous, subcutaneous and
    oral toxicity study with NIK-242 in rats. Unpublished report from
    Division of Toxicology, Omiya Research Laboratory, Nikken Chemicals
    Co., Ltd.

    Yamamoto, H., Tateishi, T., Touchi, A., Kosuge, M., Takahashi, K.,
    Takahashi, T., Nakano, S. & Kasai, Y. (1989) A 13-week oral subacute
    toxicity study of NIK-242 in rats with 4-week recovery period
    (PRL/34). Unpublished report from Omiya Research Laboratory, Nikken
    Chemicals Co., Ltd.
    


    See Also:
       Toxicological Abbreviations